Auto drivers spend the vast majority of their time peering through the windshield, only a fraction of that time glancing in their rear-view mirrors. That’s because driving is a mission-critical activity, not unlike the decision-making required to run a serious business. Why then do most executives focus their analytical attention on the past? Why not look at what’s happening now, what’s coming down the pike?
One reason is because the tools for analyzing data have all been focused on the past. That’s now changing, in part because real-time data streams continue to proliferate. Still, accessing and effectively managing such streams requires an engine that can speak the Lingua Franca of data: SQL. Enter, SQLStream. Unlike traditional relational database management systems (RDBMS), which persist data that can later be queried, SQLStream runs continuous queries on targeted data streams. This enables a real-time view of an organization’s situation.
This was the topic of our discussion in a recent episode of The Briefing Room, with Dr. Robin Bloor and Damian Black of SQLStream. We learned:
- SQLstream holds five patents covering the basic notion of relational streaming and streaming servers processing relational operations continuously; three more patents are still pending.
- Several analysts have said that 2013 will be a key year for streaming analytics, the engines for which will be central to the integration of operational intelligence platforms.
- Examples of streaming Big Data that can be analyzed for business benefit include application log files, GPS data, call record details, Web traffic and network traffic.
- One specific area of value is telematics, allowing telco companies to remotely monitor the health of devices, systems and infrastructure, then avoid outages by recognizing problems as they occur, or even effectively predicting device or system failure.
- Dr. Bloor noted that Hadoop isn’t actually a very good parallelism engine. It’s not a performance engine. Its beauty is in its versatility, and the fact that you can store any kind of data in it easily.
Full event archive here.