Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
Intel(s intc) has been angling to get its processors into the mobile network for years. Now thanks to partnership with Nokia Siemens Networks(s nok)(s si) it’s finally getting its chance.
At CTIA Wireless Thursday, NSN and Intel said they have begun jointly building a mobile edge computing architecture that would put applications servers at every cell site. The partnership is an extension of NSN’s Liquid Applications strategy, which it unveiled at Mobile World Congress with IBM(s ibm).
The idea is to transform the radio access network from a mere delivery network to one that hosts content and services. Such an architecture would not only put video and content closer to mobile subscribers, but it would make such Liquid Apps much more network-aware apps. Here’s how I described when NSN first partnered with IBM:
Mobile applications and radio infrastructure have always been walled off from one another – applications just barrel ahead onto their radio on-ramps oblivious to the highway traffic conditions ahead. What NSN proposes to do with Liquid Apps is to make those disparate portions of the network work in unison.
For example, mobile video today can be a precarious proposition. As video viewers rack up in a particular cell, the network will keep trying to cram those video streams into the same limited airwaves, The result is a backed-up network with no one getting a quality video stream – or any stream at all. By processing video at the cell site, though, the base station could make decisions how to deliver those individual video feeds based on the prevailing network conditions.
If the cell is congested, then the base station downgrades the video quality of every stream, ensuring everyone sees a decent-quality picture. And as users gradually vacate the cell, the base station could gradually boost video quality for those that remain.
At the heart of Liquid Apps is an application server utilizing Intel’s Crystal Forest platform for network infrastructure and a boatload of Xeon silicon. Those servers would perform the localized processing and content storage as well as maintain a constant collaborative link with the radio base station.
This is some pretty cool technology, but there’s a scary aspect to it, too. Technically the Liquid Apps architecture could be used to optimize all content traversing the network by acting as a traffic cop that dictates the flow of different types of data before they enter the airwaves. But Intel and NSN are also proposing that carriers sell Liquid Apps as a service to developers and content providers.
In such a scenario, a YouTube(s goog) or a Netflix(s nflx) could pay to have their content not just hosted at then cell site but also prioritized as it leaves the tower. It’s a pay to play model that doesn’t exactly sit well with net neutrality principles — content providers that choose not to participate might see their customers’ experience suffer.
U.S. carriers like Verizon(s vz)(s vod) and AT&T(s t) are already talking up two-sided revenue models in which content providers subsidize their customers’ data usage on the mobile internet. Liquid Apps could become another tool in that pay-to-play arsenal.