Proposal for P2P Wireless (Bluetooth LE) transfer of Payment URI
43 messages, 5 to 10 Feb 2015. Started by Paul Puey.
Eric Voskuil (15), Paul Puey (10), Martin Habovštiak (7), Mike Hearn (5), Roy Badami (3), Andreas Schildbach, Brian Hoffman, William Swanson
Airbitz has developed and implemented a method for communicating a bitcoin URI across Bluetooth (BLE) or any other P2P, mid range, wireless, broadcast medium. The currently documented implementation is available in our iOS and Android mobile wallet (updated Android version with BLE coming in about 1 week). We would like to have the BIP pulled into Github for review and discussion. Here is the current BIP:
BIP: TBD
Title: P2P Wireless URI transfer
Authors: Thomas Baker <tom’at’airbitz.co>, Paul Puey <paul’at’airbitz.co>
Contributors: Joey Krug <joeykrug’at’gmail.com>
Status: proposal
Type: Standards Track
Created: 2015-01-12
Table of Contents
-
Abstract -
Motivation -
Specification -
Compatibility -
Examples -
References
Abstract
This is a protocol for peer-to-peer wireless transfer of a URI request using an open broadcast or advertisement channel such as Bluetooth, Bluetooth Low Energy, or WiFi Direct. Motivation
There are disadvantages for a merchant (requester) and customer (sender) to exchange a URI request using QR codes that can be eliminated by using wireless broadcast or advertisements.
Current QR code scan method to transfer a request URI from merchant (Requester) to customer (Sender) is cumbersome. A usual scenario is a merchant with a POS terminal for order entry and a separate tablet for transacting payments with bitcoin, and a customer with a smartphone. After the order is entered, the merchant enters payment request information into the tablet, generates the QR code representing the URI, and presents this to the customer. The customer prepares to scan the QR code with their smartphone by maneuvering the camera to the tablet. The tablet screen must be relatively clean, point at the customer, and held steady. The smartphone camera lens must be clean, point at the tablet screen, come into range, and held steady to focus and wait for a QR scan. Environmental conditions such as bright outdoor sunlight, indoor spot lights, or significant distance between QR code and camera can create difficult and cumbersome experiences for users.
Using a wireless local broadcast allows the merchant to just enter the payment and wait. The tablet and smartphone are not maneuvered to align in any way. The customer observes broadcast listings, selects the appropriate one from possible simultaneous broadcasts from other POS stations nearby, examines the URI request details such as amount, and decides whether to send funds, initiating a bitcoin network transfer. The merchant and customer then receive the transaction confirmations and are done with the sale. Merchant and customer devices are kept private and secured in their own possession.
The URI and other broadcast identification (Joe’s Grill #1) only contain public information. However, a copycat broadcaster acting as MITM might duplicate the broadcast simultaneously as the merchant, attempting to lure the customer to send funds to the copycat. That attack is mitigated with this broadcast method because of the partial address in the broadcast. Specification
Requester generates a bitcoin URI request of variable length, and a limited descriptive identifier string. Requester then broadcasts the URI’s partial public address (<paddress>) plus identifier (<id>) over a publicly visible wireless channel.
Sender scans for broadcasts on their device, examines and selects the desired request by the identifier and partial address. This connects a data channel to Requester.
Requester sends full URI back over the data channel.
Sender device ensures <paddress> is part of the full URI public address and checks the full address integrity. Checking the broadcast and full URI integrity prevents a copycat device within range from copying the partial address and fooling the customer into sending funds to the copycat instead.
Below is a description of the protocol through Bluetooth Smart (Low Energy).
Requestor Sender - Bitcoin transaction roles
Peripheral Central - Bluetooth GAP definitions
Mode Mode
1 |------------->| - Requestor Advertises partial bitcoin: URI + Name
| ... |
2 |<-------------| - Subscribe then send sender's Name, requesting a response
3 |------------->| - ACK
4 |<-------------| - request Read Characteristic from peripheral
5 |------------->| - Sender receives full bitcoin: URI
1.
Peripheral advertises over a service UUID a BLE extended advertisement with a Scan Response containing the partial address of a bitcoin URI and a Name, any plain text. The entire response is limited to 26 characters. The first 10 make up the first 10 characters of the bitcoin URI public address where to send bitcoin, and must be present. The remaining characters are any plain text such as “The Habit 1” or “Starbucks-Reg 1”, more human readable information. The partial address serves as a check against a nearby attacker who may try to lure a Sender into sending payment to a separate wallet by advertising a similar Scan Response but cannot replicate a public address with the same leading 10 characters and different trailing characters. 2.
When the Central scans the advertisement, it may display the Scan Response in a human readable listing using the two pieces of information. If Central chooses this advertisement to receive the full request, it then subscribes to the service and writes the characteristic (a second UUID) with it’s own name, or a blank if not sending a name, to the Peripheral. 3.
Peripheral gets a characteristic write request of the Central’s name, and acknowledges the receipt by sending a server response. 4.
Central receives a characteristic write (from the response) and immediately requests the entire bitcoin URI by issuing a read request on that characteristic. 5.
Peripheral receives the read request and sends the entire bitcoin URI over that characteristic up to 512 bytes.
This ends the proposed specification as the bitcoin URI transfer is complete. The Sender would then normally confirm the request and decide whether to initiate the fund transfer. Compatibility
There are no prior BIPs covering this. Examples
Airbitz iOS Bluetooth Low Energy to Bluetooth Low Energy request transfer. References
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
Thanks Paul, for writing up your protocol!
First thoughts:
For a BIP standard, I think we should skip "bitcoin:" URIs entirely and publish BIP70 payment requests instead. URIs mainly stick around because of QR codes limited capacity. BIP70 would partly address the "copycat" problem by signing payment requests.
In your Motivation section, I miss some words about NFC. NFC already addresses all of the usability issues mentioned and is supported by mobile wallets since 2011. That doesn't mean your method doesn't make sense in some situations, but I think it should be explained why to prefer broadcasting payment requests over picking them up via near field radio.
On 02/05/2015 09:01 AM, Paul Puey wrote:
Quoted text (226 lines)
> Airbitz has developed and implemented a method for communicating a > bitcoin URI across Bluetooth (BLE) or any other P2P, mid range, > wireless, broadcast medium. The currently documented implementation is > available in our iOS and Android mobile wallet (updated Android version > with BLE coming in about 1 week). We would like to have the BIP pulled > into Github for review and discussion. Here is the current BIP: > > > BIP: TBD > > Title: P2P Wireless URI transfer > > Authors: Thomas Baker <tom’at’airbitz.co <http://airbitz.co>>, Paul Puey > <paul’at’airbitz.co <http://airbitz.co>> > > Contributors: Joey Krug <joeykrug’at’gmail.com <http://gmail.com>> > > Status: proposal > > Type: Standards Track > > Created: 2015-01-12 > > > Table of Contents > > * > > Abstract > > * > > Motivation > > * > > Specification > > * > > Compatibility > > * > > Examples > > * > > References > > > Abstract > > This is a protocol for peer-to-peer wireless transfer of a URI request > using an open broadcast or advertisement channel such as Bluetooth, > Bluetooth Low Energy, or WiFi Direct. > > > Motivation > > There are disadvantages for a merchant (requester) and customer (sender) > to exchange a URI request using QR codes that can be eliminated by using > wireless broadcast or advertisements. > > Current QR code scan method to transfer a request URI from merchant > (Requester) to customer (Sender) is cumbersome. A usual scenario is a > merchant with a POS terminal for order entry and a separate tablet for > transacting payments with bitcoin, and a customer with a smartphone. > After the order is entered, the merchant enters payment request > information into the tablet, generates the QR code representing the URI, > and presents this to the customer. The customer prepares to scan the QR > code with their smartphone by maneuvering the camera to the tablet. The > tablet screen must be relatively clean, point at the customer, and held > steady. The smartphone camera lens must be clean, point at the tablet > screen, come into range, and held steady to focus and wait for a QR > scan. Environmental conditions such as bright outdoor sunlight, indoor > spot lights, or significant distance between QR code and camera can > create difficult and cumbersome experiences for users. > > Using a wireless local broadcast allows the merchant to just enter the > payment and wait. The tablet and smartphone are not maneuvered to align > in any way. The customer observes broadcast listings, selects the > appropriate one from possible simultaneous broadcasts from other POS > stations nearby, examines the URI request details such as amount, and > decides whether to send funds, initiating a bitcoin network transfer. > The merchant and customer then receive the transaction confirmations and > are done with the sale. Merchant and customer devices are kept private > and secured in their own possession. > > The URI and other broadcast identification (Joe’s Grill #1) only contain > public information. However, a copycat broadcaster acting as MITM might > duplicate the broadcast simultaneously as the merchant, attempting to > lure the customer to send funds to the copycat. That attack is mitigated > with this broadcast method because of the partial address in the broadcast. > > > Specification > > Requester generates a bitcoin URI request of variable length, and a > limited descriptive identifier string. Requester then broadcasts the > URI’s partial public address (<paddress>) plus identifier (<id>) over a > publicly visible wireless channel. > > Sender scans for broadcasts on their device, examines and selects the > desired request by the identifier and partial address. This connects a > data channel to Requester. > > Requester sends full URI back over the data channel. > > Sender device ensures <paddress> is part of the full URI public address > and checks the full address integrity. Checking the broadcast and full > URI integrity prevents a copycat device within range from copying the > partial address and fooling the customer into sending funds to the > copycat instead. > > Below is a description of the protocol through Bluetooth Smart (Low Energy). > > Requestor Sender - Bitcoin transaction roles > > Peripheral Central - Bluetooth GAP definitions > > Mode Mode > > 1 |------------->| - Requestor Advertises partial bitcoin: URI + > Name > > | ... | > > 2 |<-------------| - Subscribe then send sender's Name, > requesting a response > > 3 |------------->| - ACK > > 4 |<-------------| - request Read Characteristic from peripheral > > 5 |------------->| - Sender receives full bitcoin: URI > > > 1. > > Peripheral advertises over a service UUID a BLE extended > advertisement with a Scan Response containing the partial address of > a bitcoin URI and a Name, any plain text. The entire response is > limited to 26 characters. The first 10 make up the first 10 > characters of the bitcoin URI public address where to send bitcoin, > and must be present. The remaining characters are any plain text > such as “The Habit 1” or “Starbucks-Reg 1”, more human readable > information. The partial address serves as a check against a nearby > attacker who may try to lure a Sender into sending payment to a > separate wallet by advertising a similar Scan Response but cannot > replicate a public address with the same leading 10 characters and > different trailing characters. > > 2. > > When the Central scans the advertisement, it may display the Scan > Response in a human readable listing using the two pieces of > information. If Central chooses this advertisement to receive the > full request, it then subscribes to the service and writes the > characteristic (a second UUID) with it’s own name, or a blank if not > sending a name, to the Peripheral. > > 3. > > Peripheral gets a characteristic write request of the Central’s > name, and acknowledges the receipt by sending a server response. > > 4. > > Central receives a characteristic write (from the response) and > immediately requests the entire bitcoin URI by issuing a read > request on that characteristic. > > 5. > > Peripheral receives the read request and sends the entire bitcoin > URI over that characteristic up to 512 bytes. > > This ends the proposed specification as the bitcoin URI transfer is > complete. The Sender would then normally confirm the request and decide > whether to initiate the fund transfer. > > > Compatibility > > There are no prior BIPs covering this. > > > Examples > > Airbitz iOS Bluetooth Low Energy to Bluetooth Low Energy request transfer. > > > References > > > > > > logo > *Paul Puey* CEO / Co-Founder, Airbitz Inc > +1-619-850-8624 | http://airbitz.co <http://airbitz.co/> | San Diego > <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> > > *DOWNLOAD THE AIRBITZ WALLET:* > > <https://play.google.com/store/apps/details?id=com.airbitz><https://itunes.apple.com/us/app/airbitz/id843536046> > > > > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > > > > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development >
Quoted text (4 lines)
> > For a BIP standard, I think we should skip "bitcoin:" URIs entirely and > publish BIP70 payment requests instead. >
Agreed - it's not clear to me at all that this partial address scheme is actually secure. The assumption appears to be that the MITM must match the address prefix generated by the genuine merchant. But if they can do a wireless MITM they can just substitute their own address prefix/partial address, no?
To avoid MITM attacks the sender must know who they are sending money to, and that means they must see a human understandable name that's cryptographically bound to the right public key. Displaying partial addresses to the user is not going to solve this unless users manually compare key prefixes across the screens.... which is even less convenient than a QR code.
I think it should be explained why to
Quoted text (3 lines)
> prefer broadcasting payment requests over picking them up via near field > radio. >
This is probably an artifact of Apple's restrictions on iOS. Only the iPhone 6 has NFC hardware and Apple don't expose it via any public API. It can however support Bluetooth LE.
Apple isn't a big deal in Germany because iPhone only achieved about 17% market share during the quarter when the iPhone 6 launched. Normally it's closer to 10-13%. Most other markets are similar.
However in the USA, UK, Australia and Japan iOS is still a big deal and NFC is going to be seen as a non-universal solution there. At least, until Apple catches up and provides an NFC API.
It's certainly not a problem to have a working radio based broadcast system, though the theoretician in me wonders what happens when lots of people are trying to pay simultaneously for something that has equal cost ..... e.g. buying movie tickets at a counter. NFC and QR codes prevent any kind of "oops I paid for someone elses stuff" confusion.
In practice of course Bitcoin payments are not normally popular enough for this to be a problem outside of Bitcoin community events.
The BIP70 protocol would preclude individuals from utilizing the P2P transfer spec. It would also require that a Sender have internet connectivity to get the payment protocol info. BLE could enable payment w/o internet by first transferring the URI to from Recipient to Sender. Then in the future, we could sign a Tx and send it over BLE back to the recipient (who would still need internet to verify the Tx). This is an important use case for areas with poor 3G/4G connectivity as I've experience myself.
Also, due to Android issues, NFC is incredibly clunky. The URI Sender is required to tap the screen *while* the two phones are in contact. We support NFC the same way Bitcoin Wallet does, but unless the payment recipient has a custom Android device (which a merchant might) then the usage model is worse than scanning a QR code. BLE also allows people to pay at a distance such as for a donation to a live performer. We'll look at adding this to the Motivation section.
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
From: Andreas Schildbach <andreas@sc...> - 2015-02-05 13:47:04
Thanks Paul, for writing up your protocol!
First thoughts:
For a BIP standard, I think we should skip "bitcoin:" URIs entirely and publish BIP70 payment requests instead. URIs mainly stick around because of QR codes limited capacity. BIP70 would partly address the "copycat" problem by signing payment requests.
In your Motivation section, I miss some words about NFC. NFC already addresses all of the usability issues mentioned and is supported by mobile wallets since 2011. That doesn't mean your method doesn't make sense in some situations, but I think it should be explained why to prefer broadcasting payment requests over picking them up via near field radio.
BIP70 requests can be sent over bluetooth as well, as can transactions. Bitcoin Wallet can already send money even when offline by doing this. It's transparent to the user. I mean original Bluetooth in this context - BLE has incredibly tight data constraints and isn't really meant for data transfer.
Yes Android Beam has a pretty stupid UI. You can actually tap the devices, take them away and then press, but that's not obvious at all. There have been new APIs added in recent releases that give more control over this, so it's possible we can revisit things and make the UI better these days.
The donation to live performer example is good - there's no issue of accidentally paying for someone else in this context as there's only one recipient, but many senders.
