More on Skype - Let's Talk Money - Let's Talk Big
I played phone tag for a while, but wound up talking to a friend in Canada about a Skype issue that raises a whole different angle I haven't seen anyone talk about yet. Thanks Doug, for sharing some brainstorming with me on this
Doug had read Skype protocol cracked? on ZDNet and felt that they brushed up against an issue that didn't get full attention. I hadn't seen this article yet myself, in part because there have been so many. Here's an excerpt that opens some new thoughts on -
For example, Vozin could use the protocol to build a plug-in that would allow Skype users to make calls to and accept calls from regular phones without buying the PC-to-phone service from Skype.Paglee said the effect on Skype could be devastating.
"It's safe to say that Skype is more popular than any other VoIP provider," he said. "And if this Chinese company could open the protocol and license it to other companies, and developers come up with applications that are better than Skype's, then they can really take revenue away from Skype."
What really catches my attention is a short phrase in there - Vozin could use the protocol to build a plug-in that would allow Skype users to make calls to and accept calls from regular phones without buying the PC-to-phone service from Skype.
Like many folks, yesterday I was thinking about a compatible client, even a unified client that allows for single software to access multiple IM clients like MSN, Yahoo, Skype, etc. We also all explored the idea of multiple VoIP networks and protocls. I wouldn't mind seeing a client that can operate using SIP or Skype protocol that would interoperate with Skye, Gizmo, SightSpeed and any SIP URL or PSTN number. We'd all like to get our hands on that client.I also want to point out that Aswath did his usual stellar job of keeping us all honest in this. His Skype has been cracked! So what? post reminds of the work done by Philippe Biondi and Fabrice Desclaux, which really is noteworthy and not to be overlooked. But tagging on Aswaths reference to the issue on slide 113 of the study that Skype “fully trusts anyone who speaks Skype” I think we have a real problem for Skype. Not a mindshare problem. Not a perception problem, but a financial problem,
If we accept that there's a client in the offing, let's use August as an availability timeframe since that's what was alluded to in all of the chatter yesterday. If that client can successfully and effectively "speak Skype" and be recognized as a Skype client then Doug's point raises the financial scenario.
Somewhere, in some fashion, Skype employs to allow that free callout to all North America right now. Call them supernodes, call them infrastructure, call them whatever you like. I called Doug in Canada using SkypeOut. That means I made a call via Skype and somewhere, they dumped it out onto the PSTN to reach Doug's phone. That's fine with Skype today because I'm a Skype member.
But what might the impact be of an alternate client. Let's call it, for sake of argument, a renegade client. Might the renegade client enable other services and other networks? Sure. Skype requires a user name, but could an alternate client creatively set up a multitude of SkypeOut calls to the PSTN, letting Skype pay the freight? I can think of several scenarios that could cost Skype a lot of money, to the benefit of the "renegade."
Most folks seem to think Skype should open the protocol. Aswath isn't so sure that's a good idea. While I lean toward opening the code, I lean even harder toward Skype embracing SIP. What is clear is that Skype just can't sit on their laurels and feel happy that they have a large user base. Their user base is an interesting demographic, with interesting penetration rates in different parts of the world. But users can be fickle. In an age of software solutions, a user base can vanish over the course of a week. Like birds on the wing, flocks can migrate in droves to new directions.
Skype has some options, but letting it ride and taking no action isn't one of them.
Technorati Tags: VoIP, Skype, Skype hacked, Skype financial impact, PSTN gateways
Email This!
Digg it!
Sphere it!
Del.icio.us
Reddit!
Newsvine
Comments
So this is like your kid getting on your phone and dialing 1-9XX calls that charge your account without your knowing anything about it. You get stuck with the charges on your phone bill.
Posted by: Rick | July 15, 2006 7:36 PM
I'm not sure Rick, but I'm inclined to think that sort of opportunity might exist. Imagine eBay/Skype as the parent and everyone running the renegade client as a kid making those calls. I think there's potential for a big problem.
Posted by: Ken | July 15, 2006 7:51 PM
One thing that will have to be addressed by the Chinese or any other "renegade" company -- the GIPS soundware that is patented and licensed. Skype's attractiveness has always been due to it's sound quality. An open protocol still carries the burden of anyone using a Skype protocol to license the GIPS soundware (or any other sound solution, for that matter).
Google Talk is a perfect example. Google has opened up their entire protocol and they are partnering with GIPS to assist Google's developers. Any developer who creates a client or application based on GTalk's API's has to license GIPS if they want the sound quality of GTalk.
Posted by: Rick | July 15, 2006 7:52 PM
To add to the previous point about sound -- I still don't see how a "renegade" client will be able to talk to Skype with any quality. Skype uses GIPS codecs, so the renegade would also have to use GIPS codecs. Unless, of course they reverse engineered them too, which would be difficult and illegal (to use).
Posted by: Rick | July 15, 2006 7:59 PM
Rick:
We have exchanged this point about codec before, I think. The only codec that Skype uses that one needs to license from GIPS is iSAC. Also, GIPS speech engine is not necessary at both the end points from interworking point of view.
Skype's quality is over rated. After all it has to reflect the vageries of the Internet. So the users may not even notice the theoretical poor quality of not using iSAC or GIPS speech engine.
Posted by: Aswath | July 16, 2006 4:05 AM
I am missing the whole bit of Skype missing out any revenue opportunity from a non-nefarious, but renegade client. After all, Skype is giving away the client for free. This renegade client is using authentic Skype id. So if the call is a SkypeOut call, then the caller must have valid SkypeOut credit.
Posted by: Aswath | July 16, 2006 4:09 AM
Aswath, I agree that Skype quality is overrated (but not bad), especially with better IM's now. Personally, I a use better quality IM. It's due to their pathetic network. They have absolutely no network provisioning. There's only so much GIPS can do with insurmountable latency. However, I hate to imagine what is would sound like without GIPS.
Agree on SkypeOut credits. Good point.
Posted by: Rick | July 16, 2006 8:01 AM
Aswath, I'm sure I'm articulating it poorly, but let me try another spin. You're right in that Skypeout is a legitimate ID. Waht if we think of it more as being analgous to the "escape out sequence" to pull second dial tone in the traditional PBX world. That was always a prime target fr fraudsters.
A single ID could proxy many dial tone requests I assume, although the audit mechanisms at Skype should quickly catch it.
But how far does "Skype trust Skype?" Could a new, phantom user be created by the renegade upon need, then abandoned? I don't know. I'm honestly not that deep into Skype. I just "see" ways Skype's implicit (maybe it's explicit) trust of Skype protocol as problematic.
It's fine to trust a protocol between trusted entities. That's what Skype does today. When an untrusted agent has the trust mechanism, as appears to be at laeast partially the case here, I think they become trusted by default. That's dangerous from a security perspective and can open the door for breaches never even anticipated.
Posted by: Ken | July 16, 2006 9:40 AM
I agree about the loss of trust between Skype clients. I say as much in my post that you refer. But since the clone will not become a supernode, I don't think there is a chance of IDs being compromised. So I feel that what Skypers have to look for is whether there are any renegade supernodes.
Posted by: Aswath | July 16, 2006 11:41 AM