Carl Weinschenk spoke with Akil Chomoko, the product marketing manager for Volubill.
The past few years have made one thing certain: Networks on occasion will be called upon to do more than normal. This can be both expected traffic – such as that generated by the recently concluded NCAA basketball tournament – and it can be from unexpected emergencies. Akil Chomoko tells IT Business Edge blogger Carl Weinschenk that there are several ways to make it through times of high demand.
Weinschenk: Is there a significant difference between preparing networks for an expected spike – such as the NCAA tournament – and unpredicted emergencies?
Chomoko: It is a different mindset. For spikes like special events, operators will do a variety of different things depending on the nature of the spike. If it is a football tournament or festival, for example, they may want to put additional radio access network points in place. The response may be additional Wi-Fi for offload or it may be 3G or 4G masts. Those are typical ways of planning.
In the case of extra Wi-Fi, for instance, a local manager would approach the telephone companies involved and say they are doing this for the Olympics or whatever and they should be aware and become part of the planning team and help arrange things.
Weinschenk: What about an emergency, such as Hurricane Sandy?
Chomoko: Hurricane Sandy was quite a difficult one. [Handling such situations is] very dependent on infrastructure. There are a variety of things that can be done. Some examples: In that sort of emergency situation operators can prioritize certain types of traffic and deprioritize others. They can use deep packet inspection. That’s one element. For instance, traffic to certain destinations and numbers may be allowed through whereas everything else is on low bandwidth or low priority in terms of connection. That’s pretty standard, really.
There are new ways of doing things which are more focused. For example, there is now technology available, it’s on some Sprint devices, with embedded agents in them which are designed for battery saving. They can stop particular apps, since apps in which requests go all the way to the edge affect more of the network.
Weinschenk: Why is it important to stop apps at the end device?
Chomoko: A device will keep on trying to get access to something or do something if it keeps on being blocked at the core. It keeps knocking on the door all the time. It is more effective to stop the device from continually retrying. The problem is that not all devices have embedded agents that can be controlled. Obviously the core is the first line of defense, but control can be more effective by going out to the edge.
Weinschenk: Is technology such as yours aimed at one type of network?
Chomoko: It is agnostic. It sits atop broadband, wireless and mobile networks. As a subscriber interacts with any service and uses any application, such as VOD, we police how they use the service in real time. If, for example, 1,000 VoIP calls initiate from one location at one point in time to a VoIP server, our software asks what to do. We may allow it to happen or first check if there is capacity … and then let it occur, once we check to make sure the subscriber has credit and meets other conditions.
Another approach is caching. People watch a lot of video. Why not cache it on the Internet and bring it to points closer to where it is used? We work with Akamai and Ericsson for instance to support things like March Madness.
Weinschenk: It seems that network operators need to be as proactive as possible.
Chomoko: Networks are dimensioned for standard service planning. Some redundancy is routinely built in. It is an element of spare capacity. But is very much a best effort approach on the part of operators. Now we take more of a view of controlling things and packing them and charging appropriately as a way to make money and to stretch the soup. I think every operator in the last two years started to do that.
Weinschenk: Are there standard approaches to network planning?
Chomoko: There is a network convention but the way marketing departments in most organizations promote goes way beyond what the network planners have built for. It is well known that most mobile networks are at saturation. Networks plan, but some plan better than others.