The issue of confused payments remains in other situations though.
For the coffee shop use case, it'd be nicer (I think) if we aim for a Square-style UI where the device broadcasts a (link to) a photo of the user combined with a bluetooth MAC. Then the merchant tablet can show faces of people in the shop, and can push a payment request to the users device. That device can then buzz the user, show a confirmation screen, put something on their smart watch etc or just auto-authorise the payment because the BIP70 signature is from a trusted merchant. User never even needs to touch their phone at all.
On Thu, Feb 5, 2015 at 9:06 PM, Paul Puey <paul@airbitz•co> wrote:
Quoted text (62 lines)
> The BIP70 protocol would preclude individuals from utilizing the P2P > transfer spec. It would also require that a Sender have internet > connectivity to get the payment protocol info. BLE could enable payment w/o > internet by first transferring the URI to from Recipient to Sender. Then in > the future, we could sign a Tx and send it over BLE back to the recipient > (who would still need internet to verify the Tx). This is an important use > case for areas with poor 3G/4G connectivity as I've experience myself. > > Also, due to Android issues, NFC is incredibly clunky. The URI Sender is > required to tap the screen *while* the two phones are in contact. We > support NFC the same way Bitcoin Wallet does, but unless the payment > recipient has a custom Android device (which a merchant might) then the > usage model is worse than scanning a QR code. BLE also allows people to pay > at a distance such as for a donation to a live performer. We'll look at > adding this to the Motivation section. > > [image: logo] > *Paul Puey* CEO / Co-Founder, Airbitz Inc > +1-619-850-8624 | http://airbitz.co | San Diego > <http://facebook.com/airbitz> <http://twitter.com/airbitz> > <https://plus.google.com/118173667510609425617> > <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> > <https://angel.co/paul-puey> > *DOWNLOAD THE AIRBITZ WALLET:* > <https://play.google.com/store/apps/details?id=com.airbitz> > <https://itunes.apple.com/us/app/airbitz/id843536046> > > > From: Andreas Schildbach <andreas@sc...> - 2015-02-05 13:47:04 > > Thanks Paul, for writing up your protocol! > > First thoughts: > > For a BIP standard, I think we should skip "bitcoin:" URIs entirely and > publish BIP70 payment requests instead. URIs mainly stick around because > of QR codes limited capacity. BIP70 would partly address the "copycat" > problem by signing payment requests. > > In your Motivation section, I miss some words about NFC. NFC already > addresses all of the usability issues mentioned and is supported by > mobile wallets since 2011. That doesn't mean your method doesn't make > sense in some situations, but I think it should be explained why to > prefer broadcasting payment requests over picking them up via near field > radio. > > > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is > your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development > >
Thanks for CC'ing me Mike. Having trouble receiving maillist list posts.
Even if a user could get the BIP70 URL in the URI, they would still need internet to access the URL. This BLE spec doesn't preclude BIP70, but can work with it while still allowing individuals without a certificate to broadcast a request.
The issue of confused payments becomes less so if the Recipient broadcasts a name along with the 10 digit public addr prefix. Only if there is a name conflict will the user have to be concerned with the prefix. The name can be something like
Mikes Coffee #1 and it can have a "Register #1" at the counter. A customer facing screen can also show the 10 digit prefix.
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
On Thu, Feb 5, 2015 at 12:28 PM, Mike Hearn <mike@plan99•net> wrote:
Quoted text (91 lines)
> BIP70 requests can be sent over bluetooth as well, as can transactions. > Bitcoin Wallet can already send money even when offline by doing this. It's > transparent to the user. I mean original Bluetooth in this context - BLE > has incredibly tight data constraints and isn't really meant for data > transfer. > > Yes Android Beam has a pretty stupid UI. You can actually tap the devices, > take them away and then press, but that's not obvious at all. There have > been new APIs added in recent releases that give more control over this, so > it's possible we can revisit things and make the UI better these days. > > The donation to live performer example is good - there's no issue of > accidentally paying for someone else in this context as there's only one > recipient, but many senders. > > The issue of confused payments remains in other situations though. > > For the coffee shop use case, it'd be nicer (I think) if we aim for a > Square-style UI where the device broadcasts a (link to) a photo of the user > combined with a bluetooth MAC. Then the merchant tablet can show faces of > people in the shop, and can push a payment request to the users device. > That device can then buzz the user, show a confirmation screen, put > something on their smart watch etc or just auto-authorise the payment > because the BIP70 signature is from a trusted merchant. User never even > needs to touch their phone at all. > > On Thu, Feb 5, 2015 at 9:06 PM, Paul Puey <paul@airbitz•co> wrote: > >> The BIP70 protocol would preclude individuals from utilizing the P2P >> transfer spec. It would also require that a Sender have internet >> connectivity to get the payment protocol info. BLE could enable payment w/o >> internet by first transferring the URI to from Recipient to Sender. Then in >> the future, we could sign a Tx and send it over BLE back to the recipient >> (who would still need internet to verify the Tx). This is an important use >> case for areas with poor 3G/4G connectivity as I've experience myself. >> >> Also, due to Android issues, NFC is incredibly clunky. The URI Sender is >> required to tap the screen *while* the two phones are in contact. We >> support NFC the same way Bitcoin Wallet does, but unless the payment >> recipient has a custom Android device (which a merchant might) then the >> usage model is worse than scanning a QR code. BLE also allows people to pay >> at a distance such as for a donation to a live performer. We'll look at >> adding this to the Motivation section. >> >> [image: logo] >> *Paul Puey* CEO / Co-Founder, Airbitz Inc >> +1-619-850-8624 | http://airbitz.co | San Diego >> <http://facebook.com/airbitz> <http://twitter.com/airbitz> >> <https://plus.google.com/118173667510609425617> >> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> >> <https://angel.co/paul-puey> >> *DOWNLOAD THE AIRBITZ WALLET:* >> <https://play.google.com/store/apps/details?id=com.airbitz> >> <https://itunes.apple.com/us/app/airbitz/id843536046> >> >> >> From: Andreas Schildbach <andreas@sc...> - 2015-02-05 13:47:04 >> >> Thanks Paul, for writing up your protocol! >> >> First thoughts: >> >> For a BIP standard, I think we should skip "bitcoin:" URIs entirely and >> publish BIP70 payment requests instead. URIs mainly stick around because >> of QR codes limited capacity. BIP70 would partly address the "copycat" >> problem by signing payment requests. >> >> In your Motivation section, I miss some words about NFC. NFC already >> addresses all of the usability issues mentioned and is supported by >> mobile wallets since 2011. That doesn't mean your method doesn't make >> sense in some situations, but I think it should be explained why to >> prefer broadcasting payment requests over picking them up via near field >> radio. >> >> >> >> >> ------------------------------------------------------------------------------ >> Dive into the World of Parallel Programming. The Go Parallel Website, >> sponsored by Intel and developed in partnership with Slashdot Media, is >> your >> hub for all things parallel software development, from weekly thought >> leadership blogs to news, videos, case studies, tutorials and more. Take a >> look and join the conversation now. http://goparallel.sourceforge.net/ >> _______________________________________________ >> Bitcoin-development mailing list >> Bitcoin-development@lists•sourceforge•net >> https://lists.sourceforge.net/lists/listinfo/bitcoin-development >> >> >
Quoted text (4 lines)
> > Even if a user could get the BIP70 URL in the URI, they would still need > internet to access the URL. >
The way Bitcoin Wallet does it, the bitcoin URI includes a MAC address where you can download the request from. BIP70 does not depend on internet access or HTTP, plus, you don't have to sign them.
The name field might work but requires the merchant to set it, e.g. by asking the payer what their name is, then typing it in, then the payer has to wait for it to show up. By this point it's probably faster to have scanned a QR code.
Re: security. I'll repeat what I wrote up-thread in case you didn't see it:
it's not clear to me at all that this partial address scheme is actually
Quoted text (11 lines)
> secure. The assumption appears to be that the MITM must match the address > prefix generated by the genuine merchant. But if they can do a wireless > MITM they can just substitute their own address prefix/partial address, no? > > To avoid MITM attacks the sender must know who they are sending money to, > and that means they must see a human understandable name that's > cryptographically bound to the right public key. Displaying partial > addresses to the user is not going to solve this unless users manually > compare key prefixes across the screens.... which is even less convenient > than a QR code. >
On 02/05/2015 12:28 PM, Mike Hearn wrote:
Quoted text (3 lines)
> The donation to live performer example is good - there's no issue of > accidentally paying for someone else in this context as there's only one > recipient, but many senders.
I'm not sure you could assume this, even if the payer only received one broadcast. And if the payer receives multiple, it constitutes a DOS on the scenario, potentially unintentional.
Quoted text (1 line)
> The issue of confused payments remains in other situations though.
Agree, the problem of the payer strongly identifying the receiver requires either proximity (NFC or QR code scan from the known-good source) or PKI/WoT. The problem can't be resolved through a broadcast.
Quoted text (8 lines)
> For the coffee shop use case, it'd be nicer (I think) if we aim for a > Square-style UI where the device broadcasts a (link to) a photo of the > user combined with a bluetooth MAC. Then the merchant tablet can show > faces of people in the shop, and can push a payment request to the users > device. That device can then buzz the user, show a confirmation screen, > put something on their smart watch etc or just auto-authorise the > payment because the BIP70 signature is from a trusted merchant. User > never even needs to touch their phone at all.
I'm imagining myself walking around broadcasting my photo and MAC address while hucksters push payment requests to me for approval, while recording my photo and correlating it to my address. It will pretty quickly turn in to a scenario where I need to touch something before this is turned on.
Quoted text (37 lines)
> On Thu, Feb 5, 2015 at 9:06 PM, Paul Puey <paul@airbitz•co > <mailto:paul@airbitz.co>> wrote: > > The BIP70 protocol would preclude individuals from utilizing the P2P > transfer spec. It would also require that a Sender have internet > connectivity to get the payment protocol info. BLE could enable > payment w/o internet by first transferring the URI to from Recipient > to Sender. Then in the future, we could sign a Tx and send it over > BLE back to the recipient (who would still need internet to verify > the Tx). This is an important use case for areas with poor 3G/4G > connectivity as I've experience myself. > > Also, due to Android issues, NFC is incredibly clunky. The URI > Sender is required to tap the screen *while* the two phones are in > contact. We support NFC the same way Bitcoin Wallet does, but unless > the payment recipient has a custom Android device (which a merchant > might) then the usage model is worse than scanning a QR code. BLE > also allows people to pay at a distance such as for a donation to a > live performer. We'll look at adding this to the Motivation section. > > From: Andreas Schildbach <andreas@sc...> - 2015-02-05 13:47:04 > > Thanks Paul, for writing up your protocol! > > First thoughts: > > For a BIP standard, I think we should skip "bitcoin:" URIs entirely and > publish BIP70 payment requests instead. URIs mainly stick around because > of QR codes limited capacity. BIP70 would partly address the "copycat" > problem by signing payment requests. > > In your Motivation section, I miss some words about NFC. NFC already > addresses all of the usability issues mentioned and is supported by > mobile wallets since 2011. That doesn't mean your method doesn't make > sense in some situations, but I think it should be explained why to > prefer broadcasting payment requests over picking them up via near field > radio.
Quoted text (3 lines)
> > I'm imagining myself walking around broadcasting my photo and MAC > address while hucksters push payment requests to me for approval
I hate to break it to you, but you broadcast a photo of your face every time you walk outside ;)
Bluetooth MAC addresses are random, they aren't useful identifiers. If someone can see you, a face is a far more uniquely identifying thing than a MAC.
"Payment spam" might be a problem. I can imagine a wallet requiring that such requests are signed and then spammers can be blacklisted in the usual fashion so they can't push things to your phone anymore. Anyway, a hurdle that can be jumped if/when it becomes an issue.
On 02/05/2015 12:50 PM, Mike Hearn wrote:
Quoted text (9 lines)
> I'm imagining myself walking around broadcasting my photo and MAC > address while hucksters push payment requests to me for approval > > I hate to break it to you, but you broadcast a photo of your face every > time you walk outside ;) > > Bluetooth MAC addresses are random, they aren't useful identifiers. If > someone can see you, a face is a far more uniquely identifying thing > than a MAC.
Interesting take on privacy. But the market will of course decide.
Would the merchant be broadcasting payment requests in the clear, or would they be encrypted with a public key of the spender?
Quoted text (4 lines)
> "Payment spam" might be a problem. I can imagine a wallet requiring that > such requests are signed and then spammers can be blacklisted in the > usual fashion so they can't push things to your phone anymore. Anyway, a > hurdle that can be jumped if/when it becomes an issue.
Not sure I'd shoot for a system that's guaranteed to require PKI with blacklisting and spam filtering.
e
This sounds horrible. You could basically monitor anyone with a wallet in a highly populated area and track them super easily by doing facial recognition. Yes you could photograph people but it's way more burdensome. Sorry to go off topic a little.
On Feb 5, 2015, at 3:50 PM, Mike Hearn <mike@plan99•net> wrote:
Quoted text (18 lines)
>> I'm imagining myself walking around broadcasting my photo and MAC >> address while hucksters push payment requests to me for approval > > I hate to break it to you, but you broadcast a photo of your face every time you walk outside ;) > > Bluetooth MAC addresses are random, they aren't useful identifiers. If someone can see you, a face is a far more uniquely identifying thing than a MAC. > > "Payment spam" might be a problem. I can imagine a wallet requiring that such requests are signed and then spammers can be blacklisted in the usual fashion so they can't push things to your phone anymore. Anyway, a hurdle that can be jumped if/when it becomes an issue. > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
Yes, a stellar device for mass surveillance coupled with transaction tainting.
e
Quoted text (25 lines)
> On Feb 5, 2015, at 1:19 PM, Brian Hoffman <brianchoffman@gmail•com> wrote: > > This sounds horrible. You could basically monitor anyone with a wallet in a highly populated area and track them super easily by doing facial recognition. Yes you could photograph people but it's way more burdensome. Sorry to go off topic a little. > > > On Feb 5, 2015, at 3:50 PM, Mike Hearn <mike@plan99•net> wrote: > >>> I'm imagining myself walking around broadcasting my photo and MAC >>> address while hucksters push payment requests to me for approval >> >> I hate to break it to you, but you broadcast a photo of your face every time you walk outside ;) >> >> Bluetooth MAC addresses are random, they aren't useful identifiers. If someone can see you, a face is a far more uniquely identifying thing than a MAC. >> >> "Payment spam" might be a problem. I can imagine a wallet requiring that such requests are signed and then spammers can be blacklisted in the usual fashion so they can't push things to your phone anymore. Anyway, a hurdle that can be jumped if/when it becomes an issue. >> ------------------------------------------------------------------------------ >> Dive into the World of Parallel Programming. The Go Parallel Website, >> sponsored by Intel and developed in partnership with Slashdot Media, is your >> hub for all things parallel software development, from weekly thought >> leadership blogs to news, videos, case studies, tutorials and more. Take a >> look and join the conversation now. http://goparallel.sourceforge.net/ >> _______________________________________________ >> Bitcoin-development mailing list >> Bitcoin-development@lists•sourceforge•net >> https://lists.sourceforge.net/lists/listinfo/bitcoin-development
Quoted text (5 lines)
> > This sounds horrible. You could basically monitor anyone with a wallet in > a highly populated area and track them super easily by doing facial > recognition. >
We're talking about BLE, still? The radio tech that runs in the so called "junk bands" because propagation is so poor?
