Big data analytics for better network operations

The growing demand and consumption of data services by consumers and enterprises is leading the explosion in the traffic that transcends telcos’ networks. As traffic increases, customer and network data increases as well. Furthermore, traffic carried on telco networks is becoming increasingly rich, with voice being supplanted by data services such as streaming audio and video. On some telco networks, video now accounts for as much as 60% of total traffic. LTE deployments and associated marketing campaigns will boost the consumption of data services.

This increased data traffic comes from a variety of sources (smartphones, TVs, tablets, and laptops) and multiple channels (social media, web chat, email, and voice calls). Machine-to-machine (M2M) communications will increase data service usage as customers use the network to control devices in places like the home or their car.

Meanwhile, telcos also have operational data such as billing, network, location data, and call detail records, presented in a structured format, which is typically contained in SQL databases. Semi-structured and unstructured data such as call logs, social media messages, text messages, emails, customer feedback documents, system logs, and sensor data is also present.

The rate of data growth will exceed the capacity of telcos’ existing data warehouses. To support internal decision-making processes, these new and varying datasets must be ingested by storage platforms and processed using analytic tools to gain insights. The insights will inform how to drive increased ARPU, predict and reduce customer churn, deliver improved customer experience through personalized services, and limit the operational costs associated with network management (through network design, planning, and optimization).

Telcos fully understand the need to store every form of data available to them. However, the current approach adopted by existing data warehouses is dependent on relational database management systems (RDBMS) which are inefficient in handling large data volumes, the varying nature of the data, and the demand for data processing speeds. Some telcos are increasingly supplementing their RDBMS with complex event-processing tools that stream data either into the data warehouse or directly into an analytics engine to provide real-time monitoring and analysis capabilities. However, with this approach, not all of the data can be analyzed, so more advanced technologies such as Hadoop are required to improve the storage and processing capabilities.

Telcos are well positioned to take advantage of big data analytics. Data will continue to increase and mounting business pressures make the adoption of big data technologies inevitable. However, telcos must define what business challenges they need to resolve first. They are faced with multiple challenges, most of which require the telco to exploit all forms of data to derive new insights that deliver problem resolution. For example, telcos must understand why customers are churning, and what service conditions must be averted to avoid customer churn.

While data resides at a number of levels within telcos (including financial, network, service, operational, and customer), this article focuses on how big data analytics can be applied at the network level by leveraging data from other data sources such as the customer and the environment to obtain actionable and valuable insights.