The Voice of the Future

Microsoft’s acquisition of Skype is the tech story of the day. Reactions to the still-unofficial deal are all over the place, ranging from suspicions that MS overpaid, speculations about where MS will take Skype, and what it means for mobile networks. While there are interesting issues regarding the integration of Skype with MS’ existing products, I think the payoff will come when Skype is integrated into Windows Phone, the mobile OS.

The transition of mobile networks from legacy technology to LTE is often underrated in terms of overall impact, as the common perception is that LTE is just 3G on steroids, a somewhat faster network that can support a few more smartphones thanks to increased spectral efficiency. This is the wrong way to look at LTE.

LTE is not simply a faster radio for smartphones, it’s a radical revision in the structure and architecture of mobile networks that integrates them with the Internet by extending IP everywhere. This not only alters the mobile network, it potentially alters the Internet in very significant ways as well. As long as the Internet was a system for Web surfing with a few random other applications along for the ride, it was fine to continue extending peering and transit agreements without much regard for traffic differentiation; “all packets are equal” makes sense for a network that’s dominated by a single application. Similarly, the protocols and standards that define the Internet don’t have to be very sophisticated with respect to Quality of Service if everyone is doing the same thing.

Mobile networks alter the application mix by introducing a trio of high revenue applications into the mix: The mobile Internet starts with the Web, but it adds voice, urgent text messaging, and specialized video streaming into the mix. When the mobile network is IP-based, IP has to handle these differentiated apps as well as it handles the Web, or better.

LTE is still up in the air about how to handle voice and video. The premise has been that IMS, a complex system for session control and charging that supplements IP, would handle high-value applications, but it’s not progressing as fast as people w0uld like. Some mobile operators, such as Hutchison, have broken from the pack and offered a mobile version of Skype to the their networks with somewhat mixed results.

Skype is best understood as another entrant into the mix of possible ways to handle voice (and video) over LTE networks, so the interesting questions about the Skype acquisition relate to the integration of Skype with Windows Phone. Microsoft has access to a large installed base of smartphones through its deal with Nokia, the world’s largest supplier of smartphones and a player that’s not going to depart the market quietly.

Operators are not Skype-friendly, mainly because Skype is neither network-friendly nor reliable. Analysts stress the lost revenue issues operators face with Skype, but that’s a short-sighted view because Skype’s revenues depend on PSTN integration so the revenue problem goes both ways.

The problems that Skype has on mobile IP networks are fixable with the right level of integration and in-network support for calling, mobility, and Quality of Service. The solutions will depend on operators providing facilities that are tailored for Skype, which couldn’t happen as long as Skype was seen as an enemy of the operators depending on net neutrality to make it work well. There’s no solution to be had without closer cooperation between Skype and the networks, and that sort of thing is anathema to net neutrality.

Skype over LTE can work fine as long as the networks can recognize Skype’s transport requirements and deal with them appropriately. This has billing implications as well as technical ones, and the new arrangements are best worked out by a firm like Microsoft with a vested interest in working with the operators and the handset vendors to put the tweaks in  the networks to make it work well. The inherited “end-to-end” model doesn’t have the answer to this problem, and the IMS doesn’t have the necessary traction, in part because it’s a bit over-designed. There are similar problems with video over LTE, since it requires in-network systems for transcoding at the boundary of the wireless and wired infrastructure to work well.

Microsoft could easily fail to integrate Skype correctly into Windows Phone, but I think they’ll succeed. MS has some bright wireless engineers who can patiently explain the network enhancements they need to do voice and video to the marketing and policy people who want neutrality. This is going to be very interesting.

As Dan York says, this is the end of the Skype as Bandit era:

… you are no longer fighting “against THE MAN”… you now are “THE MAN”! It’s hard to get much bigger of a megacorp than Microsoft!

Skype went as far as they could with that notion, now that they’ve joined the establishment, they can go a bit further.