My watch loses its connection to my phone if I just put it down and walk around my apartment. I'm all for reasonable paranoia, but Bluetooth isn't going to be enabling mass surveillance any time soon. It barely goes through air, let alone walls.
Anyway, whatever. I'm just bouncing around ideas for faster user interfaces. You could always switch it off or set it to be triggered by the presence of particular wifi hotspots, if you don't mind an initial bit of setup.
Back on topic - the debate is interesting, but I think to get this to the stage of being a BIP we'd need at least another wallet to implement it? Then I guess a BIP would be useful regardless of the design issues. The prefix matching still feels flaky to me but it's hard to know if you could really swipe payments out of the air in practice, without actually trying it.
BLE has an advertised range of over 100m.
http://www.bluetooth.com/Pages/low-energy-tech-info.aspx
In the case of mass surveillance that range could most likely be extended dramatically by the reviewer. I've seen WiFi ranges of over a mile with a strong (not FCC approved) receiver.
WiFi hotspots don't have strong identity or a guaranteed position, so they can't be trusted for location.
e
On Feb 5, 2015, at 1:36 PM, Mike Hearn <mike@plan99•net> wrote:
Quoted text (11 lines)
>> This sounds horrible. You could basically monitor anyone with a wallet in a highly populated area and track them super easily by doing facial recognition. > > We're talking about BLE, still? The radio tech that runs in the so called "junk bands" because propagation is so poor? > > My watch loses its connection to my phone if I just put it down and walk around my apartment. I'm all for reasonable paranoia, but Bluetooth isn't going to be enabling mass surveillance any time soon. It barely goes through air, let alone walls. > > Anyway, whatever. I'm just bouncing around ideas for faster user interfaces. You could always switch it off or set it to be triggered by the presence of particular wifi hotspots, if you don't mind an initial bit of setup. > > Back on topic - the debate is interesting, but I think to get this to the stage of being a BIP we'd need at least another wallet to implement it? Then I guess a BIP would be useful regardless of the design issues. The prefix matching still feels flaky to me but it's hard to know if you could really swipe payments out of the air in practice, without actually trying it. > >
The broadcast is ONLY done when the wallet is in Receive mode. Same as when the QR code is visible. The use of the *Name* section is specifically so that a recipient can broadcast their name/handle. Not so the recipient would broadcast the name of the Sender.
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
On Thu, Feb 5, 2015 at 12:50 PM, Mike Hearn <mike@plan99•net> wrote:
Quoted text (16 lines)
> I'm imagining myself walking around broadcasting my photo and MAC >> address while hucksters push payment requests to me for approval > > > I hate to break it to you, but you broadcast a photo of your face every > time you walk outside ;) > > Bluetooth MAC addresses are random, they aren't useful identifiers. If > someone can see you, a face is a far more uniquely identifying thing than a > MAC. > > "Payment spam" might be a problem. I can imagine a wallet requiring that > such requests are signed and then spammers can be blacklisted in the usual > fashion so they can't push things to your phone anymore. Anyway, a hurdle > that can be jumped if/when it becomes an issue. >
The implementation on Airbitz does not encourage or even let a user broadcast a photo. Just an address prefix and "name/handle". And it's only broadcast during the Receive request. Not generally while the app is running although that's up to the implementation.
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
On Thu, Feb 5, 2015 at 1:19 PM, Brian Hoffman <brianchoffman@gmail•com> wrote:
Quoted text (39 lines)
> This sounds horrible. You could basically monitor anyone with a wallet in > a highly populated area and track them super easily by doing facial > recognition. Yes you could photograph people but it's way more burdensome. > Sorry to go off topic a little. > > > On Feb 5, 2015, at 3:50 PM, Mike Hearn <mike@plan99•net> wrote: > > I'm imagining myself walking around broadcasting my photo and MAC >> address while hucksters push payment requests to me for approval > > > I hate to break it to you, but you broadcast a photo of your face every > time you walk outside ;) > > Bluetooth MAC addresses are random, they aren't useful identifiers. If > someone can see you, a face is a far more uniquely identifying thing than a > MAC. > > "Payment spam" might be a problem. I can imagine a wallet requiring that > such requests are signed and then spammers can be blacklisted in the usual > fashion so they can't push things to your phone anymore. Anyway, a hurdle > that can be jumped if/when it becomes an issue. > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is > your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development > >
Hi Paul,
The issue is in the establishment of trust. Anyone can broadcast the initial information.
e
Quoted text (14 lines)
> On Feb 5, 2015, at 2:01 PM, Paul Puey <paul@airbitz•co> wrote: > > The broadcast is ONLY done when the wallet is in Receive mode. Same as when the QR code is visible. The use of the *Name* section is specifically so that a recipient can broadcast their name/handle. Not so the recipient would broadcast the name of the Sender. > > On Thu, Feb 5, 2015 at 12:50 PM, Mike Hearn <mike@plan99•net> wrote: >>> I'm imagining myself walking around broadcasting my photo and MAC >>> address while hucksters push payment requests to me for approval >> >> I hate to break it to you, but you broadcast a photo of your face every time you walk outside ;) >> >> Bluetooth MAC addresses are random, they aren't useful identifiers. If someone can see you, a face is a far more uniquely identifying thing than a MAC. >> >> "Payment spam" might be a problem. I can imagine a wallet requiring that such requests are signed and then spammers can be blacklisted in the usual fashion so they can't push things to your phone anymore. Anyway, a hurdle that can be jumped if/when it becomes an issue. >
So if you picked up the BLE broadcast request. All you know is that *someone* within 100m is requesting bitcoin at a certain address. Not necessarily who. The *name* is both optional, and possibly just a *handle* of the user. If I'm sitting 5 ft away from someone at dinner and wanted to pay them via BLE, I might see "Monkey Dude" on my list and simply ask him "is that you?" If so, I send it. If there are two "Monkey Dude's" Then I have to bother with the address prefix, but not otherwise.
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
On Thu, Feb 5, 2015 at 1:46 PM, Eric Voskuil <eric@voskuil•org> wrote:
Quoted text (42 lines)
> BLE has an advertised range of over 100m. > > http://www.bluetooth.com/Pages/low-energy-tech-info.aspx > > In the case of mass surveillance that range could most likely be extended > dramatically by the reviewer. I've seen WiFi ranges of over a mile with a > strong (not FCC approved) receiver. > > WiFi hotspots don't have strong identity or a guaranteed position, so they > can't be trusted for location. > > e > > On Feb 5, 2015, at 1:36 PM, Mike Hearn <mike@plan99•net> wrote: > > This sounds horrible. You could basically monitor anyone with a wallet in >> a highly populated area and track them super easily by doing facial >> recognition. >> > > We're talking about BLE, still? The radio tech that runs in the so called > "junk bands" because propagation is so poor? > > My watch loses its connection to my phone if I just put it down and walk > around my apartment. I'm all for reasonable paranoia, but Bluetooth isn't > going to be enabling mass surveillance any time soon. It barely goes > through air, let alone walls. > > Anyway, whatever. I'm just bouncing around ideas for faster user > interfaces. You could always switch it off or set it to be triggered by the > presence of particular wifi hotspots, if you don't mind an initial bit of > setup. > > Back on topic - the debate is interesting, but I think to get this to the > stage of being a BIP we'd need at least another wallet to implement it? > Then I guess a BIP would be useful regardless of the design issues. The > prefix matching still feels flaky to me but it's hard to know if you could > really swipe payments out of the air in practice, without actually trying > it. > > >
Although not perfect, and it may require visual/verbal verification, I don't see what the trust issue is.
[image: logo] *Paul Puey* CEO / Co-Founder, Airbitz Inc +1-619-850-8624 | http://airbitz.co | San Diego <http://facebook.com/airbitz> <http://twitter.com/airbitz> <https://plus.google.com/118173667510609425617> <https://go.airbitz.co/comments/feed/> <http://linkedin.com/in/paulpuey> <https://angel.co/paul-puey> *DOWNLOAD THE AIRBITZ WALLET:* <https://play.google.com/store/apps/details?id=com.airbitz> <https://itunes.apple.com/us/app/airbitz/id843536046>
On Thu, Feb 5, 2015 at 2:05 PM, Eric Voskuil <eric@voskuil•org> wrote:
Quoted text (34 lines)
> Hi Paul, > > The issue is in the establishment of trust. Anyone can broadcast the > initial information. > > e > > On Feb 5, 2015, at 2:01 PM, Paul Puey <paul@airbitz•co> wrote: > > The broadcast is ONLY done when the wallet is in Receive mode. Same as > when the QR code is visible. The use of the *Name* section is specifically > so that a recipient can broadcast their name/handle. Not so the recipient > would broadcast the name of the Sender. > > On Thu, Feb 5, 2015 at 12:50 PM, Mike Hearn <mike@plan99•net> wrote: > >> I'm imagining myself walking around broadcasting my photo and MAC >>> address while hucksters push payment requests to me for approval >> >> >> I hate to break it to you, but you broadcast a photo of your face every >> time you walk outside ;) >> >> Bluetooth MAC addresses are random, they aren't useful identifiers. If >> someone can see you, a face is a far more uniquely identifying thing than a >> MAC. >> >> "Payment spam" might be a problem. I can imagine a wallet requiring that >> such requests are signed and then spammers can be blacklisted in the usual >> fashion so they can't push things to your phone anymore. Anyway, a hurdle >> that can be jumped if/when it becomes an issue. >> > >
A MITM can receive the initial broadcast and then spoof it by jamming the original. You then only see one.
e
Quoted text (27 lines)
> On Feb 5, 2015, at 2:07 PM, Paul Puey <paul@airbitz•co> wrote: > > So if you picked up the BLE broadcast request. All you know is that *someone* within 100m is requesting bitcoin at a certain address. Not necessarily who. The *name* is both optional, and possibly just a *handle* of the user. If I'm sitting 5 ft away from someone at dinner and wanted to pay them via BLE, I might see "Monkey Dude" on my list and simply ask him "is that you?" If so, I send it. If there are two "Monkey Dude's" Then I have to bother with the address prefix, but not otherwise. > >> On Thu, Feb 5, 2015 at 1:46 PM, Eric Voskuil <eric@voskuil•org> wrote: >> BLE has an advertised range of over 100m. >> >> http://www.bluetooth.com/Pages/low-energy-tech-info.aspx >> >> In the case of mass surveillance that range could most likely be extended dramatically by the reviewer. I've seen WiFi ranges of over a mile with a strong (not FCC approved) receiver. >> >> WiFi hotspots don't have strong identity or a guaranteed position, so they can't be trusted for location. >> >> e >> >> On Feb 5, 2015, at 1:36 PM, Mike Hearn <mike@plan99•net> wrote: >> >>>> This sounds horrible. You could basically monitor anyone with a wallet in a highly populated area and track them super easily by doing facial recognition. >>> >>> We're talking about BLE, still? The radio tech that runs in the so called "junk bands" because propagation is so poor? >>> >>> My watch loses its connection to my phone if I just put it down and walk around my apartment. I'm all for reasonable paranoia, but Bluetooth isn't going to be enabling mass surveillance any time soon. It barely goes through air, let alone walls. >>> >>> Anyway, whatever. I'm just bouncing around ideas for faster user interfaces. You could always switch it off or set it to be triggered by the presence of particular wifi hotspots, if you don't mind an initial bit of setup. >>> >>> Back on topic - the debate is interesting, but I think to get this to the stage of being a BIP we'd need at least another wallet to implement it? Then I guess a BIP would be useful regardless of the design issues. The prefix matching still feels flaky to me but it's hard to know if you could really swipe payments out of the air in practice, without actually trying it. >
Personally I like the simplicity of tapping two phones together to make payment - it should be quicker and easier than scanning QR codes and it's a trust model that's hard to misunderstand.
Is NFC good enough for that? I fear even with NFC it is possible to produce a device with longer range than one would expect. What happened to the idea of tapping two devices together and then comparing the timing of the tap (as detected by the phones' accelerometers) to make spoofing a transaction harder? I remember hearing about that years ago - is that still a thing?
roy
On Thu, Feb 05, 2015 at 02:10:51PM -0800, Eric Voskuil wrote:
Quoted text (41 lines)
> A MITM can receive the initial broadcast and then spoof it by jamming the original. You then only see one. > > e > > > On Feb 5, 2015, at 2:07 PM, Paul Puey <paul@airbitz•co> wrote: > > > > So if you picked up the BLE broadcast request. All you know is that *someone* within 100m is requesting bitcoin at a certain address. Not necessarily who. The *name* is both optional, and possibly just a *handle* of the user. If I'm sitting 5 ft away from someone at dinner and wanted to pay them via BLE, I might see "Monkey Dude" on my list and simply ask him "is that you?" If so, I send it. If there are two "Monkey Dude's" Then I have to bother with the address prefix, but not otherwise. > > > >> On Thu, Feb 5, 2015 at 1:46 PM, Eric Voskuil <eric@voskuil•org> wrote: > >> BLE has an advertised range of over 100m. > >> > >> http://www.bluetooth.com/Pages/low-energy-tech-info.aspx > >> > >> In the case of mass surveillance that range could most likely be extended dramatically by the reviewer. I've seen WiFi ranges of over a mile with a strong (not FCC approved) receiver. > >> > >> WiFi hotspots don't have strong identity or a guaranteed position, so they can't be trusted for location. > >> > >> e > >> > >> On Feb 5, 2015, at 1:36 PM, Mike Hearn <mike@plan99•net> wrote: > >> > >>>> This sounds horrible. You could basically monitor anyone with a wallet in a highly populated area and track them super easily by doing facial recognition. > >>> > >>> We're talking about BLE, still? The radio tech that runs in the so called "junk bands" because propagation is so poor? > >>> > >>> My watch loses its connection to my phone if I just put it down and walk around my apartment. I'm all for reasonable paranoia, but Bluetooth isn't going to be enabling mass surveillance any time soon. It barely goes through air, let alone walls. > >>> > >>> Anyway, whatever. I'm just bouncing around ideas for faster user interfaces. You could always switch it off or set it to be triggered by the presence of particular wifi hotspots, if you don't mind an initial bit of setup. > >>> > >>> Back on topic - the debate is interesting, but I think to get this to the stage of being a BIP we'd need at least another wallet to implement it? Then I guess a BIP would be useful regardless of the design issues. The prefix matching still feels flaky to me but it's hard to know if you could really swipe payments out of the air in practice, without actually trying it. > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
On Thu, Feb 5, 2015 at 2:10 PM, Eric Voskuil <eric@voskuil•org> wrote:
Quoted text (2 lines)
> A MITM can receive the initial broadcast and then spoof it by jamming the > original. You then only see one.
You are right, of course. There is no way to make Bluetooth 100% secure, since it is an over-the-air technology. You could try securing it using a CA or other identity server, but now you've excluded ad-hoc person-to-person payments. Plus, you need an active internet connection to reach the CA.
You can try using proximity as a substitute for identity, like requiring NFC to kick-start the connection, but at that point you might as well use QR codes.
This BIP is not trying to provide absolute bullet-proof security, since that's impossible given the physical limitations of the Bluetooth technology. Instead, it's trying to provide the best-possible security given those constraints. In exchange for this, we get greatly enhanced usability in common scenarios.
There are plenty of usable, real-world technologies with big security holes. Anybody with lock-picking experience will tell you this, but nobody is welding their front door shut. The ability to go in and out is worth the security risk.
