What Big Data and Virtualization Means for OSS/BSS

Big Data might just be critical to the success of SDN, NFV and CSPs’ cloud strategies, but are the incumbent OSS/BSS systems ready for the change?

Many CSP big data projects are standalone business intelligence or customer experience solutions. Unlike these ‘pure’ big data solutions, closing the loop between data, analytics and the network itself will require incumbent OSS/BSS to be far more flexible than it is today.

At a telco big data congress last year I was impressed by the scope of big data solutions being applied to both business support and operations support challenges within the CSPs. But in almost every case these solutions were ‘standalone’: They would pull in some data, analyse it, and generate results. Results were to be consumed by just one or two systems: A management reporting tool perhaps, or a Suggested Products prompt in a customer web portal.

They generated great results and both the cost and time to deliver the results was very modest. In fact the overall return-on-invest (considering cost of software, cost of servers, cost of effort, and the pay-back time) was impressive compared with a typical OSS or BSS project.

Why such a good ROI?

These projects conformed to the classic big data model: A ‘read only’ data extract, analysis of data in a ‘black box’ that doesn’t have to integrate with other live processes, and presentation of results for ‘read only’ consumption.

These projects weren’t subject to the ‘integration tax’ of a typical CSP’s enterprise IT projects; a cost that comes from the necessary integration of multiple servers and OSS/BSS systems that manage critical data and business process. Systems that support large numbers of concurrent transactions. Systems that have to integrate with real-world network equipment and customer services.

One telco once told me that, for any OSS/BSS project, they start with the assumption that system integration will cost about £300,000. Just for effort.

The vision for telco big data is for a more integrated experience: Analytics that directly affect the network state (for example, influencing SDN controller policies, or moving apps around data centres to meet demand) and customer experience (such as adjusting service policies and charging to resolve or avoid service quality issues).

The implication is that big data needs to be hooked up to incumbent OSS/BSS systems. But if the starting point is a bill for £300,000 of integration effort, well, there goes the impressive return-on-investment.

OSS/BSS vendors need to lead the way. Their products need to change so that they can play nicely in this new world of flexible, agile and real-time big data analytics.

The problem, today, is that many OSS/BSS systems, architecture and data models are proprietary and rigid. Why? Because historically CSPs data requirements, network topology and business processes have also been proprietary and rigid.

The software model followed the business model.

New business models, whereagility is a reality rather than a marketing buzzword, will require agile OSS/BSS systems.

Doing analytics while the world (the customers’ services, the network state) moves on, making anintelligent, automated decision, then changing the networks/services inreal-time is going to be hard. But this needs to be achieved to realise the full benefit of emerging technologies like SDN, NFV and cloud-based communication services.

This post was originally published in telco and media big data blog, CSP Data.