Last post, I covered the arguments behind why a P2P content distribution method using WiFi would be essential to supplement LTE networks 2014 onward. In this post, I’ll address the characteristics of components of such a system as well as try to give some examples for interesting software I have found that can be the foundation of such a system.
Before we delve into the how, let’s take a quick look at the distribution of video traffic over mobile networks. For this I am going to turn back to Cisco’s Visual Network Index for mobile traffic. Following is what I compiled based on data in Appendix A of Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2009-2014 (Feb. 2010).
This figure shows that even by 2014 when real-time video communications starting to become main-stream, almost 90% of video traffic will continue to be content streamed from limited number of content repositories delivered to smartphones, handsets and more importantly PDAs, laptops, netbooks using 3G/4G networks as transport. In the white paper Approaching the Zettabyte Era, Cisco has outlined a number of challenges video traffic will put on the Internet in general. Taking a corollary of their observations and looking at the specifics of 3G/4G wireless networks as the last mile, there are couple of methods to overcome so called last mile bottlenecks:
- Push content as download (closer to) the user so that it is available on-demand, as opposed to the traditional streaming method. A good example of this is BBC’s iPlayer that provides both downloading and streaming options.
- Deliver content via alternative, i.e., wired access networks as close as possible to the user to prevent flash-crowd impacts.
In both cases, we can argue that common solution is to have personal(ized) content storage facilities that are accessible to mobile users via WiFi. In such a system, there will be four major components:
- Terminal client that discovers personal(ized) storage, retrieves and donates content (for P2P delivery)
- Personal(ized) storage at home, at office, in public venues, as well as at the nearest suitable Internet POP
- Content distribution controller that tracks the meta-data, manages digital rights
- Operations and Administration system that controls the other three components
I have started to notice good examples of solutions that can be used as parts of such a content distribution system. Here are few that I can list:
- AirVideo by AirVideo: It allows you to convert and stream your videos on an iPhone or iPad over WiFi from a AirVideo server running on your home computer.
- iMediaShare by Bianor: It is not exactly what I was looking for but iMediaShare combines a personalized cloud storage with a DLNA/uPnP server on your iPhone or Android. This allows the user to play the content at any home entertainment device that is uPnP compatible. It can be adopted such that iMediaShare can be run on your home PC or alternatively home NAS device.
- Twonky by PacketVideo: It provides the home DLNA/uPnP server along with a content manager. It allows creating playlists of personal content as well as online resources.
Looking at the status of available tools, I see there are still missing building blocks. Probably the most essential one is an intelligent client software on the user terminal that can actively discover available content over WiFi and retrieves it from that local source as opposed to a centralized location. A 3GPP WiFi offloading mechanism, Local IP Access (LIPA) is a potential optimization technique that can be used for this purpose.
So far we haven’t discussed the most fundamental question: who will pay for such a system? Even though this type of optimization would benefit everyone; end-user with reduction in her monthly wireless broadband bill (assuming it has a cap), operator with reduction in capacity requirements for low value/pipe service, content-owner with increase in customer satisfaction for WiFi speed content streaming; I can anticipate a lot of struggle to build a solid deployment/pricing model. However, one thing is certain. This would put a great fear in the hearts of companies such as Akamai and Limelight. Following is an excerpt from Akamai CEO Paul Sagan at his company’s 1Q2010 earnings’s call:
There are certainly networks that get underprovisioned. We have seen that to some extent in some wireless networks, and the end-users with a disappointing experience. I think the challenge there is those providers either have to expand their capacity, or they are going to lose subscribers somewhere else. So you are right, we can’t do better than the last mile in the ISP networks.
I think Akamai is mistaken, there are ways to do better. On this blog and on our website, we will introduce them and provide a beta to prove its viability. Stay tuned.