Bluetooth payments add a whole new dimension to real-world Bitcoin usability. Do we shut that down because it can't be made perfect, or do we do the best we can and move forward?
-William
I would like to shortly express my opinion:
- Having BT as an alternative is good idea but it must be secure enough - Signed BIP70 should be enough. I see only two issues regarding BIP70 (but they apply also to TCP/IP, not just BT): key revocations and MITM attacks by governments. - Broadcasting faces is very bad idea IMHO. - Comparing addresses seems complicated but if hash was displayed as a unique, picture hard to be mistake or long phrase, it could be more convenient. - Maybe storing public key (I do NOT mean bitcoin address!) of merchant after successful transaction is good compromise?
Another idea: I noticed it's extremely easy to compare two strings if they are the same size (in terms of millimeters, not number of characters). If the hash of signing key was printed on a sign near the POS in specified size (90% of smallest available screen?) and phone would scale correctly, just putting the phone near the sign would be enough to instantly spot whether the hashes are same.
Maybe instead of hex/base58 hash encoding use colored barcode. But I'm not sure if it would improve things.
2015-02-05 23:49 GMT+01:00 Roy Badami <roy@gnomon•org•uk>:
Quoted text (69 lines)
> Personally I like the simplicity of tapping two phones together to > make payment - it should be quicker and easier than scanning QR codes > and it's a trust model that's hard to misunderstand. > > Is NFC good enough for that? I fear even with NFC it is possible to > produce a device with longer range than one would expect. What > happened to the idea of tapping two devices together and then > comparing the timing of the tap (as detected by the phones' > accelerometers) to make spoofing a transaction harder? I remember > hearing about that years ago - is that still a thing? > > roy > > On Thu, Feb 05, 2015 at 02:10:51PM -0800, Eric Voskuil wrote: >> A MITM can receive the initial broadcast and then spoof it by jamming the original. You then only see one. >> >> e >> >> > On Feb 5, 2015, at 2:07 PM, Paul Puey <paul@airbitz•co> wrote: >> > >> > So if you picked up the BLE broadcast request. All you know is that *someone* within 100m is requesting bitcoin at a certain address. Not necessarily who. The *name* is both optional, and possibly just a *handle* of the user. If I'm sitting 5 ft away from someone at dinner and wanted to pay them via BLE, I might see "Monkey Dude" on my list and simply ask him "is that you?" If so, I send it. If there are two "Monkey Dude's" Then I have to bother with the address prefix, but not otherwise. >> > >> >> On Thu, Feb 5, 2015 at 1:46 PM, Eric Voskuil <eric@voskuil•org> wrote: >> >> BLE has an advertised range of over 100m. >> >> >> >> http://www.bluetooth.com/Pages/low-energy-tech-info.aspx >> >> >> >> In the case of mass surveillance that range could most likely be extended dramatically by the reviewer. I've seen WiFi ranges of over a mile with a strong (not FCC approved) receiver. >> >> >> >> WiFi hotspots don't have strong identity or a guaranteed position, so they can't be trusted for location. >> >> >> >> e >> >> >> >> On Feb 5, 2015, at 1:36 PM, Mike Hearn <mike@plan99•net> wrote: >> >> >> >>>> This sounds horrible. You could basically monitor anyone with a wallet in a highly populated area and track them super easily by doing facial recognition. >> >>> >> >>> We're talking about BLE, still? The radio tech that runs in the so called "junk bands" because propagation is so poor? >> >>> >> >>> My watch loses its connection to my phone if I just put it down and walk around my apartment. I'm all for reasonable paranoia, but Bluetooth isn't going to be enabling mass surveillance any time soon. It barely goes through air, let alone walls. >> >>> >> >>> Anyway, whatever. I'm just bouncing around ideas for faster user interfaces. You could always switch it off or set it to be triggered by the presence of particular wifi hotspots, if you don't mind an initial bit of setup. >> >>> >> >>> Back on topic - the debate is interesting, but I think to get this to the stage of being a BIP we'd need at least another wallet to implement it? Then I guess a BIP would be useful regardless of the design issues. The prefix matching still feels flaky to me but it's hard to know if you could really swipe payments out of the air in practice, without actually trying it. >> > > >> ------------------------------------------------------------------------------ >> Dive into the World of Parallel Programming. The Go Parallel Website, >> sponsored by Intel and developed in partnership with Slashdot Media, is your >> hub for all things parallel software development, from weekly thought >> leadership blogs to news, videos, case studies, tutorials and more. Take a >> look and join the conversation now. http://goparallel.sourceforge.net/ > >> _______________________________________________ >> Bitcoin-development mailing list >> Bitcoin-development@lists•sourceforge•net >> https://lists.sourceforge.net/lists/listinfo/bitcoin-development > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development
On 02/05/2015 02:08 PM, Paul Puey wrote:
Quoted text (2 lines)
> Although not perfect, and it may require visual/verbal verification, > I don't see what the trust issue is.
I agree that with manual verification between the parties the worst problem becomes DOS, which is certainly not catastrophic.
But the objective is to the extent possible improve upon the cumbersome process of QR code, NFC signal, or textual address scanning. Given that there would be no way to know you are under attack, with the exception of manual confirmation, it would seem unwise to ever rely on the automation. If the automation cannot be relied upon, it may actually make matters worse. People would either take their chances by relying on it or go through a more complex process.
In terms of the difficulty of an attack, it's important to recognize that all attacks (DOS, privacy, integrity) in this scenario can be fully-automated and executed over the air by a black box at some distance:
* DOS is possible by rebroadcasting a similar request.
* Privacy is compromised by monitoring for payment requests and correlating them to location and potentially images of parties.
* Integrity is compromised by either: (1) Rebroadcasting a similar transaction with a bogus address but with the same leading characters; can't be spent but you lose your money. (2) Rebroadcasting with a valid address that doesn't match the leading characters, in the expectation that the user doesn't check manually.
Regarding possible mitigation via BIP-70:
A BIP-70 signed payment request in the initial broadcast can resolve the integrity issues, but because of the public nature of the broadcast coupled with strong public identity, the privacy compromise is much worse. Now transactions are cryptographically tainted.
This is also the problem with BIP-70 over the web. TLS and other security precautions aside, an interloper on the communication, desktop, datacenter, etc., can capture payment requests and strongly correlate transactions to identities in an automated manner. The payment request must be kept private between the parties, and that's hard to do.
So the initial broadcast needs privacy, but then of course it cannot be a broadcast - it need to be a narrow cast. That brings us back to proximity-based establishment.
I think that you could get away with this for a while, simply because of the narrow fields we are working in presently. But in a bitcoin world it would be very problematic. For this reason I wouldn't want to encourage standardization on this approach.
e
On 02/05/2015 02:10 PM, Eric Voskuil wrote:
Quoted text (63 lines)
> A MITM can receive the initial broadcast and then spoof it by jamming > the original. You then only see one. > > e > > On Feb 5, 2015, at 2:07 PM, Paul Puey <paul@airbitz•co > <mailto:paul@airbitz.co>> wrote: > >> So if you picked up the BLE broadcast request. All you know is that >> *someone* within 100m is requesting bitcoin at a certain address. Not >> necessarily who. The *name* is both optional, and possibly just a >> *handle* of the user. If I'm sitting 5 ft away from someone at dinner >> and wanted to pay them via BLE, I might see "Monkey Dude" on my list >> and simply ask him "is that you?" If so, I send it. If there are two >> "Monkey Dude's" Then I have to bother with the address prefix, but not >> otherwise. >> >> On Thu, Feb 5, 2015 at 1:46 PM, Eric Voskuil <eric@voskuil•org >> <mailto:eric@voskuil.org>> wrote: >> >> BLE has an advertised range of over 100m. >> >> http://www.bluetooth.com/Pages/low-energy-tech-info.aspx >> >> In the case of mass surveillance that range could most likely be >> extended dramatically by the reviewer. I've seen WiFi ranges of >> over a mile with a strong (not FCC approved) receiver. >> >> WiFi hotspots don't have strong identity or a guaranteed position, >> so they can't be trusted for location. >> >> e >> >> On Feb 5, 2015, at 1:36 PM, Mike Hearn <mike@plan99•net >> <mailto:mike@plan99.net>> wrote: >> >>> This sounds horrible. You could basically monitor anyone with >>> a wallet in a highly populated area and track them super >>> easily by doing facial recognition. >>> >>> >>> We're talking about BLE, still? The radio tech that runs in the >>> so called "junk bands" because propagation is so poor? >>> >>> My watch loses its connection to my phone if I just put it down >>> and walk around my apartment. I'm all for reasonable paranoia, >>> but Bluetooth isn't going to be enabling mass surveillance any >>> time soon. It barely goes through air, let alone walls. >>> >>> Anyway, whatever. I'm just bouncing around ideas for faster user >>> interfaces. You could always switch it off or set it to be >>> triggered by the presence of particular wifi hotspots, if you >>> don't mind an initial bit of setup. >>> >>> Back on topic - the debate is interesting, but I think to get >>> this to the stage of being a BIP we'd need at least another >>> wallet to implement it? Then I guess a BIP would be useful >>> regardless of the design issues. The prefix matching still feels >>> flaky to me but it's hard to know if you could really swipe >>> payments out of the air in practice, without actually trying it. >>> >>> >>
For peer-to-peer payments, how common do we think that the payment is of an ad hoc nature rather than to a known contact?
If I want to pay my friends/colleagues/etc over a restaurant table there's no reason why I couldn't already have their public keys in my contact list - then it would be pretty straightforward to have a watertight mechanism where I would know who I was paying. You could probably even relatively securely bootstrap a key exchange over SMS, relying only on the contacts already having each other in their phonebooks.
As for comsumer-to-merchant transactions where the merchant is a bricks and mortar merchant, IMHO it absolutely has to be "pay that terminal over there". It's the trust model we all currently use - whether paying cash or card - and it's the only trust model that works IMHO (and customers and businesses alike are well aware of the risks of a fraudster standing behind the counter pretending to be an employee accepting payment - and by and large are pretty good at mitigating it). OTOH as we've discussed here before there are many use cases where the custoemr doesn't actually know or care about the name of the shop or bar they walked into but is pretty damn sure that they need to make payment to the person over there behind the counter.
Granted, there are cases taht dont' fall into either of the above - but they're the cases that are (a) harder to figure out how to authenticate and consequently (b) the use cases that are going to be most subject to attempted fraud.
roy
On Thu, Feb 05, 2015 at 03:02:56PM -0800, William Swanson wrote:
Quoted text (42 lines)
> On Thu, Feb 5, 2015 at 2:10 PM, Eric Voskuil <eric@voskuil•org> wrote: > > A MITM can receive the initial broadcast and then spoof it by jamming the > > original. You then only see one. > > You are right, of course. There is no way to make Bluetooth 100% > secure, since it is an over-the-air technology. You could try securing > it using a CA or other identity server, but now you've excluded ad-hoc > person-to-person payments. Plus, you need an active internet > connection to reach the CA. > > You can try using proximity as a substitute for identity, like > requiring NFC to kick-start the connection, but at that point you > might as well use QR codes. > > This BIP is not trying to provide absolute bullet-proof security, > since that's impossible given the physical limitations of the > Bluetooth technology. Instead, it's trying to provide the > best-possible security given those constraints. In exchange for this, > we get greatly enhanced usability in common scenarios. > > There are plenty of usable, real-world technologies with big security > holes. Anybody with lock-picking experience will tell you this, but > nobody is welding their front door shut. The ability to go in and out > is worth the security risk. > > Bluetooth payments add a whole new dimension to real-world Bitcoin > usability. Do we shut that down because it can't be made perfect, or > do we do the best we can and move forward? > > -William > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development >
Quoted text (10 lines)
> A BIP-70 signed payment request in the initial broadcast can resolve the > integrity issues, but because of the public nature of the broadcast > coupled with strong public identity, the privacy compromise is much > worse. Now transactions are cryptographically tainted. > > This is also the problem with BIP-70 over the web. TLS and other > security precautions aside, an interloper on the communication, desktop, > datacenter, etc., can capture payment requests and strongly correlate > transactions to identities in an automated manner. The payment request > must be kept private between the parties, and that's hard to do.
What about using encryption with forward secrecy? Merchant would generate signed request containing public ECDH part, buyer would send back transaction encrypted with ECDH and his public ECDH part. If receiving address/amount is meant to be private, use commit protocol (see ZRTP/RedPhone) and short authentication phrase (which is hard to spoof thanks to commit protocol - see RedPhone)?
On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote:
Quoted text (17 lines)
>> A BIP-70 signed payment request in the initial broadcast can resolve the >> integrity issues, but because of the public nature of the broadcast >> coupled with strong public identity, the privacy compromise is much >> worse. Now transactions are cryptographically tainted. >> >> This is also the problem with BIP-70 over the web. TLS and other >> security precautions aside, an interloper on the communication, desktop, >> datacenter, etc., can capture payment requests and strongly correlate >> transactions to identities in an automated manner. The payment request >> must be kept private between the parties, and that's hard to do. > > What about using encryption with forward secrecy? Merchant would > generate signed request containing public ECDH part, buyer would send > back transaction encrypted with ECDH and his public ECDH part. If > receiving address/amount is meant to be private, use commit protocol > (see ZRTP/RedPhone) and short authentication phrase (which is hard to > spoof thanks to commit protocol - see RedPhone)?
Hi Martin,
The problem is that you need to verify the ownership of the public key. A MITM can substitute the key. If you don't have verifiable identity associated with the public key (PKI/WoT), you need a shared secret (such as a secret phrase). But the problem is then establishing that secret over a public channel.
You can bootstrap a private session over the untrusted network using a trusted public key (PKI/WoT). But the presumption is that you are already doing this over the web (using TLS). That process is subject to attack at the CA. WoT is not subject to a CA attack, because it's decentralized. But it's also not sufficiently deployed for some scenarios.
e
On 02/05/2015 03:34 PM, Roy Badami wrote:
Quoted text (10 lines)
> For peer-to-peer payments, how common do we think that the payment is > of an ad hoc nature rather than to a known contact? > > If I want to pay my friends/colleagues/etc over a restaurant table > there's no reason why I couldn't already have their public keys in my > contact list - then it would be pretty straightforward to have a > watertight mechanism where I would know who I was paying. You could > probably even relatively securely bootstrap a key exchange over SMS, > relying only on the contacts already having each other in their > phonebooks.
In this case there is no need for P2P communication, just pay to an address you already have for the other party. If you want to avoid address reuse, use stealth addressing.
But yes, if you don't have a stealth address for the other party you can certainly communicate in private as peers where you trust that you share a public key. The core issue here is really bootstrapping of that trust in an ad hoc manner.
Quoted text (11 lines)
> As for comsumer-to-merchant transactions where the merchant is a > bricks and mortar merchant, IMHO it absolutely has to be "pay that > terminal over there". It's the trust model we all currently use - > whether paying cash or card - and it's the only trust model that works > IMHO (and customers and businesses alike are well aware of the risks > of a fraudster standing behind the counter pretending to be an > employee accepting payment - and by and large are pretty good at > mitigating it). OTOH as we've discussed here before there are many > use cases where the custoemr doesn't actually know or care about the > name of the shop or bar they walked into but is pretty damn sure that > they need to make payment to the person over there behind the counter.
Yes, proximity is practically the universal solution to the problem of the payer identifying the correct seller in any face-to-face scenario.
