[ANNOUNCE] ParrotTalk release/design change considerations

On Thu, Oct 26, 2017 at 12:00 AM, henry <henry@...> wrote:

I disabled vatId authorization in version 3, located in SessionOperations>> processIWant | SessionOperations>>processIAm: .

I was asked to disseminate my news to Pharo Users, hello there. I was asked to describe ParrotTalk well, provide use cases and adopters of its use. Alright, I will give an attempt.

ParrotTalk is an encrypted connection framework. Currently allowing anonymous 2048-bit key negotiation to establish user-provided encryption cipher and user-provided encoding and decoding, both through a provided SessionAgentMap to a starting SessionAgent server. Please look in the test case ThunkHelloWorldTest for building these maps and running a connection iwth data passing after encryption is established. There is a 4-way negotiation, from ProtocolOffered/Accepted to Go/GoToo. In using RSA 2048 signature validation and DH 2048 primes to establish the key used within the selected Cipher. The Cipher and Encoder are selected by name through the negotiation protocol. Currently three Ciphers are selectable, AESede, DESede, and DES. There are two encoders tested, asn1der, and Bytes. This protocol is described here, in this document.

https://github.com/ZiroZimbarra/callistohouse/blob/master/docs/ParrotTalkFrameDesign-3.5.pdf

For as to use cases, this encrypted connection has no third party, man-in-the-middle situation by not using Certificates. As such, this is a tight implementation of NSA-proof encryption without explicit authorization beyond knowledge of a host:port. The use cases involve any communication desired to be encrypted with such high encryption. The support will last my lifetime, so we have a settled solution, here in the third version, provided here. It requires version 111 of Cryptography, as a prerequisite. Both run on Squeak and Pharo.

http://www.squeaksource.com/Cryptography/Cryptography-zzz.111.mcz

http://www.squeaksource.com/Cryptography/ParrotTalk-HenryHouse.3.mcz

The current use is with my hubbub system, a promise-based distributed object implementation. I am working to bring ParrotTalk to Java and allow hubbub to operate interdependently between Squeak, Pharo, Java and any other languages which can support ParrotTalk and STON. My latest efforts with hubbub are to bring STON as the Layer 6 encoding. Hubbub depends on eLinda.

Took me a while to track down what eLinda was.  I presume its an implementation of Linda "a model of coordination and communication among several parallel processes operating upon objects stored in and retrieved from shared, virtual, associative memory" 
https://en.wikipedia.org/wiki/Linda_(coordination_language)

http://www.cs.yale.edu/publications/techreports/tr984.pdf

cheers -ben 

On Fri, Nov 3, 2017 at 4:06 AM, henry <[hidden email]> wrote:

I have a question about my derivation of a MAC key. It is very complicated and I am considering ways to simplify for use by potential language implementations. Here is the current derivation in Squeak/Pharo. Should I just use the sharedKey as the SHA1HMAC key or some simplified hash of the sharedKey, without all the pad hashings I am doing? What do you think? I want to hear from you.

Hi Henry,

I'd guess not many people here are *really* competent enough in crypto to confidently advise you.  I know I'm not.  I only know enough to know crypto is *hard*.

* https://www.happybearsoftware.com/you-are-dangerously-bad-at-cryptography.html

* https://www.schneier.com/blog/archives/2011/04/schneiers_law.html

* https://www.schneier.com/essays/archives/1997/01/why_cryptography_is.html

I hope that doesn't discourage you, and I'm interested to hear a more useful reply from anyone else. 

cheers -ben

P.S. I just bumped into what looks like an interesting intellectual exercise...

http://cryptopals.com/

 

makeHMAC

            | sharedKey hashPadder macKey |

           sharedKey := diffieHellman sharedKeyPadPositiveByteArray.

           hashPadder := self class.

           macKey := MD5 hashMessage: (

                      (hashPad der hash: sharedKey pad: 16rCC),

                      (hashPad der hash: sharedKey pad: 16rBB),

                      (hashPad der hash: sharedKey pad: 16rAA),

                      (hashPad der hash: sharedKey pad: 16r99)).

           macKey := macKey, (MD5 hashMessage: (

                      (hashPadder hash: sharedKey pad: 16r88),

                      (hashPad der hash: sharedKey pad: 16r77),

                      (hashPad der hash: sharedKey pad: 16r66),

                      (hashPad der hash: sharedKey pad: 16r55))).

           macKey := macKey, (MD5 hashMessage: (

                      (hashPad der hash: sharedKey pad: 16r44),

                      (hashPad der hash: sharedKey pad: 16r33),

                      (hashPad der hash: sharedKey pad: 16r22),

                      (hashPad der hash: sharedKey pad: 16r11))).

           ^  SHA1 new hmac key: macKey.

Would a better way forwards is simply this?

makeHMAC

           macKey := diffieHellman sharedKeyPadPositiveByteArray.

           ^  SHA1 new hmac key: macKey.

Thank you.

