“Each year Gartner gives buzzwords a reality check by publishing its Hype Cycle report. The latest one focusing on big data shows that the industry is just reaching the peak of its hype right now, with vendors flocking to the market, customers getting anxious that they don’t yet fully understand the technology, and the expectations about what big data can do for an organization being over-inflated.”
As can be seen from the diagrams below from the Gartner Group, Big Data has been in the “Peak of inflated expectations” for their last two Hype Cycles.
Gartner Hype Cycle 2012
Gartner Hype Cycle 2013
However, there is still confusion that pervades the industry with many different variants and interpretations of what big data really is. To be honest, the phrase has almost become passé. Maybe it’s just because the term seems silly to me. The “big” data of today will actually be “tiny” data in years to come. Nonetheless, the intent certainly still has merits for CSPs. The definition of big data is often characterised within the constructs of the following four dimensions:
- Volume – The amount of data
- Velocity – The speed of change (or delta) of new data in and/or out
- Variety – The different sources of data (not to mention types / schema) and then throw in the additional dimension of
- Veracity – The reliability of the data
By any of the first three measures, CSPs are seeing rapid expansion in their B/OSS data sets. And as we’ve discussed previously, the integrity of data can be the difference between the B/OSS being useful or not. I’m not going to try to describe the relative virtues of MapReduce, NoSQL, etc (nor could I do justice to such a comparison), but the big three V’s of big data effectively equate to too big, too fast and too varied for the traditional RDBMS-based OSS of the past. The big V’s also present a major challenge to the rat’s nests of integration that have brought OSS to the point of being almost unchangeable at some CSPs.
Which of the nascent big data engines will take OSS / BSS into the future of virtualised networks with their increased volume and velocity, not to mention the search for alternative data source varieties in the never-ending quest for big value from CSP’s big data?
At the end of the day, whether your analytic engine is termed big data or not, if it’s delivering big value, actionable intelligence from your data then it’s doing its job.
Just out of interest, is a skills availability, or lack thereof, hindering the ability for your organisation to become more active in a big data analytics approach?