The recent RSA conference in San Francisco was awash in talk of big data, but it was clear there was some disagreement about what people mean by big data and some outright skepticism about it being the answer.
The recent RSA Conference in San Francisco was awash in talk of big data, but it was clear there was some disagreement about what people mean by big data and some outright skepticism about it being the answer.
Big data was a key theme in the opening keynote by Art Coviello, executive chairman of host company RSA (the security division of EMC), which announced a big data product in January. The new analytic tools are necessary, Coviello says, because of the scope of the challenge.
[ MORE FROM THE CONFERENCE: Cloud forensics: In a lawsuit, can your cloud provider get key evidence you need? ]
After all, organizations are swimming in security data. In a panel discussion at the show Ramin Safai, chief information security officer at Jefferies & Co., said his investment bank with 5,000 employees captures 25GB of security-related data every day. Buried in that they typically find 50 issues to examine more closely, two of which end up requiring real attention.
Do we need fancy new big data tools based on technologies like Hadoop to help contend with the deluge? Not necessarily, some speakers said. There are lots of ways to collect data and you can use tools like Splunk to correlate it, said David Hannigan, information security officer at Zappos, who participated in a CISO panel discussion.
But fellow panelist Stephen Moloney, manager of enterprise information security at Humana, is looking for more. He would like a big data tool to bring together information from disparate sources and give him the big picture. "Instead of a snapshot of the Grand Canyon, I want to see it from 30,000 feet."
Isn't that was security information and event management (SIEM) systems were supposed to do?
Session moderator Rick Holland, a senior analyst at Forrester, asked the audience of about 100 to raise their hands if they had deployed SIEM. About 60% of the group had. "How many consider their SIEM successful?" he asked. Only a few kept their hands up. "So, if we can't get small data right with SIEM, why do we think we can get big security data right?" Holland asked, to the laughter of the group.
Moloney, however, is still holding out hope for SIEM. "We're building out our SIEM and collecting all the data we can," he says. "We have a large security operations group that understands it very well. They're constantly retuning the sources to make it more valuable."
Hannigan agreed that tuning is critical, and that people are key. If they don't understand what they're seeing it isn't going to matter if your tool is based on small or big data.