OSS self-analysis

A lot of people asked me if it was frustrating not having a clear specific diagnosis, but I didn’t mind, I just chose the most optimistic diagnosis.”
Karen Duffy

Predictive analytics has become a bit of a cliche for the OSS industry, one which I’ve succumbed to using too (see these previous blogs as but a few examples – 1, 2, 3, 4 and 5). As number 4, “The touchpoint explosion,” indicates, most reasonably large networks are already too complex for humans to manage in real-time and the advent of network virtualisation and the Internet of Things (IoT) are only likely to exacerbate this situation.

Today’s blog takes a slightly different perspective on how OSS can make use of predictive analytics. Rather than network health, service take-up, service churn, etc today we’ll look at managing network capacity and resource utilisation. In a virtualised world, the OSS will even be managing the resources that they reside on themselves.

Data centres have been considered a growth industry for a number of years now and I expect that will continue to remain true in terms of ever-expanding service volumes (demand). However, as hardware costs reduce and automations help drive efficiencies, I can also foresee commoditisation of all but the most niche of data centre services.

With commoditisation comes reductions in profit margins, which in turn brings an increased focus on doing more with the infrastructure you have rather than just blindly expending CAPEX on more, more, more.

Automation of data centre services has evolved to the point where dynamic allocation of resources is commonplace. But dynamic allocation poses challenges for forecasting demand, as services are established and torn down on the whim of your customer collective. This is where predictive analytics will be required to provide more sophisticated forecasting models from the vast amounts of data collected from the data centre’s elastic IT systems and matching against dynamically available resources on that existing infrastructure.

Just one of the levers of the predictive analytics model will be the forecasting and scaling of OSS capacity to manage the changing service, resource allocation and infrastructure models.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.