As we move into an era characterized by rapid generation of large volumes of data, many are looking at how such data can be leveraged more effectively to give their organization a competitive advantage.
To many telecom service providers, big data is a natural evolution of business intelligence -- one that started in the early 2000, with office applications being used to analyze customer billing and business support information. As telecom services matured, business intelligence and business support solutions evolved from focusing on systems and database integration, for the purpose of reporting historical organization performance, to providing timely information of how the organization is performing. The next stage of the evolution of business intelligence and analytics would be to provide real-time and forward-looking information that could be automated into existing decision processes.
With data being generated at an unprecedented scale from a variety of sources and in a variety of formats, many CSPs are faced with the challenge of managing, organizing, storing and computing this flood of data in a timely manner. As large amounts of structured and unstructured data are generated in the network from numerous sources, like billing systems, CRM, machine-to-machine data sources, OSS and social media, CSPs have to be able to store this information to quickly analyze it without loss of data fidelity. That is where the promise of big data comes in.
Many companies have started providing a myriad of solutions to help CSPs with their data woes. These range from Hadoop and MapReduce to in-database analytics and in-memory database to text mining. Over the past few years vendors like IBM, Oracle and EMC have gone on an acquisition spree to acquire big data and analytical companies to beef up their current portfolio to compete with newer players like Splunk, Cloudera Ontology and Tableau. This trend of incorporating big data capabilities is expected to continue as more solution providers jump onto the bandwagon and new companies emerge to cater to specific data and analytical needs of CSPs.
Although internet companies have been some of the earliest to adopt big data, other industries will need to follow suit quickly. This is especially true for CSPs faced with increasing competition from over-the-top players. CSPs' traditional business models are under threat and they will have to evolve to meet with challenges associated with the digital era. Although growth of mobile data and the current provisioning of LTE data services may provide some respite from declining traditional revenues, CSPs will have to look for ways to monetize their other assets, like network information and subscriber information, to maintain profitability in the future.
However, CSPs are expected to face several challenges associated with implementing big data and the associated analytics. First, the implementation is daunting for many CSPs both in terms of cost and implementation time. CSPs have to prioritize their capital expenditure, and big data initiatives might not make it to the top of CTOs' priority lists. Second, the lack of internal big data and analytical talent and the high cost of acquiring and retaining such talent is expected to be one of the largest obstacle of successfully implementing these solutions. To be able to effectively extract the most from big data solutions, these professionals would need to be conversant with IT, statistical and business domain knowledge. Such a combination of skill sets will be in short supply.
To help meet the challenges faced by CSPs, it is expected that niche solutions providers will emerge to provide "out of the box" and accessible solutions (solutions that would not require a user to hold a PhD) to complement the offerings from larger big data solution providers, which target the more sophisticated needs of some CSPs.
Mark Koh is a senior industrial analyst with Frost & Sullivan's ICT practice, Asia Pacific. For more information contact [email protected]