News for hackers. Stuff that matters.
Re:Skype's Principal Architect explains why peer-to-peer was... by crazygringo on Monday June 24, 2013 @12:42AM (#5931027)
I'm curious... I understand that the Microsoft servers are functioning as supernodes now.

But when one computer calls another, and neither are behind NAT/firewall/etc., is the actual audio stream going directly between computers, or are they both sending the actual audio to/from the Microsoft servers?

Obviously Microsoft could choose to redirect it through their servers whenever desired, but I assume that audio communication is still direct, whenever possible?

Re:Skype's Principal Architect explains why peer-to-peer was... by stephengillie on Monday June 24, 2013 @01:45AM (#5931240)
I don't think the scenario you're presenting is realistic. Either an exposed PC on the internet has its own firewall, or it's already exploited.
Re:Skype's Principal Architect explains why peer-to-peer was... by zurn on Monday June 24, 2013 @02:25AM (#5931334)
Nope - centralised network firewalls never became popular for consumer users. Microsoft started shipping a default-on host firewall in XP SP2, those let the user get prompted on what apps to allow. Lately corporations have been moving to host firewalls too since the "crunchy on the outside, soft on the inside" intranet mindset is increasingly incompatible with the modern world.
Re:Skype's Principal Architect explains why peer-to-peer was... by tokenizerrr on Monday June 24, 2013 @01:54AM (#5931260)
Same lan network?
Re:Skype's Principal Architect explains why peer-to-peer was... by stephengillie on Monday June 24, 2013 @02:13AM (#5931306)
Are you skyping people down the hall instead of walking to visit them?

I can see a feasible argument for this when you're talking about routing it through your WAN, but if you're that concerned about internal security then you should roll your own voicechat solution that you can keep on your own self-hosted (non-cloud) internal servers.

Re:Skype's Principal Architect explains why peer-to-peer was... by riquito on Monday June 24, 2013 @02:42AM (#5931370)
It's common practice to use a chat system inside a company, and instead of people shouting or constantly walking to their peers everyone write from their pc instead of visit the others (for most of the little requests).
Re:Skype's Principal Architect explains why peer-to-peer was... by lucaspiller on Monday June 24, 2013 @04:10AM (#5931623)
A lot of companies still give each employee a telephone on their desk too.
Re:Skype's Principal Architect explains why peer-to-peer was... by jaredmcateer on Monday June 24, 2013 @09:10AM (#5932620)
My company did, I unplugged it and put it under my desk when it became apparent 99% of the people calling me were telemarketers and recruiters cold calling the entire directory.
Re:Skype's Principal Architect explains why peer-to-peer was... by pwnna on Monday June 24, 2013 @04:22AM (#5931666)
IRC!
Re:Skype's Principal Architect explains why peer-to-peer was... by tokenizerrr on Tuesday June 25, 2013 @09:56AM (#5939426)
Yup. Web development company with a few floors of employees. Everyone has it installed, and it works fantastic for quick questions or sending someone a big database dump (beacuse it currently goes over LAN).
Re:Skype's Principal Architect explains why peer-to-peer was... by throwawaykf on Monday June 24, 2013 @02:08AM (#5931290)
It really isn't that hard to test. Any of a million packet sniffing or network monitoring tools could be used to trivially settle this.

I cannot understand why none of the alarmist articles about Skype (some even by so-called "cypherpunks") has ever even bothered to run such a simple check.

FWIW, I ran my own simple test using netstat when Skype first centralized the supernodes. I walked away satisfied when I still saw primarily P2P activity during voice/video calls to an IP that belonged to the remote party. (It used UPnP port forwarding to traverse the NAT/firewall, if anybody is curious.) But then, I'm the trusting kind and dug no further. And it's been a couple years since. If there truly are "hackers" on here, I encourage them to try their own experiments and share them.

Re:Skype's Principal Architect explains why peer-to-peer was... by pbhjpbhj on Monday June 24, 2013 @08:12AM (#5932369)
>And it's been a couple years since. //

It was bought by MS in 2011. It's literally the last couple of years that people are concerned about. It's recent changes that have moved from distributed to central routing.

From the OP article (apparently by Skype [past] principal architect):

>"[...] people who were concerned about privacy -- including many of us in the security industry -- used Skype for secure communications.

>Both eBay and Silverlake maintained this architecture, as well as the Luxembourg HQ.

>Since being acquired by Microsoft, however, the service has been re-architected to run through MSFT-owned servers, rendering encryption functionally meaningless and making it just as easy as POTS to monitor.

>None of this -- neither the $7.5B acquisition itself, nor the decision to move to a datacenter model -- make strategic or business sense for Microsoft as far as I can tell." //

Re:Skype's Principal Architect explains why peer-to-peer was... by throwawaykf on Monday June 24, 2013 @03:28PM (#5935100)
Right, I ran my tests shortly after reading that Skype supernodes were getting centralized, which was after the MS acquisition. I just handwaved with "couple of years" because I couldn't remember exactly when.

Back then my conclusion was that actual voice / video traffic was still direct P2P, although it kept connections open to other IPs, probably for signaling/stun/relaying. I haven't had a chance to run the test again because, well, I'm on a mobile device :-)

But my bigger point was that on a forum ostensibly for hackers, there should not be so much speculation and innuendo about something that can be pretty easily tested.

Re:Skype's Principal Architect explains why peer-to-peer was... by Elepsis on Monday June 24, 2013 @01:24PM (#5934331)
This article is written confusingly, but he's actually quoting some other person in that bit of text. He then proceeds to reply.
Re:Skype's Principal Architect explains why peer-to-peer was... by Locke1689 on Monday June 24, 2013 @01:23PM (#5934328)
I believe that's a quote from another article that he seeks to refute. The article doesn't make sense, otherwise.