Quoted text (4 lines)
> Granted, there are cases taht dont' fall into either of the above - > but they're the cases that are (a) harder to figure out how to > authenticate and consequently (b) the use cases that are going to be > most subject to attempted fraud.
When identification is required (show me some id before I pay you) it equates to the BIP-70 scenario in the bitcoin model. This is also required in order guard against repudiation (give me a signed receipt).
Quoted text (43 lines)
> On Thu, Feb 05, 2015 at 03:02:56PM -0800, William Swanson wrote: >> On Thu, Feb 5, 2015 at 2:10 PM, Eric Voskuil <eric@voskuil•org> wrote: >>> A MITM can receive the initial broadcast and then spoof it by jamming the >>> original. You then only see one. >> >> You are right, of course. There is no way to make Bluetooth 100% >> secure, since it is an over-the-air technology. You could try securing >> it using a CA or other identity server, but now you've excluded ad-hoc >> person-to-person payments. Plus, you need an active internet >> connection to reach the CA. >> >> You can try using proximity as a substitute for identity, like >> requiring NFC to kick-start the connection, but at that point you >> might as well use QR codes. >> >> This BIP is not trying to provide absolute bullet-proof security, >> since that's impossible given the physical limitations of the >> Bluetooth technology. Instead, it's trying to provide the >> best-possible security given those constraints. In exchange for this, >> we get greatly enhanced usability in common scenarios. >> >> There are plenty of usable, real-world technologies with big security >> holes. Anybody with lock-picking experience will tell you this, but >> nobody is welding their front door shut. The ability to go in and out >> is worth the security risk. >> >> Bluetooth payments add a whole new dimension to real-world Bitcoin >> usability. Do we shut that down because it can't be made perfect, or >> do we do the best we can and move forward? >> >> -William >> >> ------------------------------------------------------------------------------ >> Dive into the World of Parallel Programming. The Go Parallel Website, >> sponsored by Intel and developed in partnership with Slashdot Media, is your >> hub for all things parallel software development, from weekly thought >> leadership blogs to news, videos, case studies, tutorials and more. Take a >> look and join the conversation now. http://goparallel.sourceforge.net/ >> _______________________________________________ >> Bitcoin-development mailing list >> Bitcoin-development@lists•sourceforge•net >> https://lists.sourceforge.net/lists/listinfo/bitcoin-development >>
That's exactly what I though when seeing the RedPhone code, but after I studied the commit protocol I realized it's actually secure and convenient way to do it. You should do that too. :)
Shortly, how it works: The initiator of the connection sends commit message containing the hash of his temporary public ECDH part, second party sends back their public ECDH part and then initiator sends his public ECDH part in open. All three messages are hashed together and the first two bytes are used to select two words from a shared dictionary which are displayed on the screen of both the initiator and the second party. The parties communicate those two words and verify they match.
If an attacker wants to do MITM, he has a chance of choosing right public parts 1:65536. There is no way to brute-force it, since that would be noticed immediately. If instead of two words based on the first two bytes, four words from BIP39 wordlist were chosen, it would provide entropy of 44 bits which I believe should be enough even for paranoid people.
How this would work in Bitcoin payment scenario: user's phone broadcasts his name, merchant inputs amount and selects the name from the list, commit message is sent (and then the remaining two messages), merchant spells four words he sees on the screen and buyer confirms transaction after verifying that words match.
2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>:
Quoted text (35 lines)
> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>> A BIP-70 signed payment request in the initial broadcast can resolve the >>> integrity issues, but because of the public nature of the broadcast >>> coupled with strong public identity, the privacy compromise is much >>> worse. Now transactions are cryptographically tainted. >>> >>> This is also the problem with BIP-70 over the web. TLS and other >>> security precautions aside, an interloper on the communication, desktop, >>> datacenter, etc., can capture payment requests and strongly correlate >>> transactions to identities in an automated manner. The payment request >>> must be kept private between the parties, and that's hard to do. >> >> What about using encryption with forward secrecy? Merchant would >> generate signed request containing public ECDH part, buyer would send >> back transaction encrypted with ECDH and his public ECDH part. If >> receiving address/amount is meant to be private, use commit protocol >> (see ZRTP/RedPhone) and short authentication phrase (which is hard to >> spoof thanks to commit protocol - see RedPhone)? > > Hi Martin, > > The problem is that you need to verify the ownership of the public key. > A MITM can substitute the key. If you don't have verifiable identity > associated with the public key (PKI/WoT), you need a shared secret (such > as a secret phrase). But the problem is then establishing that secret > over a public channel. > > You can bootstrap a private session over the untrusted network using a > trusted public key (PKI/WoT). But the presumption is that you are > already doing this over the web (using TLS). That process is subject to > attack at the CA. WoT is not subject to a CA attack, because it's > decentralized. But it's also not sufficiently deployed for some scenarios. > > e >
On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote:
Quoted text (3 lines)
> That's exactly what I though when seeing the RedPhone code, but after > I studied the commit protocol I realized it's actually secure and > convenient way to do it. You should do that too. :)
I was analyzing the model as you described it to me. A formal analysis of the security model of a particular implementation, based on inference from source code, is a bit beyond what I signed up for. But I'm perfectly willing to comment on your description of the model if you are willing to indulge me.
Quoted text (8 lines)
> Shortly, how it works: > The initiator of the connection sends commit message containing the > hash of his temporary public ECDH part, second party sends back their > public ECDH part and then initiator sends his public ECDH part in > open. All three messages are hashed together and the first two bytes > are used to select two words from a shared dictionary which are > displayed on the screen of both the initiator and the second party. > The parties communicate those two words and verify they match.
How do they compare words if they haven't yet established a secure channel?
Quoted text (12 lines)
> If an attacker wants to do MITM, he has a chance of choosing right > public parts 1:65536. There is no way to brute-force it, since that > would be noticed immediately. If instead of two words based on the > first two bytes, four words from BIP39 wordlist were chosen, it would > provide entropy of 44 bits which I believe should be enough even for > paranoid people. > > How this would work in Bitcoin payment scenario: user's phone > broadcasts his name, merchant inputs amount and selects the name from > the list, commit message is sent (and then the remaining two > messages), merchant spells four words he sees on the screen and buyer > confirms transaction after verifying that words match.
So the assumption is that there exists a secure (as in proximity-based) communication channel?
e
Quoted text (36 lines)
> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>> A BIP-70 signed payment request in the initial broadcast can resolve the >>>> integrity issues, but because of the public nature of the broadcast >>>> coupled with strong public identity, the privacy compromise is much >>>> worse. Now transactions are cryptographically tainted. >>>> >>>> This is also the problem with BIP-70 over the web. TLS and other >>>> security precautions aside, an interloper on the communication, desktop, >>>> datacenter, etc., can capture payment requests and strongly correlate >>>> transactions to identities in an automated manner. The payment request >>>> must be kept private between the parties, and that's hard to do. >>> >>> What about using encryption with forward secrecy? Merchant would >>> generate signed request containing public ECDH part, buyer would send >>> back transaction encrypted with ECDH and his public ECDH part. If >>> receiving address/amount is meant to be private, use commit protocol >>> (see ZRTP/RedPhone) and short authentication phrase (which is hard to >>> spoof thanks to commit protocol - see RedPhone)? >> >> Hi Martin, >> >> The problem is that you need to verify the ownership of the public key. >> A MITM can substitute the key. If you don't have verifiable identity >> associated with the public key (PKI/WoT), you need a shared secret (such >> as a secret phrase). But the problem is then establishing that secret >> over a public channel. >> >> You can bootstrap a private session over the untrusted network using a >> trusted public key (PKI/WoT). But the presumption is that you are >> already doing this over the web (using TLS). That process is subject to >> attack at the CA. WoT is not subject to a CA attack, because it's >> decentralized. But it's also not sufficiently deployed for some scenarios. >> >> e >>
I believe, we are still talking about transactions of physical people in physical world. So yes, it's proximity based - people tell the words by mouth. :)
In case of RedPhone, you read those words verbally over not-yet-verified channel relying on difficulty of spoofing your voice. Also the app remembers the public keys, so you don't need to verify second time.
I suggest you to try RedPhone (called Signal on iPhone) yourself. It's free/open source, Internet-based and end-to-end encrypted. You may find it useful some day. Also I'm willing to help you with trying it after I wake up. (~8 hours: Send me private e-mail if you want to.)
Dňa 6. februára 2015 1:22:23 CET používateľ Eric Voskuil <eric@voskuil•org> napísal:
Quoted text (96 lines)
> >On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote: >> That's exactly what I though when seeing the RedPhone code, but after >> I studied the commit protocol I realized it's actually secure and >> convenient way to do it. You should do that too. :) > >I was analyzing the model as you described it to me. A formal analysis >of the security model of a particular implementation, based on >inference >from source code, is a bit beyond what I signed up for. But I'm >perfectly willing to comment on your description of the model if you >are >willing to indulge me. > >> Shortly, how it works: >> The initiator of the connection sends commit message containing the >> hash of his temporary public ECDH part, second party sends back their >> public ECDH part and then initiator sends his public ECDH part in >> open. All three messages are hashed together and the first two bytes >> are used to select two words from a shared dictionary which are >> displayed on the screen of both the initiator and the second party. > >> The parties communicate those two words and verify they match. > >How do they compare words if they haven't yet established a secure >channel? > >> If an attacker wants to do MITM, he has a chance of choosing right >> public parts 1:65536. There is no way to brute-force it, since that >> would be noticed immediately. If instead of two words based on the >> first two bytes, four words from BIP39 wordlist were chosen, it would >> provide entropy of 44 bits which I believe should be enough even for >> paranoid people. >> >> How this would work in Bitcoin payment scenario: user's phone >> broadcasts his name, merchant inputs amount and selects the name from >> the list, commit message is sent (and then the remaining two >> messages), merchant spells four words he sees on the screen and buyer >> confirms transaction after verifying that words match. > >So the assumption is that there exists a secure (as in proximity-based) >communication channel? > >e > >> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>> A BIP-70 signed payment request in the initial broadcast can >resolve the >>>>> integrity issues, but because of the public nature of the >broadcast >>>>> coupled with strong public identity, the privacy compromise is >much >>>>> worse. Now transactions are cryptographically tainted. >>>>> >>>>> This is also the problem with BIP-70 over the web. TLS and other >>>>> security precautions aside, an interloper on the communication, >desktop, >>>>> datacenter, etc., can capture payment requests and strongly >correlate >>>>> transactions to identities in an automated manner. The payment >request >>>>> must be kept private between the parties, and that's hard to do. >>>> >>>> What about using encryption with forward secrecy? Merchant would >>>> generate signed request containing public ECDH part, buyer would >send >>>> back transaction encrypted with ECDH and his public ECDH part. If >>>> receiving address/amount is meant to be private, use commit >protocol >>>> (see ZRTP/RedPhone) and short authentication phrase (which is hard >to >>>> spoof thanks to commit protocol - see RedPhone)? >>> >>> Hi Martin, >>> >>> The problem is that you need to verify the ownership of the public >key. >>> A MITM can substitute the key. If you don't have verifiable identity >>> associated with the public key (PKI/WoT), you need a shared secret >(such >>> as a secret phrase). But the problem is then establishing that >secret >>> over a public channel. >>> >>> You can bootstrap a private session over the untrusted network using >a >>> trusted public key (PKI/WoT). But the presumption is that you are >>> already doing this over the web (using TLS). That process is subject >to >>> attack at the CA. WoT is not subject to a CA attack, because it's >>> decentralized. But it's also not sufficiently deployed for some >scenarios. >>> >>> e >>>
- -- Odoslané z môjho Android zariadenia pomocou K-9 Mail.
The trust can be considered bootstrapped by visual verification of the address prefix. If we are really concerned about someone jamming a Bluetooth signal in a coffeeshop then the UI can encourage verification of the prefix. Much like how regular Bluetooth requires 'pairing' via entering a 4-6 digit code.
Paul Puey CEO / Co-Founder, Airbitz Inc 619.850.8624 | http://airbitz.co | San Diego
On Feb 5, 2015, at 3:46 PM, Eric Voskuil <eric@voskuil•org> wrote:
On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote:
Quoted text (17 lines)
>> A BIP-70 signed payment request in the initial broadcast can resolve the >> integrity issues, but because of the public nature of the broadcast >> coupled with strong public identity, the privacy compromise is much >> worse. Now transactions are cryptographically tainted. >> >> This is also the problem with BIP-70 over the web. TLS and other >> security precautions aside, an interloper on the communication, desktop, >> datacenter, etc., can capture payment requests and strongly correlate >> transactions to identities in an automated manner. The payment request >> must be kept private between the parties, and that's hard to do. > > What about using encryption with forward secrecy? Merchant would > generate signed request containing public ECDH part, buyer would send > back transaction encrypted with ECDH and his public ECDH part. If > receiving address/amount is meant to be private, use commit protocol > (see ZRTP/RedPhone) and short authentication phrase (which is hard to > spoof thanks to commit protocol - see RedPhone)?
Hi Martin,
The problem is that you need to verify the ownership of the public key. A MITM can substitute the key. If you don't have verifiable identity associated with the public key (PKI/WoT), you need a shared secret (such as a secret phrase). But the problem is then establishing that secret over a public channel.
You can bootstrap a private session over the untrusted network using a trusted public key (PKI/WoT). But the presumption is that you are already doing this over the web (using TLS). That process is subject to attack at the CA. WoT is not subject to a CA attack, because it's decentralized. But it's also not sufficiently deployed for some scenarios.
e
Commit protocol provides both better user experience and better security.
Dňa 6. februára 2015 1:49:12 CET používateľ Paul Puey <paul@airbitz•co> napísal:
Quoted text (56 lines)
>The trust can be considered bootstrapped by visual verification of the >address prefix. If we are really concerned about someone jamming a >Bluetooth signal in a coffeeshop then the UI can encourage verification >of the prefix. Much like how regular Bluetooth requires 'pairing' via >entering a 4-6 digit code. > > > >Paul Puey CEO / Co-Founder, Airbitz Inc >619.850.8624 | http://airbitz.co | San Diego > > > > >On Feb 5, 2015, at 3:46 PM, Eric Voskuil <eric@voskuil•org> wrote: > >On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>> A BIP-70 signed payment request in the initial broadcast can resolve >the >>> integrity issues, but because of the public nature of the broadcast >>> coupled with strong public identity, the privacy compromise is much >>> worse. Now transactions are cryptographically tainted. >>> >>> This is also the problem with BIP-70 over the web. TLS and other >>> security precautions aside, an interloper on the communication, >desktop, >>> datacenter, etc., can capture payment requests and strongly >correlate >>> transactions to identities in an automated manner. The payment >request >>> must be kept private between the parties, and that's hard to do. >> >> What about using encryption with forward secrecy? Merchant would >> generate signed request containing public ECDH part, buyer would send >> back transaction encrypted with ECDH and his public ECDH part. If >> receiving address/amount is meant to be private, use commit protocol >> (see ZRTP/RedPhone) and short authentication phrase (which is hard to >> spoof thanks to commit protocol - see RedPhone)? > >Hi Martin, > >The problem is that you need to verify the ownership of the public key. >A MITM can substitute the key. If you don't have verifiable identity >associated with the public key (PKI/WoT), you need a shared secret >(such >as a secret phrase). But the problem is then establishing that secret >over a public channel. > >You can bootstrap a private session over the untrusted network using a >trusted public key (PKI/WoT). But the presumption is that you are >already doing this over the web (using TLS). That process is subject to >attack at the CA. WoT is not subject to a CA attack, because it's >decentralized. But it's also not sufficiently deployed for some >scenarios. > >e
- -- Odoslané z môjho Android zariadenia pomocou K-9 Mail.
