PGP Creator's Zfone Encrypts VoIP

Philip Zimmermann, creator of PGP wrote in to tell me about Zfone, his new system for encrypting any SIP VoIP voice stream. His first release is Mac & Linux only. I tested it with him using Gizmo as our client and it was pretty trivial to use. While it should work on most any SIP compatible VoIP client, he hopes that clients like OpenWengo and Gizmo will incorporate Zfone directly into the UI. Zfone has no centralization, and has been submitted to the IETF. He hasn't yet determined a license, but he believes strongly in releasing source code for all encryption products. A windows client is forthcoming.

Comments

Important Stuff

[WebHostingGuy]

This is important stuff as more and more phone traffic is routing open in the internet. While most people do not believe their emails are totally private, when it comes to talking on the phone I believe there is a perception (and assumption) that no one else is listening. SIP, Asterisk and all the flavors of VOIP is changing telecom and encryption is necessary.

Re:My only concern

[Aspirator]

I presume that this will be released outside the US, and allowed to migrate in.

Even then I'm sure that there will be attempts by the US Government to
prevent its use.

This is a significant advance in the search for privacy.

It would also..

[TheAxeMaster]

It would also almost totally negate any ISP's attempt at shaping VOIP traffic to try and get people to buy their service instead. This has been somewhat of a question in recent months, but if you can encrypt your stream, then there's not much chance they can slow your packets. I'm all for the increased security as well. Now if we can only get them to cut down on the spam....

Re:My only concern

[Anonymous Coward]

I'd say it's dissapointing that the post had to link to a definition of Luddite in the first place.

What ever happened to PGP Phone?

[WarlockD]

The MIT Website has taken it down, but I remember it working somewhat well between two IP address.

Was it just too far ahead of its time?

Re:My only concern

[lucm]

Many country have much more severe regulations against encryption technologies than the USA. Just a few years ago there was an incredibly severe law in France, stating that even domestic encryption was under the government control.

Believe it or not, you actually live in a (somehow) free country.

Re:Releasing the source

[HolyCrapSCOsux]

Wouldn't that kinda be the point?

Checksums, distro needs sigged.

[Anonymous Coward]

As there is no cryptographic signature on the package, these are my sums
as received. Please compare and post if yours are different.

SHA1 (zfone-linux.tar.gz) = aa9ac66a5dce43cff2639787f30e939078b47ebe
MD5 (zfone-linux.tar.gz) = c6a47feca0fd5cb5bf72a8f6a1e8f207

PRZ, please sign your packages! Thanks, World.

I agree, it is hugely important

[maillemaker]

Hopefully, this will be the straw that breaks the camel's back.

Ultimately, ALL traffic should be encrypted, whether it is VOIP, email, web browsing, whatever.

The guy is right when on his home page he talks about how it is so difficult to implement this sort of stuff as an add-on for emails, managing keys and the like. It's why no one does it. Of course, there has always been a computing overhead, also, which is why only pages that "need" to be secured currently are. But as horsepower goes up, those limitations should go away.

Ultimately, it should be a matter of course before all traffic that goes in our out of your computer is encrypted by default.

Hopefully this is the start of something huge!

Steve

Re:I agree, it is hugely important

[utlemming]

I tend to agree that with CPU's increasing in speed and power, that the cost of encyrpting goes down. But also the cost of breaking the same codes goes down. Right now it takes 2^120 rounds to break AES encrpytion while trying to brute force. So in order for the higher speeds of the CPU argument to hold to encrypt everything, encryption technology will have to go up. Right now 128 bit is acceptable, but as speeds increase, then the encrpytion will have to be changed as well. And that will require faster processors.

Re:I agree, it is hugely important

[Anonymous Coward]

Ultimately, ALL traffic should be encrypted, whether it is VOIP, email, web browsing, whatever.

No, it shouldn't. Just because encryption is a very useful hammer does not mean that everything is a nail.

Thanks to Moore's Law, encryption is cheap -- but it's still not free. That's OK for things like E-mail, where the two end-systems handle only a handful of messages at a time. But if the Web suddenly switched from HTTP to HTTPS overnight, Web servers would collapse left and right from having to juggle thousands of simultaneous encrypted connections. Plus, the overhead for setting up an encrypted link is much higher than an unencrypted link, thanks to things like session key negotiation; for short-lived sessions like most HTTP traffic, the cost of setting up the link would dwarf the cost of actually transmitting the data.

There are a lot of networking applications where the contents of the packets is just plain not worth protecting from sniffers. Things like multiplayer games and most Web surfing usually fall under that category. There's no reason to force the added overhead on those applications just because we can.

Re:I agree, it is hugely important

[iabervon]

Encrypting everything doesn't help security at all. It only helps if you have some idea of who the intended recipient is. For most traffic, that's an application-specific concern, and so a general encryption mechanism is useless. For example, if I want to send email to a gmail user, and I care about it being private, I have to find a private key for the user. It doesn't help to just encrypt (great, nobody at all can read it), or encrypt it for secure transit to Google (why should I trust Google more than other random intermediaries?). I need the actual user's PGP key, and that requires an application specific identity specification (i.e., the email address), with a user-specific way of determining that I've got the right address for the communication.

