Mobile Application Performance Management (APM) and Bandwidth Management – Some Perspectives

Although the issues in mobile APM – bandwidth, application chattiness, latency, network quality, protocol, and congestion – are the same, there are some interesting differences when one looks a little closer. In the enterprise world, the cost of “wired” bandwidth has come down over the years. Companies following best practices have learned how to manage, prioritize, and allocate bandwidth to business critical applications via QoS mechanisms. The segment of network where bandwidth is generally a problem is in the last hop to the facilities hosting the end user devices such as PCs/laptops. The bandwidth at the data centers and in the backbone network is generally abundant.

If one has to make a broad statement, we can say that “wired” bandwidth and its capacity management (both for the enterprise and the service provider) are under reasonable control. This management should lessen the bandwidth problems in APM.

In the mobile world, bandwidth has always been scarce because of limited spectrum made available to mobile service providers (MSPs). Although the demand for voice bandwidth is plateaued, there is an explosion in the demand for data bandwidth because of immense popularity of Smartphones and Apps. Also, unlike in the “wired” world where last hop to the end user is generally distinct between consumers and business, the air bandwidth from the MSP are shared among consumers and business users in the mobile world.

Although the mobile data bandwidth available to Smartphones has been increasing because of introduction of 4G/LTE technology (where one can get 10 Mb/s– 20 Mb/s) downstream, it is a rat race to keep up with the pace of new Apps, especially in the video streaming category. The MSPs are in this unenviable task of managing this mobile bandwidth capacity. Not surprisingly, most MSPs have already capped the monthly bandwidth usage to the user (there are still concerns on the stress put on the signaling network by certain type of applications which is not addressed adequately). Although capping traffic is an easy first step, we can expect more challenges faced by the MSPs in the capacity management of the mobile bandwidth.

Hence we can expect bandwidth issues to continue to plague the application performance in the mobile world. We also expect chattiness/latency and packet loss issues to continue to impact application performance. The APM techniques and methodologies used in the “wired” world can continue to be used for mobile applications. But capturing of mobile traffic stream from a Smartphone for analysis purposes is much more difficult than from LAN attached PC.

The MSPs have to start understand the type, traffic volume, and usage pattern of the mobile Apps flowing through their network. There are tools from vendors such as Procera Networks, Allot Communications, Cisco, Opnet Technologies, and others for conducting application traffic analysis. Scalability of these tools in terms of capacity and the sheer volume of the performance metrics may pose a problem. MSPs also have to deal with regulatory issues and privacy concerns regarding inspecting App traffic.

This challenging task of mobile bandwidth management can also open new doors of opportunity for MSPs. Examples are providing QoS for Apps and charging for it, charging based on time of the day, collecting charges from the destination portals, etc. We have surely entered a very interesting and exciting phase of mobile data communications.