Slashdot Mirror
More Skype Back Door Speculation
An anonymous reader writes "According to reports, there may be a back door built into Skype, which allows connections to be bugged. The company has declined to expressly deny the allegations. At a meeting with representatives of ISPs and the Austrian regulator on lawful interception of IP based services held on 25th June, high-ranking officials at the Austrian interior ministry revealed that it is not a problem for them to listen in on Skype conversations."
Unless you think it's a good thing that some people can snoop on others conversations, this should be a really good reason to embrace free software.
Swedish plasma phys. PhD student; MSc EE; knows maths, programming, electronics; finance interest; seeks opportunities
All you have to know to monitor someone's Skype is their password. Login with Skype on another machine, set status to invisible. Anything they type or receive in chat you receive.
1. For IM: Jabber (non-US server) + OTR Plugin + Tor. ... and we don't do waterboarding here) (I hope)
2. For everything else (email/vpn/storage) services as provided by www.xerobank.com will do you good.
3. TrueCrypt Full Drive Encryption. (Check your local laws - under Dutch law they cannot force me to give up the passwords
You can be sure that these people are also trying to:
You can be equally certain that they are not doing it right and that the backdoors they are trying to put in make your system less secure.
Running open source software is your best bet, but even there, you aren't completely protected.
Assume all communication that uses any kind of monitorable infrastructure is bugged. The capacity is there, and the desire is there.
It is the way of things.
-- http://frobnosticate.com
This is going to be a problem with any so called "secure" communication system that relies on source secret clients and unpublished protocols.
There are many ways to build such clients to "assist" external intercept, since they often have to first communicate with some central server to locate users. They could for example have a command that forces the client to always route back through the server (like they do for NAT), and use a simple data transformation rather than full encryption so casual packing snooping makes it "appear" encrypted when it is actually not.
They might also have flaws in their implimentation, particularly with key exchange, that allows an invisible man in the middle. The ZRTP stuff developed by Phil Zimmerman that we use in GNU Telephony secure calling uses extra steps to compute a sas to validate there are not fake public session keys given out by a man in the middle, for one example of how such flaws can effect otherwise "secure in appearence" systems.
Of course, even secure peer-reviewed protocols and foss clients do not gaurantee security. For example, one can tether a bunch of ZRTP softphones to an Asterisk server using PBX enrollment, but this enables and requires said server to decrypt all traffic as it passes through, as it acts as a "trusted" man-in-the-middle.
In the end, the best solution, even with ZRTP, remains using pure peer-to-peer (end-to-end) media connections, and when needed transparent proxy media exchange; the latter for dealing with NAT. In ZRTP, sas negotiation assures any such proxy used for NAT "remains" transparent.
In the case of Skype, source secret clients that can report false call information and source secret protocols are a clear recipe for disaster.
I don't think competitive code is as much of a threat as simply knowing what the code does is a threat.
I have read through a good portion of the PDF and I agree that the analysis of the breakdown and all of the measures Skype makes to conceal what it's doing are both impressive and worrisome. I suppose, perhaps, an alternative measure might be for some sort of "computing trustworthiness" scale to be created where the worst offenders (like Skype) are ranked as "potentially dangerous" until they [Skype] clears the matter up.
I suppose in the presence of such a [subjective?] scale, there would be a huge list of programs and applications deemed to be offensive in this way, but perhaps a black list such as this could be useful in attempting to get software a bit more open than it is today? After all, if you could cite an application as "2 out of 10" on the trustworthiness scale as a reason not to purchase, people might begin to take notice. I think a scale like this, whether subjective or not, would enable the technically uninterested to read these 'executive summaries' of information and make better decisions -- making it easier for the public to make more informed choices.
Would lawsuits result? Of course. But the lawsuits against RBLs once happened frequently before people decided it was better to just take measures to stay off the lists. Consumer Reports once found itself at the receiving end of legal actions and demands (and probably still does) but in the end, it's worth the effort they make as they are generally accepted as a trustworthy source. We need a Consumer Reports for software that exposes the privacy and security concerns that different software poses.
I know this stuff about Skype has given me reason to pause, but that's just me... I can sort of read and understand most of what I read here. But how about the rest of the uninformed? How can we get the point across to them?
Two words: Network Effect. All the alternatives I have reviewed are harder than skype. Harder to download, setup, use, the list goes on.
Result: Skype is popular - they nailed delivery to the "masses". No screwing around with the microphone, NAT/firewalls, SIP providers, names etc etc. The average joe can just download and install it in just two url clicks, type in a name and begin to use it. Done deal.
All the open source VOIP (most of them SIP) I have seen completely miss this most important point, and so all their development effort is ultimately wasted - walled themselves off to the technically proficient crowd and not benefiting from the network effect.
Why must EVERY conversation on privacy boil down to a few tired questions about "open source" alternatives ?
Because open source alternatives shouldn't have backdoors. And if it does they can be identified and closed. The only reason the conversation is tiresome is because proprietary software seems to have a perpetual stream of backdoors that keep keep bringing it up.
What, like if the source code is open, then that will prevent backdoors ? Erm hello, the client software isn't the problem, it's the network of Skype servers the bloody data passes through that is the weak point in the equation.
Nobody intelligent is asking for an oss skype client. They are asking for an oss replacement to the entire skype service. For precisely the reason you stated.
So who do you trust more with your privacy ? A multi million dollar company, or some nerd in his moms basement, acting as a VOIP connectivity server.
If that nerd is just hosting as a connection service, and the voip data stream itself is end-to-end encrypted and is actually transmitted directly to the recipient, then I trust the nerd in the basement more, because he never even sees the stream, and even if he did, its encrypted.
At least as long as I know I'm -really- using the public key of the called party to encrypt it, that is. But that is biggest weakness of almost all internet uses of encryption.
In my case, I'd chose option "none of the above", but really ... open source is not the answer to ALL the worlds ills.
Not all of them. But it is the answer to this one.
What keeps me with Skype is that I can have US telephone number. So no matter where I am my friends and family can call me.
If there was another service which allowed me to have a US telephone number for incoming calls and let me call any other POTS number I'd use it.
Nothing in the world is more dangerous than sincere ignorance and conscientious stupidity.
Therefore, if the Chinese have no problem with Skype, Skype must have a back door.