Saturday, March 28, 2015

Mobile Traffic Gets Asymmetrical

Oddly enough, for networks designed for symmetrical traffic, Internet traffic now drives bandwidth demand on mobile networks, and that traffic is highly asymmetrical, one reason access to Wi-Fi and other non-traditional networks has become so important.

Telefónica O2’s customers, as an example, are using 60 percent more data than they were 12
months ago, and 600 percent more than at the end of 2010, according to Real Wireless.

The uplink:downlink ratio, on a mobile network supporting 3G or 4G, is now about 1:7, and much of the downstream traffic is bandwidth-intensive video.

Eventually, most believe, new Internet of Things sensor traffic will add other types of load, namely demand for smaller messages of lower bandwidth, but requiring low power performance and high reliability.

LTE was not optimized for this type of usage, so Wi-Fi and other specialized connections
will likely be important.

That is why spectrum sharing, and the move towards dynamic spectrum allocation, are cornerstones of Ofcom’s plan to open up more spectrum for mobile broadband. Currently, around 29 percent of spectrum is shared between public and private sector users, and increasing that percentage is vital to achieving the government goal of opening up 500 MHz of new sub-5 GHz frequencies.

Cloud-based services featuring constant streaming of data and content, rather than
more periodic application and data downloads, also are affecting thinking about the design of networks.

The GSMA believes that mobile cloud traffic will account for 70 percent of total by 2020, as compared to 35 percent in 2013.

While the dominant design of a mobile network will continue to be based on support for roughly symmetrical traffic, future requirements will be for support of asymmetrical traffic.

No comments:

Will AI Fuel a Huge "Services into Products" Shift?

As content streaming has disrupted music, is disrupting video and television, so might AI potentially disrupt industry leaders ranging from ...