Skype to Users: Microsoft Good, My Bad, We Fixed, All Better Now

Problems with popular P2P VoIP phone purveyor Skype caused problems for millions of users last week. They couldn't connect to the service to phone home, work, or anyplace else. Skype's initial posting at its Heartbeat blog seemed to suggest that the fault was partially Microsoft's, but in its latest posting, Skype puts the blame squarely on itself. Selections from the full posting follow (ellipses [...] indicate omitted sections):

1. Are we blaming Microsoft for what happened?

We don’t blame anyone but ourselves. The Microsoft Update patches were merely a catalyst — a trigger — for a series of events that led to the disruption of Skype, not the root cause of it. And Microsoft has been very helpful and supportive throughout.[...]

2. What was different about this set of Microsoft update patches?

In short – there was nothing different about this set of Microsoft patches. During a joint call soon after problems were detected, Skype and Microsoft engineers went through the list of patches that had been pushed out. We ruled each one out as a possible cause for Skype’s problems.[...]

3. How come previous Microsoft update patches didn’t cause disruption?

That’s because the update patches were not the cause of the disruption.[...]

4. Has the bug been fixed? Should Skype users worry about future Microsoft Update patches and reboots?

Yes, the bug has been squashed. The parameters of the P2P network have been tuned to be smarter about how similar situations should be handled. Once we found the algorithmic fix to ensure continued operation in the face of high numbers of client reboots, the efforts focused squarely on stabilising the P2P core. The fix means that we’ve tuned Skype’s P2P core so that it can cope with simultaneous P2P network load and core size changes similar to those that occurred on August 16. We’d like to reassure our users across the globe that we’ve done everything we need to do to make sure this doesn’t happen again. We’ve already introduced a number of improvements to our software to ensure our users will not be similarly affected – in the unlikely possibility of this combination of events recurring.[...]

So, let's sum up. In short, Skype is saying 'Don't blame Microsoft, blame us. We found the problem and fixed it.'  Read the whole post.

Is That All There Is? A Closer Look at How Skype Works 

Even after Skype's second post, lots of people have been wondering if that's all there is to it. After all, in the past, Skype's had no problems handling lots of login requests triggered by past Windows Update or other reboots, but this time, its self-healing mechanism didn't work properly. What went wrong this time?

According to Skype, there was a "previously unseen fault in the P2P network resource allocation algorithm Skype used. Consequently, the P2P network’s self-healing function didn’t work quickly enough. Skype’s peer-to-peer core was not properly tuned to cope with the load and core size changes that occurred on August 16. The reboots resulting from software patching merely served as a catalyst. This combination of factors created a situation where the self-healing needed outside intervention and assistance by our engineers."

First of all, let's make it clear: Microsoft (and Windows Update) did not cause Skype to fail. Period. So, if Windows Update merely exposed a problem (rather than causing it), exactly what is it about Skype's P2P design that helped the failure to take place? To understand that, it's time to learn about Supernodes.

The Supernode Connection

In section 3 of its latest Heartbeat posting on the problem, Skype refers to problems being caused by its supernodes rebooting, without explaining exactly what a 'supernode' is. So, what exactly is a 'supernode'? As Skype discusses in the universities section of its security resource center, any PC running Skype that is not behind a firewall and is left running for several days at a time can be automatically promoted to supernode status.

Skype's supernodes are used to enable Skype users to find each other through storing directory information. Promoting PCs directly on the Internet to supernodes enables Skype to provide service to computers behind firewalls and routers that use NAT (network address translation). So, if you use Skype and your PC is behind a firewall or connected to a router, be sure to thank those brave Skype users whose PCs are outside those protections.

The flaw discovered by Skype last week affected Skype's ability to recover from having a large number of its supernodes offline because of the automatic rebooting triggered by Windows Update. In other words, if your PC is a supernode, it's part of the infrastructure of the Skype P2P service, not merely a client. For users who've been wondering if Skype is really a distributed network rather than merely a P2P network (scroll down and read the comments), the fact that some clients are also supernodes that assist other clients, and the fact that some clients are merely clients, suggests Skype has some of the characteristics of a distributed network.

Don't Wanna Be a Supernode? You Do Have a Choice

You didn't know your PC could become a supernode? You granted permission for your system to be elevated to supernode status when you accepted the Skype end-user license agreement, which reads (in part):

Utilization of Your Computer. Skype Software may utilize the processor and bandwidth of the computer (or other applicable device) you are utilizing, for the limited purpose of facilitating the communication between Skype Software users.

However, you can opt out of supernode status if you want to. Becoming a supernode has been known to saturate Fast Ethernet networks and to overwhelm low-end routers connected to supernodes, so Skype provides a variety of ways (open the 'How can we prevent our network from hosting supernodes?' section) to disable the option to become a supernode.

Supernodes and Skype - a Few Questions for the Future

Now that Skype users understand more about how the service works, it's going to be interesting to see what happens with the supernodes that make it all work. Will some Skype users opt to disable supernode services or move their systems behind a firewall or NAT router to preserve bandwidth? Will other Skype users set up an "always-on" system that's not protected to do their share to provide supernode service for the network? Skype's software changes are designed to prevent a recurrence of last week's service meltdown, but what happens to Skype if users realize that their systems are part of the infrastructure - and don't like the idea? Stay tuned.

Edited for minor grammar problems.

COMMENTS:

5

TAGS: 

Skype, VOIP, P2P, supernode

Comments Print Email