The Three Vectors of Mobile Analytics

Today’s announcement of JDSU’s acquisition of Arieso, a UK based, location analytics company, demonstrates an industry movement to obtain new sources of intelligence about the performance and usage of mobile networks.    The migration from voice to data services has left mobile operators blind to how networks, applications and services are being used by consumers – a gap that the vendor community is anxious to fill.

To fill this gap, it turns out that there are three main sources of data within mobile networks from which analytics and intelligence can be built through network and device monitoring – each have their own benefits and considerable challenges.    

JDSU’s primary source of data to date, has been from probes in the network – products that capture packets, backhaul the data and then analyze content.   This method excels in providing insight to a specific problem or issue.  The challenge is that, as a general network monitoring tool, it can generate huge volumes of traffic as the packet captures are backhauled across the network.   Assuming you can solve for this (by increasing backhaul capacity, or pre-processing raw data at the probe site as some DPI vendors are doing), you still have the challenge of managing the large volume of resultant data that shows up in your data center.

Arieso sources data not from the network directly, but from the existing OSS systems.   These systems typically receive data and diagnostics from components in the network such as cell switches, SGSN and GGSN systems.   Tapping into the diagnostic information from these systems provides an alternatively to probes (although backhauling large volumes of diagnostic data from switches isn’t free either).   The bigger challenge is that analysis of this information is vendor and product specific.   This can create a large integration cost for sourcing and decoding the diagnostic information from each type of OSS system, switch or device and has to be planned in advance.  A probe however, can be dropped into a specific location in the network at a moment’s notice.  

The third source of information is from the handset or smartphone itself – obtaining diagnostic information which is then periodically uploaded from the smartphone to an analytics engine.    This provides the highest fidelity information because it proves how the user actually experienced the service (rather than a view from the network) and takes into account the mobile device behavior in addition to the network.   Aside from the privacy & consent issues for gathering diagnostics from the handset, because you don’t want drain the battery, this information is not typically real-time.   Consequentially, unlike probes and OSS integration, it provides a high quality but post-event view of how both the network and the smartphone performed.

For mobile operators, obtaining the performance and usage data they need will likely require a blend of information sources to get to a real-time network view, understand performance over-time and solve individual consumer problems.

For JDSU, the acquisition makes sense because it recognizes that a complete solution requires multiple sources.   The opportunity for JDSU now, is to  combine these sources into a coherent real-time view of the network with sufficient granularity to allow mobile operators to identify and solve problems as they occur.