- HH

-------- Original Message --------

Subject: Re: [Pharo-users] [Pharo-dev] [ANNOUNCE] ParrotTalk release/design change considerations

Local Time: November 2, 2017 3:50 PM

UTC Time: November 2, 2017 7:50 PM

From: [hidden email]

To: henry <[hidden email]>

Squeak-dev <[hidden email]>, Pharo-users list <[hidden email]>, Pharo Development List <[hidden email]>

I have developed the ParrotTalk Protocol, documented in part here[1], while I have two implementations: 1 in Squeak/Pharo [2 a/b] and the other in Java [3 a/b]. Particulars of some decisions are absent from the specification: namely the MAC key and ivSequence derivations, as well as constrained traffic signing; they are in the implementations. I will update the 3.6 slideshow to specify these items.

Thank you.

- HH

[1] - ParrotTalk Protocol : http://jmp.sh/OqlYpyg

Squeak/Pharo implementation

[2 a] - http://www.squeaksource.com/Cryptography/Cryptography-HenryHouse.113.mcz

[2 b] - http://www.squeaksource.com/Cryptography/ParrotTalk-HenryHouse.9.mcz

Java implementation

[3 a] - https://github.com/ZiroZimbarra/ASN1

[3 b] - https://github.com/ZiroZimbarra/ParrotTalk

- HH

-------- Original Message --------

Subject: Re: [Pharo-users] [Pharo-dev] [ANNOUNCE] ParrotTalk release/design change considerations

Local Time: November 2, 2017 3:43 PM

UTC Time: November 2, 2017 7:43 PM

From: [hidden email]

To: Squeak-dev <[hidden email]>

Pharo-users list <[hidden email]>, Pharo Development List <[hidden email]>

More celebration requested. I got encrypted data passing between Squeak/Pharo and Java: AES with CBC and IVs and SHA1 HMAC. It works!  NSA-proof crypto rocking Squeak, Pharo & Java!

*Celebrates again, drinks for everyone*

- HH

-------- Original Message --------

Subject: Re: [Pharo-users] [Pharo-dev] [ANNOUNCE] ParrotTalk release/design change considerations

Local Time: October 31, 2017 1:54 PM

UTC Time: October 31, 2017 5:54 PM

From: [hidden email]

To: henry <[hidden email]>

Ben Coman <[hidden email]>, Squeak-dev <[hidden email]>, Pharo-users list <[hidden email]>, Pharo Development List <[hidden email]>

I wanted to share my good news, though I have not yet gotten Pharo, Squeak and Java talking. What I was able to accomplish, after finishing the port of Cryptography's ASN1Types, Module and Streams is to get Java's rendezvous and key exchange working. After I got that connecting, I looked at the ciphers, ivSequences and MAC and got that working. In Java, I am now able to encrypt and decrypt "hello world".

*Celebrates*

- HH

-------- Original Message --------

Subject: Re: [Pharo-users] [Pharo-dev] [ANNOUNCE] ParrotTalk release/design change considerations

Local Time: October 25, 2017 12:00 PM

UTC Time: October 25, 2017 4:00 PM

From: [hidden email]

To: Ben Coman <[hidden email]>, Squeak-dev <[hidden email]>, Pharo-users list <[hidden email]>

Pharo Development List <[hidden email]>

I disabled vatId authorization in version 3, located in SessionOperations>>processIWan t | SessionOperations>>processIAm: .

I was asked to disseminate my news to Pharo Users, hello there. I was asked to describe ParrotTalk well, provide use cases and adopters of its use. Alright, I will give an attempt.

ParrotTalk is an encrypted connection framework. Currently allowing anonymous 2048-bit key negotiation to establish user-provided encryption cipher and user-provided encoding and decoding, both through a provided SessionAgentMap to a starting SessionAgent server. Please look in the test case ThunkHelloWorldTest for building these maps and running a connection iwth data passing after encryption is established. There is a 4-way negotiation, from ProtocolOffered/Accepted to Go/GoToo. In using RSA 2048 signature validation and DH 2048 primes to establish the key used within the selected Cipher. The Cipher and Encoder are selected by name through the negotiation protocol. Currently three Ciphers are selectable, AESede, DESede, and DES. There are two encoders tested, asn1der, and Bytes. This protocol is described here, in this document.

https://github.com/ZiroZimbarra/callistohouse/blob/master/docs/ParrotTalkFrameDesign-3.5.pdf

For as to use cases, this encrypted connection has no third party, man-in-the-middle situation by not using Certificates. As such, this is a tight implementation of NSA-proof encryption without explicit authorization beyond knowledge of a host:port. The use cases involve any communication desired to be encrypted with such high encryption. The support will last my lifetime, so we have a settled solution, here in the third version, provided here. It requires version 111 of Cryptography, as a prerequisite. Both run on Squeak and Pharo.

http://www.squeaksource.com/Cryptography/Cryptography-zzz.111.mcz

http://www.squeaksource.com/Cryptography/ParrotTalk-HenryHouse.3.mcz

The current use is with my hubbub system, a promise-based distributed object implementation. I am working to bring ParrotTalk to Java and allow hubbub to operate interdependently between Squeak, Pharo, Java and any other languages which can support ParrotTalk and STON. My latest efforts with hubbub are to bring STON as the Layer 6 encoding. Hubbub depends on eLinda.

http://www.squeaksource.com/Cryptography/elinda-HenryHouse.14.mcz

http://www.squeaksource.com/Oceanside/hubbub-HenryHouse.38.mcz

Hubbub currently fails to load correctly in Pharo. In Squeak the two ParrotRemoteTestCases: LookupTestCase and OperationalTestCase really screw everything up, do not save an image after running. I never run these tests until I get a better handle of serialization. Currently, do to a shift in progress to STON, the serialization tests fail, just the way we like it. Fix the tests then the system starts to work right.

