The best way to guarantee a quality video experience over the Internet is to stream a movie or show from as close as possible to the customer. This works well in broadband and WIFI delivery, but achieving proximity in mobile networks is no easy matter. Qwilt aims to change that, with the release of its Mobile Video Fabric.
In my review of the new Ooyala video data yesterday, it is clear that mobile video consumption is becoming increasingly important. Ericsson foresees even bigger growth to come. Today, the company says 40% of the 1.1 Exabytes of monthly smartphone data traffic is due to video. It expects this to balloon to 50% of 12 Exabytes per month by 2019.
Techniques for optimizing the delivery of video in a physical IP network, though complicated, are being implemented and deployed today to good effect. Unfortunately, mobile networks pose a whole set of new challenges for video delivery that are proving even more difficult to conquer.
In video streaming, network latency is critical in determining the quantity of video that can be delivered (this is known as the video throughput.) Latency is simply the delay between a request being made and it being received by the intended recipient. This is important because the gap between the request being made and received is, in effect, idle time. Shorten the latency and you can increase the amount of video delivered (increase the throughput.)
When delivering video from the Internet on a mobile network, most of the latency occurs in the Internet, before it enters the mobile network. Qwilt’s Mobile Video Fabric (MVF) solution allows videos requested from sites like Netflix, Hulu and YouTube to be cached just inside the mobile network’s core. This caching occurs in a transparent fashion, without video requestor or provider knowing it is happening. According to Dan Sahar, VP and Co-Founder of Qwilt, by caching this video in the mobile core, throughput can be increased by as much as 35%.
The next challenge is to cache the video as close to the mobile device originating the video request. That means putting caches at the cell towers. Doing this allows the video cache to adjust to a particular cell tower radio network (RAN) situation. For example, cell towers Levi’s Stadium in Santa Clara may all be receiving a lot of requests for highlights of an in-progress Niners game while those near CenturyLink Field in Seattle are dealing with Seahawks requests.
According to Mr. Sahar, Qwilt will deliver a software solution as part of MVF next year that will enable video caching at the cell tower level. The software will be able to run on existing equipment or using off-shelf-hardware.
To cope with the massive increase in video envisioned by Ericsson increasing bandwidth with new technologies like LTE is going to be critical. However, given the finite nature and high cost of wireless capacity, using existing bandwidth more efficiently is critical as well. Video caching could well be an important part of the overall network solution, helping operators to control network capital and operational costs.
Why it matters
Video is driving a rapid expansion in the use of mobile networks.
To cope with the massive increase in demand for bandwidth, operators will need to expand capacity with new technologies like LTE.
However, using existing bandwidth more efficiently by leveraging new technologies like video caching is going to be required as well.