Although consumer to merchant is a use case for BLE I would argue that NFC has a higher chance of providing a better user experience in most cases since, at least on Android, a user can tap their phone without even having a wallet running. The URI handler will launch the wallet for them. However a dedicated, user facing, screen can give certainty that the user is connecting to the correct recipient.
1. Because it can show an address prefix 2. It can display the users nickname/handle upon connecting which is only sent to the merchant upon a point to point connection. Not a broadcast.
The Airbitz wallet already does this on the recipient side. A popup shows the most recent person connecting to the recipient.
Paul Puey CEO / Co-Founder, Airbitz Inc 619.850.8624 | http://airbitz.co | San Diego
On Feb 5, 2015, at 3:34 PM, Roy Badami <roy@gnomon•org•uk> wrote:
For peer-to-peer payments, how common do we think that the payment is of an ad hoc nature rather than to a known contact?
If I want to pay my friends/colleagues/etc over a restaurant table there's no reason why I couldn't already have their public keys in my contact list - then it would be pretty straightforward to have a watertight mechanism where I would know who I was paying. You could probably even relatively securely bootstrap a key exchange over SMS, relying only on the contacts already having each other in their phonebooks.
As for comsumer-to-merchant transactions where the merchant is a bricks and mortar merchant, IMHO it absolutely has to be "pay that terminal over there". It's the trust model we all currently use - whether paying cash or card - and it's the only trust model that works IMHO (and customers and businesses alike are well aware of the risks of a fraudster standing behind the counter pretending to be an employee accepting payment - and by and large are pretty good at mitigating it). OTOH as we've discussed here before there are many use cases where the custoemr doesn't actually know or care about the name of the shop or bar they walked into but is pretty damn sure that they need to make payment to the person over there behind the counter.
Granted, there are cases taht dont' fall into either of the above - but they're the cases that are (a) harder to figure out how to authenticate and consequently (b) the use cases that are going to be most subject to attempted fraud.
roy
Quoted text (43 lines)
> On Thu, Feb 05, 2015 at 03:02:56PM -0800, William Swanson wrote: >> On Thu, Feb 5, 2015 at 2:10 PM, Eric Voskuil <eric@voskuil•org> wrote: >> A MITM can receive the initial broadcast and then spoof it by jamming the >> original. You then only see one. > > You are right, of course. There is no way to make Bluetooth 100% > secure, since it is an over-the-air technology. You could try securing > it using a CA or other identity server, but now you've excluded ad-hoc > person-to-person payments. Plus, you need an active internet > connection to reach the CA. > > You can try using proximity as a substitute for identity, like > requiring NFC to kick-start the connection, but at that point you > might as well use QR codes. > > This BIP is not trying to provide absolute bullet-proof security, > since that's impossible given the physical limitations of the > Bluetooth technology. Instead, it's trying to provide the > best-possible security given those constraints. In exchange for this, > we get greatly enhanced usability in common scenarios. > > There are plenty of usable, real-world technologies with big security > holes. Anybody with lock-picking experience will tell you this, but > nobody is welding their front door shut. The ability to go in and out > is worth the security risk. > > Bluetooth payments add a whole new dimension to real-world Bitcoin > usability. Do we shut that down because it can't be made perfect, or > do we do the best we can and move forward? > > -William > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development >
On 02/05/2015 04:49 PM, Paul Puey wrote:
Quoted text (2 lines)
> The trust can be considered bootstrapped by visual verification of the > address prefix.
Another (unspendable) address can trivially match the prefix. Imagine someone walking around in a mall with a phone in the pocket with a malicious app, just disrupting business by causing money to be burned. Manual verification doesn't fix this attack.
Quoted text (2 lines)
> If we are really concerned about someone jamming a Bluetooth signal > in a coffeeshop then the UI can encourage verification of the prefix.
I don't think it would be great to constrain a standard implementation to low cost purchases or the need for manual verification, but again manual prefix verification isn't actually a solution.
Quoted text (2 lines)
> Much like how regular Bluetooth requires 'pairing' via entering a 4-6 > digit code.
An appeal to the security of BT bootstrapping isn't exactly flattering.
You know I love Airbitz, and I know you guys are extremely privacy conscious. I personally would have no problem using this feature under certain circumstances. My question is only whether it would be wise to standardize on the proposal as-is.
e
Quoted text (38 lines)
> On Feb 5, 2015, at 3:46 PM, Eric Voskuil <eric@voskuil•org > <mailto:eric@voskuil.org>> wrote: > > On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>> A BIP-70 signed payment request in the initial broadcast can resolve the >>> integrity issues, but because of the public nature of the broadcast >>> coupled with strong public identity, the privacy compromise is much >>> worse. Now transactions are cryptographically tainted. >>> >>> This is also the problem with BIP-70 over the web. TLS and other >>> security precautions aside, an interloper on the communication, desktop, >>> datacenter, etc., can capture payment requests and strongly correlate >>> transactions to identities in an automated manner. The payment request >>> must be kept private between the parties, and that's hard to do. >> >> What about using encryption with forward secrecy? Merchant would >> generate signed request containing public ECDH part, buyer would send >> back transaction encrypted with ECDH and his public ECDH part. If >> receiving address/amount is meant to be private, use commit protocol >> (see ZRTP/RedPhone) and short authentication phrase (which is hard to >> spoof thanks to commit protocol - see RedPhone)? > > Hi Martin, > > The problem is that you need to verify the ownership of the public key. > A MITM can substitute the key. If you don't have verifiable identity > associated with the public key (PKI/WoT), you need a shared secret (such > as a secret phrase). But the problem is then establishing that secret > over a public channel. > > You can bootstrap a private session over the untrusted network using a > trusted public key (PKI/WoT). But the presumption is that you are > already doing this over the web (using TLS). That process is subject to > attack at the CA. WoT is not subject to a CA attack, because it's > decentralized. But it's also not sufficiently deployed for some scenarios. > > e >
On 02/05/2015 04:36 PM, Martin Habovštiak wrote:
Quoted text (3 lines)
> I believe, we are still talking about transactions of physical > people in physical world. So yes, it's proximity based - people > tell the words by mouth. :)
Notice from my original comment:
Quoted text (3 lines)
>>>> A MITM can substitute the key. If you don't have verifiable >>>> identity associated with the public key (PKI/WoT), you need >>>> a shared secret (such as a secret phrase).
I said this could only be accomplished using a shared secret or a trusted public key. Exchanging a value that is derived from a pair of public keys is a distinction without a difference. The problem remains that the parties must have a secure/out-of-band channel for communicating this value.
The fact that they are face-to-face establishes this channel, but that brings us back to the original problem, as it requires manual verification - as in visual/audible scanning of the two values for comparison. At that point the visual comparison of the address, or some value derived from it, is simpler.
Quoted text (4 lines)
> In case of RedPhone, you read those words verbally over not-yet- > verified channel relying on difficulty of spoofing your voice. Also > the app remembers the public keys, so you don't need to verify > second time.
This is reasonable, but wouldn't help in the case of an ad-hoc connection between parties who don't know each other well.
Quoted text (5 lines)
> I suggest you to try RedPhone (called Signal on iPhone) yourself. > It's free/open source, Internet-based and end-to-end encrypted. You > may find it useful some day. Also I'm willing to help you with > trying it after I wake up. (~8 hours: Send me private e-mail if > you want to.)
I appreciate the offer. I really don't trust *any* smartphone as a platform for secure communication/data. But encrypting on the wire does of course shrink the attack surface and increase the attacker's cost.
e
Quoted text (1 line)
> Dňa 6. februára 2015 1:22:23 CET používateľ Eric Voskuil
<eric@voskuil•org> napísal:
Quoted text (97 lines)
>> On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote: >>> That's exactly what I though when seeing the RedPhone code, but after >>> I studied the commit protocol I realized it's actually secure and >>> convenient way to do it. You should do that too. :) > >> I was analyzing the model as you described it to me. A formal analysis >> of the security model of a particular implementation, based on >> inference >>from source code, is a bit beyond what I signed up for. But I'm >> perfectly willing to comment on your description of the model if you >> are >> willing to indulge me. > >>> Shortly, how it works: >>> The initiator of the connection sends commit message containing the >>> hash of his temporary public ECDH part, second party sends back their >>> public ECDH part and then initiator sends his public ECDH part in >>> open. All three messages are hashed together and the first two bytes >>> are used to select two words from a shared dictionary which are >>> displayed on the screen of both the initiator and the second party. > >>> The parties communicate those two words and verify they match. > >> How do they compare words if they haven't yet established a secure >> channel? > >>> If an attacker wants to do MITM, he has a chance of choosing right >>> public parts 1:65536. There is no way to brute-force it, since that >>> would be noticed immediately. If instead of two words based on the >>> first two bytes, four words from BIP39 wordlist were chosen, it would >>> provide entropy of 44 bits which I believe should be enough even for >>> paranoid people. >>> >>> How this would work in Bitcoin payment scenario: user's phone >>> broadcasts his name, merchant inputs amount and selects the name from >>> the list, commit message is sent (and then the remaining two >>> messages), merchant spells four words he sees on the screen and buyer >>> confirms transaction after verifying that words match. > >> So the assumption is that there exists a secure (as in proximity-based) >> communication channel? > >> e > >>> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>>> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>> A BIP-70 signed payment request in the initial broadcast can >> resolve the >>>>>> integrity issues, but because of the public nature of the >> broadcast >>>>>> coupled with strong public identity, the privacy compromise is >> much >>>>>> worse. Now transactions are cryptographically tainted. >>>>>> >>>>>> This is also the problem with BIP-70 over the web. TLS and other >>>>>> security precautions aside, an interloper on the communication, >> desktop, >>>>>> datacenter, etc., can capture payment requests and strongly >> correlate >>>>>> transactions to identities in an automated manner. The payment >> request >>>>>> must be kept private between the parties, and that's hard to do. >>>>> >>>>> What about using encryption with forward secrecy? Merchant would >>>>> generate signed request containing public ECDH part, buyer would >> send >>>>> back transaction encrypted with ECDH and his public ECDH part. If >>>>> receiving address/amount is meant to be private, use commit >> protocol >>>>> (see ZRTP/RedPhone) and short authentication phrase (which is hard >> to >>>>> spoof thanks to commit protocol - see RedPhone)? >>>> >>>> Hi Martin, >>>> >>>> The problem is that you need to verify the ownership of the public >> key. >>>> A MITM can substitute the key. If you don't have verifiable identity >>>> associated with the public key (PKI/WoT), you need a shared secret >> (such >>>> as a secret phrase). But the problem is then establishing that >> secret >>>> over a public channel. >>>> >>>> You can bootstrap a private session over the untrusted network using >> a >>>> trusted public key (PKI/WoT). But the presumption is that you are >>>> already doing this over the web (using TLS). That process is subject >> to >>>> attack at the CA. WoT is not subject to a CA attack, because it's >>>> decentralized. But it's also not sufficiently deployed for some >> scenarios. >>>> >>>> e >>>> > >
Thanks for all the feedback Eric. You know we value all that you have to say. That's what this forum is for. We're looking for great ideas to harden this protocol and we're not closed to better ideas and we'll improve it as suggestions come up.
Paul Puey CEO / Co-Founder, Airbitz Inc 619.850.8624 | http://airbitz.co | San Diego
On Feb 5, 2015, at 5:05 PM, Eric Voskuil <eric@voskuil•org> wrote:
Quoted text (3 lines)
> On 02/05/2015 04:49 PM, Paul Puey wrote: > The trust can be considered bootstrapped by visual verification of the > address prefix.
Another (unspendable) address can trivially match the prefix. Imagine someone walking around in a mall with a phone in the pocket with a malicious app, just disrupting business by causing money to be burned. Manual verification doesn't fix this attack.
Quoted text (2 lines)
> If we are really concerned about someone jamming a Bluetooth signal > in a coffeeshop then the UI can encourage verification of the prefix.
I don't think it would be great to constrain a standard implementation to low cost purchases or the need for manual verification, but again manual prefix verification isn't actually a solution.
Quoted text (2 lines)
> Much like how regular Bluetooth requires 'pairing' via entering a 4-6 > digit code.
An appeal to the security of BT bootstrapping isn't exactly flattering.
You know I love Airbitz, and I know you guys are extremely privacy conscious. I personally would have no problem using this feature under certain circumstances. My question is only whether it would be wise to standardize on the proposal as-is.
e
Quoted text (37 lines)
> On Feb 5, 2015, at 3:46 PM, Eric Voskuil <eric@voskuil•org > <mailto:eric@voskuil.org>> wrote: > > On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>> A BIP-70 signed payment request in the initial broadcast can resolve the >>> integrity issues, but because of the public nature of the broadcast >>> coupled with strong public identity, the privacy compromise is much >>> worse. Now transactions are cryptographically tainted. >>> >>> This is also the problem with BIP-70 over the web. TLS and other >>> security precautions aside, an interloper on the communication, desktop, >>> datacenter, etc., can capture payment requests and strongly correlate >>> transactions to identities in an automated manner. The payment request >>> must be kept private between the parties, and that's hard to do. >> >> What about using encryption with forward secrecy? Merchant would >> generate signed request containing public ECDH part, buyer would send >> back transaction encrypted with ECDH and his public ECDH part. If >> receiving address/amount is meant to be private, use commit protocol >> (see ZRTP/RedPhone) and short authentication phrase (which is hard to >> spoof thanks to commit protocol - see RedPhone)? > > Hi Martin, > > The problem is that you need to verify the ownership of the public key. > A MITM can substitute the key. If you don't have verifiable identity > associated with the public key (PKI/WoT), you need a shared secret (such > as a secret phrase). But the problem is then establishing that secret > over a public channel. > > You can bootstrap a private session over the untrusted network using a > trusted public key (PKI/WoT). But the presumption is that you are > already doing this over the web (using TLS). That process is subject to > attack at the CA. WoT is not subject to a CA attack, because it's > decentralized. But it's also not sufficiently deployed for some scenarios. > > e
Quoted text (8 lines)
> In this case there is no need for P2P communication, just pay to an > address you already have for the other party. If you want to avoid > address reuse, use stealth addressing. > > But yes, if you don't have a stealth address for the other party you can > certainly communicate in private as peers where you trust that you share > a public key. The core issue here is really bootstrapping of that trust > in an ad hoc manner.
Something interactive might still be nicer, though, to avoid the risk of paying to an address that the payee no longer has the private key for. "Nooo!! Don't pay to that address. I lost my old phone so I generated a new wallet."
roy
2015-02-06 2:29 GMT+01:00 Eric Voskuil <eric@voskuil•org>:
Quoted text (22 lines)
> On 02/05/2015 04:36 PM, Martin Habovštiak wrote: >> I believe, we are still talking about transactions of physical >> people in physical world. So yes, it's proximity based - people >> tell the words by mouth. :) > > Notice from my original comment: > >>>>> A MITM can substitute the key. If you don't have verifiable >>>>> identity associated with the public key (PKI/WoT), you need >>>>> a shared secret (such as a secret phrase). > > I said this could only be accomplished using a shared secret or a > trusted public key. Exchanging a value that is derived from a pair of > public keys is a distinction without a difference. The problem remains > that the parties must have a secure/out-of-band channel for > communicating this value. > > The fact that they are face-to-face establishes this channel, but that > brings us back to the original problem, as it requires manual > verification - as in visual/audible scanning of the two values for > comparison. At that point the visual comparison of the address, or some > value derived from it, is simpler.
