#MWC15 I just arrived in Barcelona for Mobile World Congress, the premiere mobile industry event. This year, they expect up to 100,000 exhibitors and customers to attend the event, up from over 70,000 last year. I took three flights taking over 18 hours to get here. My flights were packed with other attendees all trying to get here like me. I had a significant delay at one point because they had to change the actual plane we were supposed to be flying on. There were long lines and it took quite a while to get through some of the lines for the plane, passport control, and getting my show badge.
Every time, the system was not designed to handle such a large influx of people going through the system, and every time, I was delayed getting the expected service and reaching my final destination. This is much like our need for TCP optimization technologies on the Internet. Service providers have to manage congested networks due to seen (large scheduled events) and unforeseen (natural disasters and malfunctions) conditions. During these situations, they need to find ways to still deliver a quality service to their customers. Calls still need to be made. Application still need to update efficiently and in a timely manner.
Currently, many service providers use technologies like video optimization and caching to maintain a high quality of experience for their customers. Unfortunately, these technologies may not be ideal as the Internet evolves. Traffic is growing at a high rate with the general availability of 4G LTE networks and 5G is around the corner. In addition, encrypted traffic is on the rise, increasing over four-fold in Europe in the past year because of security and privacy concerns. These solutions must see the content to be effective and the encryption prevents their use.
TCP optimization, on the other hand, leverages the TCP protocol and does not depend on applications or content. It is designed to improve the flow of traffic through adjustments to the TCP protocol parameters based on expected and observed network conditions. This means that the flow from an application on the Internet to the subscriber can be optimized on one side for the low latency high speed characteristics of the Internet and given a different set of parameters based on high latency and slower access networks like cellular radios.
The TCP optimization technology manages the optimal delivery of the content by acting as a TCP proxy to handle both sides of the connection separately. If I were able to somehow apply this technology to my trip to Barcelona and manage the flow of 100,000 people through the week-long event, it would feel like I got 120% increased efficiency (based on real improvements to content delivery over live wireless networks) out of my efforts surrounding this show.