If you are not on the edge you are taking up too much space!

Insight

On the 3rd November CW was at Bird & Bird for the latest meeting of the Virtual Networks Special Interest Group (SIG). The legal firm’s contemporary offices in the City of London provided the perfect backdrop for a discussion on what is an extremely current trend within the mobile industry; Mobile Edge Computing (MEC).

Ning Wang from the University of Surrey kicked off proceedings outlining how the University’s 5G Innovation Centre plans to leverage MEC to support operations in various mobile applications, both content and non-content based. It was the role that might be played within Device-to-Device networking in particular that really caught the attention of the delegates. In not-spot regions MEC could enable network coverage by utilising knowledge about the mobility of local mobile users to identify nearby devices or ‘helper candidates’ and request access to their network. It’s not an exact science of course however studies indicate that human mobility can be predicted with an accuracy of up to 90%!

Matt Stagg from EE, provided some insight into what the operator have been working on in this space including their trials of edge video orchestration as part of the initiative to make Wembley the most connected stadium in the world. Matt also gave his opinion on what is needed for the adoption of MEC, crucially that it be an extension of the cloud computing and not a replication.

Julian Roldan Ramos from Intel noted that while the industry is searching for that killer use case to justify investing in the hardware, this may not be the right approach. Mobile Edge is effectively just a specific use case of ‘Edge’ and there are a few ‘Edge’ NFV use cases. MEC is also using many of the underlying technologies of NFV, such as virtualization and standard hardware. Therefore the hardware is essentially already there and MEC could be easily integrated into those Edge Based NFV use cases as a software component. If implemented on a wide scale  networks could then become completely software defined and dynamic in nature.

The final talk of the day from Nishanth Sastry of Kings College London took a slightly different spin on the topic. Sastry and his team spent 9 months analysing nearly 6 million users of BBC iPlayer, unveiling key consumption patterns.  Typically they found that users prefer to watch serialised content of shorter duration on catch-up. These analytics can now be deployed to cache content that users show an affinity for at the very edge of the network. For example if someone watches 3 out of 4 episodes of Dr Who, it is highly likely they will watch the 4th as well. This could offload huge swathes of traffic from the network and as a result make vast energy savings.

If you would like to engage further with the Virtual Networks SIG follow #CWVN on twitter and tweet your questions and insights to @cambwireless

To view presentations from the day visit the CW resources page.