I have never been against manual verification. What I'm trying to say is let's just make manual verification easier and more secure. Comparison of address is simpler for the coder but also simpler to attack. It has these problems: - Addresses broadcasted in plaintext (privacy issue) - Amounts broadcasted in plaintext (privacy issue) - Address is long - takes lot of time to verify (user experience issue) - Address prefix can be brute-forced, if too short or used to make "black hole" address if longer (vandalism issue)
Commit protocol can be used for both the encryption and the authentication while user experience is not bad and everything is still secure.
Quoted text (122 lines)
> >> In case of RedPhone, you read those words verbally over not-yet- >> verified channel relying on difficulty of spoofing your voice. Also >> the app remembers the public keys, so you don't need to verify >> second time. > > This is reasonable, but wouldn't help in the case of an ad-hoc > connection between parties who don't know each other well. > >> I suggest you to try RedPhone (called Signal on iPhone) yourself. >> It's free/open source, Internet-based and end-to-end encrypted. You >> may find it useful some day. Also I'm willing to help you with >> trying it after I wake up. (~8 hours: Send me private e-mail if >> you want to.) > > I appreciate the offer. I really don't trust *any* smartphone as a > platform for secure communication/data. But encrypting on the wire does > of course shrink the attack surface and increase the attacker's cost. > > e > >> Dňa 6. februára 2015 1:22:23 CET používateľ Eric Voskuil > <eric@voskuil•org> napísal: > >>> On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote: >>>> That's exactly what I though when seeing the RedPhone code, but after >>>> I studied the commit protocol I realized it's actually secure and >>>> convenient way to do it. You should do that too. :) >> >>> I was analyzing the model as you described it to me. A formal analysis >>> of the security model of a particular implementation, based on >>> inference >>>from source code, is a bit beyond what I signed up for. But I'm >>> perfectly willing to comment on your description of the model if you >>> are >>> willing to indulge me. >> >>>> Shortly, how it works: >>>> The initiator of the connection sends commit message containing the >>>> hash of his temporary public ECDH part, second party sends back their >>>> public ECDH part and then initiator sends his public ECDH part in >>>> open. All three messages are hashed together and the first two bytes >>>> are used to select two words from a shared dictionary which are >>>> displayed on the screen of both the initiator and the second party. >> >>>> The parties communicate those two words and verify they match. >> >>> How do they compare words if they haven't yet established a secure >>> channel? >> >>>> If an attacker wants to do MITM, he has a chance of choosing right >>>> public parts 1:65536. There is no way to brute-force it, since that >>>> would be noticed immediately. If instead of two words based on the >>>> first two bytes, four words from BIP39 wordlist were chosen, it would >>>> provide entropy of 44 bits which I believe should be enough even for >>>> paranoid people. >>>> >>>> How this would work in Bitcoin payment scenario: user's phone >>>> broadcasts his name, merchant inputs amount and selects the name from >>>> the list, commit message is sent (and then the remaining two >>>> messages), merchant spells four words he sees on the screen and buyer >>>> confirms transaction after verifying that words match. >> >>> So the assumption is that there exists a secure (as in proximity-based) >>> communication channel? >> >>> e >> >>>> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>>>> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>>> A BIP-70 signed payment request in the initial broadcast can >>> resolve the >>>>>>> integrity issues, but because of the public nature of the >>> broadcast >>>>>>> coupled with strong public identity, the privacy compromise is >>> much >>>>>>> worse. Now transactions are cryptographically tainted. >>>>>>> >>>>>>> This is also the problem with BIP-70 over the web. TLS and other >>>>>>> security precautions aside, an interloper on the communication, >>> desktop, >>>>>>> datacenter, etc., can capture payment requests and strongly >>> correlate >>>>>>> transactions to identities in an automated manner. The payment >>> request >>>>>>> must be kept private between the parties, and that's hard to do. >>>>>> >>>>>> What about using encryption with forward secrecy? Merchant would >>>>>> generate signed request containing public ECDH part, buyer would >>> send >>>>>> back transaction encrypted with ECDH and his public ECDH part. If >>>>>> receiving address/amount is meant to be private, use commit >>> protocol >>>>>> (see ZRTP/RedPhone) and short authentication phrase (which is hard >>> to >>>>>> spoof thanks to commit protocol - see RedPhone)? >>>>> >>>>> Hi Martin, >>>>> >>>>> The problem is that you need to verify the ownership of the public >>> key. >>>>> A MITM can substitute the key. If you don't have verifiable identity >>>>> associated with the public key (PKI/WoT), you need a shared secret >>> (such >>>>> as a secret phrase). But the problem is then establishing that >>> secret >>>>> over a public channel. >>>>> >>>>> You can bootstrap a private session over the untrusted network using >>> a >>>>> trusted public key (PKI/WoT). But the presumption is that you are >>>>> already doing this over the web (using TLS). That process is subject >>> to >>>>> attack at the CA. WoT is not subject to a CA attack, because it's >>>>> decentralized. But it's also not sufficiently deployed for some >>> scenarios. >>>>> >>>>> e >>>>> >> >> >
On 02/06/2015 12:59 AM, Roy Badami wrote:
Quoted text (13 lines)
>> In this case there is no need for P2P communication, just pay to an >> address you already have for the other party. If you want to avoid >> address reuse, use stealth addressing. >> >> But yes, if you don't have a stealth address for the other party you can >> certainly communicate in private as peers where you trust that you share >> a public key. The core issue here is really bootstrapping of that trust >> in an ad hoc manner. > > Something interactive might still be nicer, though, to avoid the risk > of paying to an address that the payee no longer has the private key > for. "Nooo!! Don't pay to that address. I lost my old phone so I > generated a new wallet."
Certainly, which brings us back to proximity.
Which reminds me - it's important to keep in mind the scenario that arises when there is no person present to represent the receiver. Such as a vending machine purchase.
Proximity in these cases is insufficient, as the receiver is not able to prevent application of a fraudulent NFC device or replacement of a static QR code. In these cases BIP-70 becomes essential.
e
Martin,
I like your idea for the commit protocol in that it resolves the vandalous address substitution attack. However, I don't see a way to prevent privacy loss without adverse impact to the scenario.
Anyone could perform the handshake and thereby obtain the payment request. Therefore to prevent inadvertent disclosure the customer must visually confirm the "phrase" and then verbally tell the merchant to proceed by sending the payment request.
One might argue that it's sufficient to preserve the integrity of the transaction while suffering the privacy loss, especially given that a hijacked handshake should never result in a completed transaction - unless of course the hijacker pays.
But imagine someone purchasing their meds. HIPAA requires the checkout queue to form behind a yellow line. That speaks directly to this question.
e
On 02/06/2015 01:07 AM, MⒶrtin HⒶboⓋštiak wrote:
Quoted text (160 lines)
> 2015-02-06 2:29 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >> On 02/05/2015 04:36 PM, Martin Habovštiak wrote: >>> I believe, we are still talking about transactions of physical >>> people in physical world. So yes, it's proximity based - people >>> tell the words by mouth. :) >> >> Notice from my original comment: >> >>>>>> A MITM can substitute the key. If you don't have verifiable >>>>>> identity associated with the public key (PKI/WoT), you need >>>>>> a shared secret (such as a secret phrase). >> >> I said this could only be accomplished using a shared secret or a >> trusted public key. Exchanging a value that is derived from a pair of >> public keys is a distinction without a difference. The problem remains >> that the parties must have a secure/out-of-band channel for >> communicating this value. >> >> The fact that they are face-to-face establishes this channel, but that >> brings us back to the original problem, as it requires manual >> verification - as in visual/audible scanning of the two values for >> comparison. At that point the visual comparison of the address, or some >> value derived from it, is simpler. > > I have never been against manual verification. What I'm trying to say > is let's just make manual verification easier and more secure. > Comparison of address is simpler for the coder but also simpler to > attack. It has these problems: > - Addresses broadcasted in plaintext (privacy issue) > - Amounts broadcasted in plaintext (privacy issue) > - Address is long - takes lot of time to verify (user experience issue) > - Address prefix can be brute-forced, if too short or used to make > "black hole" address if longer (vandalism issue) > > Commit protocol can be used for both the encryption and the > authentication while user experience is not bad and everything is > still secure. > >> >>> In case of RedPhone, you read those words verbally over not-yet- >>> verified channel relying on difficulty of spoofing your voice. Also >>> the app remembers the public keys, so you don't need to verify >>> second time. >> >> This is reasonable, but wouldn't help in the case of an ad-hoc >> connection between parties who don't know each other well. >> >>> I suggest you to try RedPhone (called Signal on iPhone) yourself. >>> It's free/open source, Internet-based and end-to-end encrypted. You >>> may find it useful some day. Also I'm willing to help you with >>> trying it after I wake up. (~8 hours: Send me private e-mail if >>> you want to.) >> >> I appreciate the offer. I really don't trust *any* smartphone as a >> platform for secure communication/data. But encrypting on the wire does >> of course shrink the attack surface and increase the attacker's cost. >> >> e >> >>> Dňa 6. februára 2015 1:22:23 CET používateľ Eric Voskuil >> <eric@voskuil•org> napísal: >> >>>> On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>> That's exactly what I though when seeing the RedPhone code, but after >>>>> I studied the commit protocol I realized it's actually secure and >>>>> convenient way to do it. You should do that too. :) >>> >>>> I was analyzing the model as you described it to me. A formal analysis >>>> of the security model of a particular implementation, based on >>>> inference >>> >from source code, is a bit beyond what I signed up for. But I'm >>>> perfectly willing to comment on your description of the model if you >>>> are >>>> willing to indulge me. >>> >>>>> Shortly, how it works: >>>>> The initiator of the connection sends commit message containing the >>>>> hash of his temporary public ECDH part, second party sends back their >>>>> public ECDH part and then initiator sends his public ECDH part in >>>>> open. All three messages are hashed together and the first two bytes >>>>> are used to select two words from a shared dictionary which are >>>>> displayed on the screen of both the initiator and the second party. >>> >>>>> The parties communicate those two words and verify they match. >>> >>>> How do they compare words if they haven't yet established a secure >>>> channel? >>> >>>>> If an attacker wants to do MITM, he has a chance of choosing right >>>>> public parts 1:65536. There is no way to brute-force it, since that >>>>> would be noticed immediately. If instead of two words based on the >>>>> first two bytes, four words from BIP39 wordlist were chosen, it would >>>>> provide entropy of 44 bits which I believe should be enough even for >>>>> paranoid people. >>>>> >>>>> How this would work in Bitcoin payment scenario: user's phone >>>>> broadcasts his name, merchant inputs amount and selects the name from >>>>> the list, commit message is sent (and then the remaining two >>>>> messages), merchant spells four words he sees on the screen and buyer >>>>> confirms transaction after verifying that words match. >>> >>>> So the assumption is that there exists a secure (as in proximity-based) >>>> communication channel? >>> >>>> e >>> >>>>> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>>>>> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>>>> A BIP-70 signed payment request in the initial broadcast can >>>> resolve the >>>>>>>> integrity issues, but because of the public nature of the >>>> broadcast >>>>>>>> coupled with strong public identity, the privacy compromise is >>>> much >>>>>>>> worse. Now transactions are cryptographically tainted. >>>>>>>> >>>>>>>> This is also the problem with BIP-70 over the web. TLS and other >>>>>>>> security precautions aside, an interloper on the communication, >>>> desktop, >>>>>>>> datacenter, etc., can capture payment requests and strongly >>>> correlate >>>>>>>> transactions to identities in an automated manner. The payment >>>> request >>>>>>>> must be kept private between the parties, and that's hard to do. >>>>>>> >>>>>>> What about using encryption with forward secrecy? Merchant would >>>>>>> generate signed request containing public ECDH part, buyer would >>>> send >>>>>>> back transaction encrypted with ECDH and his public ECDH part. If >>>>>>> receiving address/amount is meant to be private, use commit >>>> protocol >>>>>>> (see ZRTP/RedPhone) and short authentication phrase (which is hard >>>> to >>>>>>> spoof thanks to commit protocol - see RedPhone)? >>>>>> >>>>>> Hi Martin, >>>>>> >>>>>> The problem is that you need to verify the ownership of the public >>>> key. >>>>>> A MITM can substitute the key. If you don't have verifiable identity >>>>>> associated with the public key (PKI/WoT), you need a shared secret >>>> (such >>>>>> as a secret phrase). But the problem is then establishing that >>>> secret >>>>>> over a public channel. >>>>>> >>>>>> You can bootstrap a private session over the untrusted network using >>>> a >>>>>> trusted public key (PKI/WoT). But the presumption is that you are >>>>>> already doing this over the web (using TLS). That process is subject >>>> to >>>>>> attack at the CA. WoT is not subject to a CA attack, because it's >>>>>> decentralized. But it's also not sufficiently deployed for some >>>> scenarios. >>>>>> >>>>>> e >>>>>> >>> >>> >>
I'm not sure if I was clear enough. Handshake should be used to establish authenticated AND encrypted communication using ECDH (or just DH, but I think it's easier to use ECDH, since required functions are already used in Bitcoin protocol), like RedPhone does. BTW knowledge of verification string is useless to the attacker.
Yes, the customer must verify it verbally and the merchant shouldn't send the transaction before verification. Other possibility is that in case of differing verification strings new address is generated, so attacker doesn't know the address. But in this case, amount is leaked and there is quite high probability it can be found in the Blockchain. Anyway, I don't believe the transaction can be made securely without such interaction except with white-listing public keys, so I see no reason why interaction should be problematic.
We don't have such strict regulations but I agree that security is important. Currently I think that verbal verification and manual confirmation is the best way to achieve high security and reasonable user-friendliness.
