Communications Service Providers (CSPs) have traditionally managed large volumes of data about their customers and networks, from sources such as transactional data, network performance, service quality and Web interactions.
The first analytics challenge facing CSPs is the rapidly accelerating volume of raw data available. As traffic increases, the amount of network data increases as well. A network with 20 million subscribers typically generates several billion signaling events-per-day, with peak volumes of several hundreds of thousands of events-per-second. CSPs now need to take advantage of real-time event streams, such as network signaling on the control plane and user plane data, to generate insights into customer experience, for example. The deluge of data is creating significant interest in Big Data techniques for arranging and analyzing these high-volume event streams.
The second analytics challenge facing CSPs is that of the increasing velocity of the data itself and of the speed required to create actionable insights from such data. Decision-makers often use historical information to predict the future in rapidly changing environments by using data warehouses and Business Intelligence tools. While historical information can be useful, it is a representation of the past and may not be an accurate indicator of the future. Historical data needs to be taken into account as part of the context of the overall problem, but this rear-view must be tempered with what is happening right now. Operators increasingly need to generate real-time insights from sources, such as network feeds, so that they can, for example, detect when high-value customers experience dropped calls and take immediate remedial action.
The third analytics challenge is that of data variety. Data is scattered across a CSP’s organization, in every imaginable format. An increasing volume of data comes from third-party sources – social networks being an important example. Not only is it a challenge simply to integrate and aggregate all this disparate data, but the variety of information sources and formats makes the task of correlating this data in real-time particularly difficult.
Such data volume, velocity and variety challenges are overwhelming decision-makers who are under increasing pressure to make the right decisions quickly. The question is, what is the right approach to address these challenges, and why should CSPs invest in solutions for these problems?