The shift away from vatId authorization should not adversely affect hubbub, we hope! When STON starts to work, and I think I need to turn off jsonMode, then traffic between vats can start to be analyzed and figure out how resolving is working with redirectors invoked remotely.

- HH

-------- Original Message --------

Subject: Re: [Pharo-dev] [ANNOUNCE] ParrotTalk release/design change considerations

Local Time: October 25, 2017 12:43 AM

UTC Time: October 25, 2017 4:43 AM

From: [hidden email]

To: henry <[hidden email]>, Pharo Development List <[hidden email]>

Stephane Ducasse <[hidden email]>

If I google... parrottalk protocol

nothing seems related.  So what is ParrotTalk? 

P.S... since this is not "part of" pharo, [pharo-users] would be a better place for discussion than [pharo-dev].

Could you re-announce there, with some description of use cases and who/where the protocol is used?

cheers -ben

On Wed, Oct 25, 2017 at 1:42 AM, henry <[hidden email]> wrote:

Please excuse all the low-level detail, if inappropriate to the discussion. The protocol changes are here specified.

- HH

On Tue, Oct 24, 2017 at 13:40, henry <[hidden email]> wrote:

In order to bring ParrotTalk-3.6 support, alongside historical 3.5 support, the frame header stays the same and the processing of the ProtocolOffered would select "ParrotTalk-3.6" to use the compact protocol of 

- ProtocolOffered { offered, preferred }

- ProtocolAccepted { accepted }

- IWant|GiveInfo|ReplyInfo { v atId, domain, publicKey, cryptoProtocols, dataEncoders }

- IAm|ReplyInfo { vatId, domain, publicKey, cryptoProtocol, dataEncoder, dhParam }

- Go { cryptoProtocol, dataEncoder, dhParam, signature }

- GoToo { signature }

Using eLinda :

http://www.squeaksource.com/Cryptography/elinda-HenryHouse.14.mcz

Above the FrameBuffer, below the SessionOperations for each version of the protocol, an ELindaSession could be inserted. This session think would have the stacks for each protocol version SessionOperation registered by protocol in a Tuple, for eventual callback. When the ProtocolOffered comes in we publish the tuple frame, with set selected version, to route to the correct SessionOperation to support each version of the protocol.

With a non-specific vatId required, anonymous connections would be supported.

- HH

On Tue, Oct 24, 2017 at 12:52, Stephane Ducasse <[hidden email]> wrote:

Hi henry thanks for this announce. Can you tell where such protocols are used? Stef On Tue, Oct 24, 2017 at 6:33 PM, henry wrote: > Hi all, > > I am happy to announce the release of version 3.5 of ParrotTalk, for Squeak > and Pharo, found here: > > http://www.squeaksource.com/Cryptography/ParrotTalk-zzz.2.mcz > > It follows this specification: > https://github.com/ZiroZimbarra/callistohouse/blob/master/docs/ParrotTalkFrameDesign-3.5.pdf > > One item of note, in version 3.5, the system connecting to a server, sending > the IWant msg, must know the vatId of the system being connected to. I am > considering changing this to version 3.6 by removing one round-trip in > messaging. Therefore, these messages would be combined: IWant/GiveInfo, > IAm/ReplyInfo. I will keep ProtocolOffered and ProtocolAccepted to allow > eLindaSession to support both versions: 3.5 and 3.6. > > Thoughts please? > > - HH

Hi Henry.

I decided to ask this questions here.

What the advantage to use your security solution instead of legacy secure sockets which is available in the image? (implemented with ZdcSecureSocketStream)

2017-10-24 18:33 GMT+02:00 henry <[hidden email]>:

Hi all,

I am happy to announce the release of version 3.5 of ParrotTalk, for Squeak and Pharo, found here:

http://www.squeaksource.com/Cryptography/ParrotTalk-zzz.2.mcz

It follows this specification:

https://github.com/ZiroZimbarra/callistohouse/blob/master/docs/ParrotTalkFrameDesign-3.5.pdf

One item of note, in version 3.5, the system connecting to a server, sending the IWant msg, must know the vatId of the system being connected to. I am considering changing this to version 3.6 by removing one round-trip in messaging. Therefore, these messages would be combined: IWant/GiveInfo, IAm/ReplyInfo. I will keep ProtocolOffered and ProtocolAccepted to allow eLindaSession to support both versions: 3.5 and 3.6.

Thoughts please?

- HH