2015-02-10 17:55 GMT+01:00 Eric Voskuil <eric@voskuil•org>:
Quoted text (183 lines)
> Martin, > > I like your idea for the commit protocol in that it resolves the > vandalous address substitution attack. However, I don't see a way to > prevent privacy loss without adverse impact to the scenario. > > Anyone could perform the handshake and thereby obtain the payment > request. Therefore to prevent inadvertent disclosure the customer must > visually confirm the "phrase" and then verbally tell the merchant to > proceed by sending the payment request. > > One might argue that it's sufficient to preserve the integrity of the > transaction while suffering the privacy loss, especially given that a > hijacked handshake should never result in a completed transaction - > unless of course the hijacker pays. > > But imagine someone purchasing their meds. HIPAA requires the checkout > queue to form behind a yellow line. That speaks directly to this question. > > e > > On 02/06/2015 01:07 AM, MⒶrtin HⒶboⓋštiak wrote: >> 2015-02-06 2:29 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>> On 02/05/2015 04:36 PM, Martin Habovštiak wrote: >>>> I believe, we are still talking about transactions of physical >>>> people in physical world. So yes, it's proximity based - people >>>> tell the words by mouth. :) >>> >>> Notice from my original comment: >>> >>>>>>> A MITM can substitute the key. If you don't have verifiable >>>>>>> identity associated with the public key (PKI/WoT), you need >>>>>>> a shared secret (such as a secret phrase). >>> >>> I said this could only be accomplished using a shared secret or a >>> trusted public key. Exchanging a value that is derived from a pair of >>> public keys is a distinction without a difference. The problem remains >>> that the parties must have a secure/out-of-band channel for >>> communicating this value. >>> >>> The fact that they are face-to-face establishes this channel, but that >>> brings us back to the original problem, as it requires manual >>> verification - as in visual/audible scanning of the two values for >>> comparison. At that point the visual comparison of the address, or some >>> value derived from it, is simpler. >> >> I have never been against manual verification. What I'm trying to say >> is let's just make manual verification easier and more secure. >> Comparison of address is simpler for the coder but also simpler to >> attack. It has these problems: >> - Addresses broadcasted in plaintext (privacy issue) >> - Amounts broadcasted in plaintext (privacy issue) >> - Address is long - takes lot of time to verify (user experience issue) >> - Address prefix can be brute-forced, if too short or used to make >> "black hole" address if longer (vandalism issue) >> >> Commit protocol can be used for both the encryption and the >> authentication while user experience is not bad and everything is >> still secure. >> >>> >>>> In case of RedPhone, you read those words verbally over not-yet- >>>> verified channel relying on difficulty of spoofing your voice. Also >>>> the app remembers the public keys, so you don't need to verify >>>> second time. >>> >>> This is reasonable, but wouldn't help in the case of an ad-hoc >>> connection between parties who don't know each other well. >>> >>>> I suggest you to try RedPhone (called Signal on iPhone) yourself. >>>> It's free/open source, Internet-based and end-to-end encrypted. You >>>> may find it useful some day. Also I'm willing to help you with >>>> trying it after I wake up. (~8 hours: Send me private e-mail if >>>> you want to.) >>> >>> I appreciate the offer. I really don't trust *any* smartphone as a >>> platform for secure communication/data. But encrypting on the wire does >>> of course shrink the attack surface and increase the attacker's cost. >>> >>> e >>> >>>> Dňa 6. februára 2015 1:22:23 CET používateľ Eric Voskuil >>> <eric@voskuil•org> napísal: >>> >>>>> On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>> That's exactly what I though when seeing the RedPhone code, but after >>>>>> I studied the commit protocol I realized it's actually secure and >>>>>> convenient way to do it. You should do that too. :) >>>> >>>>> I was analyzing the model as you described it to me. A formal analysis >>>>> of the security model of a particular implementation, based on >>>>> inference >>>> >from source code, is a bit beyond what I signed up for. But I'm >>>>> perfectly willing to comment on your description of the model if you >>>>> are >>>>> willing to indulge me. >>>> >>>>>> Shortly, how it works: >>>>>> The initiator of the connection sends commit message containing the >>>>>> hash of his temporary public ECDH part, second party sends back their >>>>>> public ECDH part and then initiator sends his public ECDH part in >>>>>> open. All three messages are hashed together and the first two bytes >>>>>> are used to select two words from a shared dictionary which are >>>>>> displayed on the screen of both the initiator and the second party. >>>> >>>>>> The parties communicate those two words and verify they match. >>>> >>>>> How do they compare words if they haven't yet established a secure >>>>> channel? >>>> >>>>>> If an attacker wants to do MITM, he has a chance of choosing right >>>>>> public parts 1:65536. There is no way to brute-force it, since that >>>>>> would be noticed immediately. If instead of two words based on the >>>>>> first two bytes, four words from BIP39 wordlist were chosen, it would >>>>>> provide entropy of 44 bits which I believe should be enough even for >>>>>> paranoid people. >>>>>> >>>>>> How this would work in Bitcoin payment scenario: user's phone >>>>>> broadcasts his name, merchant inputs amount and selects the name from >>>>>> the list, commit message is sent (and then the remaining two >>>>>> messages), merchant spells four words he sees on the screen and buyer >>>>>> confirms transaction after verifying that words match. >>>> >>>>> So the assumption is that there exists a secure (as in proximity-based) >>>>> communication channel? >>>> >>>>> e >>>> >>>>>> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>>>>>> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>>>>> A BIP-70 signed payment request in the initial broadcast can >>>>> resolve the >>>>>>>>> integrity issues, but because of the public nature of the >>>>> broadcast >>>>>>>>> coupled with strong public identity, the privacy compromise is >>>>> much >>>>>>>>> worse. Now transactions are cryptographically tainted. >>>>>>>>> >>>>>>>>> This is also the problem with BIP-70 over the web. TLS and other >>>>>>>>> security precautions aside, an interloper on the communication, >>>>> desktop, >>>>>>>>> datacenter, etc., can capture payment requests and strongly >>>>> correlate >>>>>>>>> transactions to identities in an automated manner. The payment >>>>> request >>>>>>>>> must be kept private between the parties, and that's hard to do. >>>>>>>> >>>>>>>> What about using encryption with forward secrecy? Merchant would >>>>>>>> generate signed request containing public ECDH part, buyer would >>>>> send >>>>>>>> back transaction encrypted with ECDH and his public ECDH part. If >>>>>>>> receiving address/amount is meant to be private, use commit >>>>> protocol >>>>>>>> (see ZRTP/RedPhone) and short authentication phrase (which is hard >>>>> to >>>>>>>> spoof thanks to commit protocol - see RedPhone)? >>>>>>> >>>>>>> Hi Martin, >>>>>>> >>>>>>> The problem is that you need to verify the ownership of the public >>>>> key. >>>>>>> A MITM can substitute the key. If you don't have verifiable identity >>>>>>> associated with the public key (PKI/WoT), you need a shared secret >>>>> (such >>>>>>> as a secret phrase). But the problem is then establishing that >>>>> secret >>>>>>> over a public channel. >>>>>>> >>>>>>> You can bootstrap a private session over the untrusted network using >>>>> a >>>>>>> trusted public key (PKI/WoT). But the presumption is that you are >>>>>>> already doing this over the web (using TLS). That process is subject >>>>> to >>>>>>> attack at the CA. WoT is not subject to a CA attack, because it's >>>>>>> decentralized. But it's also not sufficiently deployed for some >>>>> scenarios. >>>>>>> >>>>>>> e >>>>>>> >>>> >>>> >>> >
On 02/10/2015 09:16 AM, MⒶrtin HⒶboⓋštiak wrote:
Quoted text (5 lines)
> I'm not sure if I was clear enough. Handshake should be used to > establish authenticated AND encrypted communication using ECDH (or > just DH, but I think it's easier to use ECDH, since required functions > are already used in Bitcoin protocol), like RedPhone does. BTW > knowledge of verification string is useless to the attacker.
Yes, I think this was clear from your description.
Quoted text (5 lines)
> Yes, the customer must verify it verbally and the merchant shouldn't > send the transaction before verification. Other possibility is that in > case of differing verification strings new address is generated, so > attacker doesn't know the address. But in this case, amount is leaked > and there is quite high probability it can be found in the Blockchain.
Yes, for each handshake the payment request would need to contain a different address, mitigating some of the privacy loss.
Quoted text (3 lines)
> Anyway, I don't believe the transaction can be made securely without > such interaction except with white-listing public keys, so I see no > reason why interaction should be problematic.
It can be done securely and privately by transfer of a shared secret through a private channel.
Quoted text (4 lines)
> We don't have such strict regulations but I agree that security is > important. Currently I think that verbal verification and manual > confirmation is the best way to achieve high security and reasonable > user-friendliness.
I think for a broadcast model (e.g. Bluetooth only) that is the only want to ensure integrity and privacy. A narrow cast can use proximity to establish trust.
Quoted text (184 lines)
> 2015-02-10 17:55 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >> Martin, >> >> I like your idea for the commit protocol in that it resolves the >> vandalous address substitution attack. However, I don't see a way to >> prevent privacy loss without adverse impact to the scenario. >> >> Anyone could perform the handshake and thereby obtain the payment >> request. Therefore to prevent inadvertent disclosure the customer must >> visually confirm the "phrase" and then verbally tell the merchant to >> proceed by sending the payment request. >> >> One might argue that it's sufficient to preserve the integrity of the >> transaction while suffering the privacy loss, especially given that a >> hijacked handshake should never result in a completed transaction - >> unless of course the hijacker pays. >> >> But imagine someone purchasing their meds. HIPAA requires the checkout >> queue to form behind a yellow line. That speaks directly to this question. >> >> e >> >> On 02/06/2015 01:07 AM, MⒶrtin HⒶboⓋštiak wrote: >>> 2015-02-06 2:29 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>>> On 02/05/2015 04:36 PM, Martin Habovštiak wrote: >>>>> I believe, we are still talking about transactions of physical >>>>> people in physical world. So yes, it's proximity based - people >>>>> tell the words by mouth. :) >>>> >>>> Notice from my original comment: >>>> >>>>>>>> A MITM can substitute the key. If you don't have verifiable >>>>>>>> identity associated with the public key (PKI/WoT), you need >>>>>>>> a shared secret (such as a secret phrase). >>>> >>>> I said this could only be accomplished using a shared secret or a >>>> trusted public key. Exchanging a value that is derived from a pair of >>>> public keys is a distinction without a difference. The problem remains >>>> that the parties must have a secure/out-of-band channel for >>>> communicating this value. >>>> >>>> The fact that they are face-to-face establishes this channel, but that >>>> brings us back to the original problem, as it requires manual >>>> verification - as in visual/audible scanning of the two values for >>>> comparison. At that point the visual comparison of the address, or some >>>> value derived from it, is simpler. >>> >>> I have never been against manual verification. What I'm trying to say >>> is let's just make manual verification easier and more secure. >>> Comparison of address is simpler for the coder but also simpler to >>> attack. It has these problems: >>> - Addresses broadcasted in plaintext (privacy issue) >>> - Amounts broadcasted in plaintext (privacy issue) >>> - Address is long - takes lot of time to verify (user experience issue) >>> - Address prefix can be brute-forced, if too short or used to make >>> "black hole" address if longer (vandalism issue) >>> >>> Commit protocol can be used for both the encryption and the >>> authentication while user experience is not bad and everything is >>> still secure. >>> >>>> >>>>> In case of RedPhone, you read those words verbally over not-yet- >>>>> verified channel relying on difficulty of spoofing your voice. Also >>>>> the app remembers the public keys, so you don't need to verify >>>>> second time. >>>> >>>> This is reasonable, but wouldn't help in the case of an ad-hoc >>>> connection between parties who don't know each other well. >>>> >>>>> I suggest you to try RedPhone (called Signal on iPhone) yourself. >>>>> It's free/open source, Internet-based and end-to-end encrypted. You >>>>> may find it useful some day. Also I'm willing to help you with >>>>> trying it after I wake up. (~8 hours: Send me private e-mail if >>>>> you want to.) >>>> >>>> I appreciate the offer. I really don't trust *any* smartphone as a >>>> platform for secure communication/data. But encrypting on the wire does >>>> of course shrink the attack surface and increase the attacker's cost. >>>> >>>> e >>>> >>>>> Dňa 6. februára 2015 1:22:23 CET používateľ Eric Voskuil >>>> <eric@voskuil•org> napísal: >>>> >>>>>> On 02/05/2015 04:04 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>>> That's exactly what I though when seeing the RedPhone code, but after >>>>>>> I studied the commit protocol I realized it's actually secure and >>>>>>> convenient way to do it. You should do that too. :) >>>>> >>>>>> I was analyzing the model as you described it to me. A formal analysis >>>>>> of the security model of a particular implementation, based on >>>>>> inference >>>>> >from source code, is a bit beyond what I signed up for. But I'm >>>>>> perfectly willing to comment on your description of the model if you >>>>>> are >>>>>> willing to indulge me. >>>>> >>>>>>> Shortly, how it works: >>>>>>> The initiator of the connection sends commit message containing the >>>>>>> hash of his temporary public ECDH part, second party sends back their >>>>>>> public ECDH part and then initiator sends his public ECDH part in >>>>>>> open. All three messages are hashed together and the first two bytes >>>>>>> are used to select two words from a shared dictionary which are >>>>>>> displayed on the screen of both the initiator and the second party. >>>>> >>>>>>> The parties communicate those two words and verify they match. >>>>> >>>>>> How do they compare words if they haven't yet established a secure >>>>>> channel? >>>>> >>>>>>> If an attacker wants to do MITM, he has a chance of choosing right >>>>>>> public parts 1:65536. There is no way to brute-force it, since that >>>>>>> would be noticed immediately. If instead of two words based on the >>>>>>> first two bytes, four words from BIP39 wordlist were chosen, it would >>>>>>> provide entropy of 44 bits which I believe should be enough even for >>>>>>> paranoid people. >>>>>>> >>>>>>> How this would work in Bitcoin payment scenario: user's phone >>>>>>> broadcasts his name, merchant inputs amount and selects the name from >>>>>>> the list, commit message is sent (and then the remaining two >>>>>>> messages), merchant spells four words he sees on the screen and buyer >>>>>>> confirms transaction after verifying that words match. >>>>> >>>>>> So the assumption is that there exists a secure (as in proximity-based) >>>>>> communication channel? >>>>> >>>>>> e >>>>> >>>>>>> 2015-02-06 0:46 GMT+01:00 Eric Voskuil <eric@voskuil•org>: >>>>>>>> On 02/05/2015 03:36 PM, MⒶrtin HⒶboⓋštiak wrote: >>>>>>>>>> A BIP-70 signed payment request in the initial broadcast can >>>>>> resolve the >>>>>>>>>> integrity issues, but because of the public nature of the >>>>>> broadcast >>>>>>>>>> coupled with strong public identity, the privacy compromise is >>>>>> much >>>>>>>>>> worse. Now transactions are cryptographically tainted. >>>>>>>>>> >>>>>>>>>> This is also the problem with BIP-70 over the web. TLS and other >>>>>>>>>> security precautions aside, an interloper on the communication, >>>>>> desktop, >>>>>>>>>> datacenter, etc., can capture payment requests and strongly >>>>>> correlate >>>>>>>>>> transactions to identities in an automated manner. The payment >>>>>> request >>>>>>>>>> must be kept private between the parties, and that's hard to do. >>>>>>>>> >>>>>>>>> What about using encryption with forward secrecy? Merchant would >>>>>>>>> generate signed request containing public ECDH part, buyer would >>>>>> send >>>>>>>>> back transaction encrypted with ECDH and his public ECDH part. If >>>>>>>>> receiving address/amount is meant to be private, use commit >>>>>> protocol >>>>>>>>> (see ZRTP/RedPhone) and short authentication phrase (which is hard >>>>>> to >>>>>>>>> spoof thanks to commit protocol - see RedPhone)? >>>>>>>> >>>>>>>> Hi Martin, >>>>>>>> >>>>>>>> The problem is that you need to verify the ownership of the public >>>>>> key. >>>>>>>> A MITM can substitute the key. If you don't have verifiable identity >>>>>>>> associated with the public key (PKI/WoT), you need a shared secret >>>>>> (such >>>>>>>> as a secret phrase). But the problem is then establishing that >>>>>> secret >>>>>>>> over a public channel. >>>>>>>> >>>>>>>> You can bootstrap a private session over the untrusted network using >>>>>> a >>>>>>>> trusted public key (PKI/WoT). But the presumption is that you are >>>>>>>> already doing this over the web (using TLS). That process is subject >>>>>> to >>>>>>>> attack at the CA. WoT is not subject to a CA attack, because it's >>>>>>>> decentralized. But it's also not sufficiently deployed for some >>>>>> scenarios. >>>>>>>> >>>>>>>> e >>>>>>>> >>>>> >>>>> >>>> >>