Re:Why not encrypt by default

[Noksagt]

• Technological
  
  • Encryption implementation isn't free.
    
  • Encryption maintenance isn't free.
    
  • Unencrypted traffic is easier to sniff (which may be legitimately important).
    
  • Encrypyed traffic has a higher CPU overhead (which isn't always made up for).
    
  • Some people prefer to have one really good encryption program (SSH or a VPN) to route all traffic over.
    
  
  
• Legal
  
  • Encryption can't always be exported from every country to every other country.
    
  • Sometimes it may be illegal to encrypt traffic.
    
  
  

Re:Phil and Jon

[merreborn]

I'm sorry, did I miss the story about DVD John breaking the public key encryption model? And blowfish? And the cypher du-jour?

He's released cracks for various pieces of software, but it's not like the guy's actually broken actual strong encryption algos.

http://en.wikipedia.org/wiki/Jon_Johansen#Other_pr ojects

Re:Why not encrypt by default

[Anonymous Coward]

If you want that encrypted you're told to send it over SSH.

I think you've answered your own question. Unix mantra: do one thing and do it well. SSH is the encrypted transport layer. No other communication layer needs encryption.

This is exactly why tar doesn't compress, and gzip doesn't make archives. If you want that tar archive compressed you're told to gzip it.

Re:Encryption is hard

[Bogtha]

Wanna encrypt VNC?

ssh -L 5900:localhost:5900 somehost.org
vnc to localhost:5900

Was that hard? No.

That's because you skipped the hardest step - the out-of-band communication necessary to establish key authenticity. The first time you connect to a host via ssh, it displays a fingerprint and asks you if it should be trusted. Most people don't value encryption enough to even bother installing the relevant software, what makes you think that they value encryption enough to figure out a way of validating the fingerprint securely?

Why not use OpenSSL?

[Cthefuture]

I know the API isn't the greatest and the documentation completely sucks but someone with OpenSSL knowledge could put together a SIP-friendly API in a couple hours.

At least then you're using a public, well hammered on API and would have a multitude of algorithms to choose from. I mean OpenSSL is used in tons of stuff and gets lots of field testing.

I have never understood the point of PGP with its proprietary crap formats when there are open, standardized formats for the stuff it is typically used for (S/MIME, X509, PKCS#12, etc.) and that are supported in standard applications rather than require some goofy PGP add-on.

Re:traffic shaping reality check

[wolrahnaes]

The mention of 911 gives me an idea for an interesting angle to ensure ISPs can't neuter VoIP.....claim that by doing so they're endangering lives in the event of a 911 call.

Re:Phil and Jon

[Jah-Wren Ryel]

He's released cracks for various pieces of software, but it's not like the guy's actually broken actual strong encryption algos.

And such 'cracks' are the best way to attack otherwise strong crypto-systems - don't try to crack an algorithm -- crack the implementation. Look for the vulnerabilities in the systems that use strong cryptography and find the back-doors, or break in a hole in the wall, but trying to go mano-a-mano with the entire crypto community isn't a smart thing and you are exactly right -- that isn't what DVD-Jon has ever done.

Not that I would imply that CSS is a strong algorithm, it ain't. But the new stuff for BLU-HD-RAY uses AES and the stuff that the Zim-man is using to security VOIP also uses tried and true crypto algorithms. That doesn't mean there won't be flaws in the implementations that can be exploited and Jon-Jon He's Our Man, If He Can't Exploit It, No One Can!!! Yeah Jon. Or something like that.

Re:Why not encrypt by default

[Matt Perry]

I think you've answered your own question. Unix mantra: do one thing and do it well. SSH is the encrypted transport layer. No other communication layer needs encryption.

So we should manually establish an SSH connection from our mail server to every MTA that we are communicating with? We should also establish a manual SSH connection for IM connections to every IM server that we connect with? SSH isn't a panacea.

What is the big deal?

[Guspaz]

Why do some people jump all over protocol-specific encryption as helping terrorists, or other such nonsense?

There is a great deal of concern over Skype being encrypted. People say it can be used by terrorists for encrypted communication. The thing is, throw up a VoIP server of some kind (Even the free ones like Ventrilo or TeamSpeak), and connect to it using something like Hamachi. Bam! All your UDP voice traffic is encrypted.

Heck, you can even do it with TCP. SSH tunnels encrypting two-way Shoutcast streams. Huzzah! Encrypted two-way voice communication! Heck, pump the shoutcast stream over HTTPS and that'd be encrypted too.

So, this is why I don't get it. Why complain about Skype when there will always be ways to encrypt voice traffic over the internet? Programs like Hamachi (Encrypted P2P VPN solution with an IM-like interface) make it insanely easy to set up more secure solutions than Skype, and there is always SSH tunnels as a fallback.

So how does this relate to the current situation? Well, people are sure to complain that this new program somehow helps terrorists. So I'm just saying that that is BS.

I've got an easy answer...

[ImaLamer]

Because it isn't always needed.

VNC? What if I'm only using it over a Cat-5 cable on a private network? Who am I encrypting it from?

You've always got FreeNet. But you aren't using it are you?