- Date published:
- Author:Brian Wood
Less is more -- or so it would seem for firms weighing in on the topic of "big data security".
And if "less" is 40 terabytes PER DAY of security data, I shudder to wonder how much "more" would be.
The article is from Ellen Messmer of NetworkWorld.
Emphasis in red added by me.
Brian Wood, VP Marketing
Do enterprise security teams want "Big Data Security"?
At RSA Conference, security professionals indicate their security-related data is pretty big already
A big theme at the RSA Conference this year is a concept known as Big Data Security -- the idea that massive amounts of data related to both network security and of business context should be stockpiled to be analyzed to pinpoint malware, rogue insiders and stealthy attacks aimed at stealing sensitive data.
But the four chief information security officers (CISO) on a conference panel about Big Data Security indicated that they already have their hands full with plenty of security data to analyze on a daily basis to defend their corporate networks, and for now, that's enough. Their viewpoints suggest it may be a while before enterprises, even very large ones, clamor for the kind of Big Data Security deployments that IBM, HP and RSA, the security division of EMC, now insist is the next big thing in corporate security.
But the HP, IBM and RSA concept about Big Data Security is also largely experimental, too, and would require use of security-event information management (SIEM) products from any of these vendors combined with analysis of massive databases of both security-event information as well as business context which combined could help pinpoint attacks.
The CISOs who spoke on the RSA Conference panel -- Ramin Safai from investment firm Jeffries & Co; Alex Tosheff from eBay's X.commerce company; Carter Lee from the e-commerce company Overstock.com; and Praveen Money from managed services provider DataShield, indicated their businesses are constantly under attack of one kind or another, and they and their staff work hard to ward off compromises, identify attackers to recognize their styles and take actions to keep them out of corporate networks.
At these firms, security data related to attacks and suspicious events is big enough. "It's 40 terabytes," said Lee, about the daily security data intake, though even more metadata around these attacks is stored by Overstock.com.
"It's an arms race," said Tosheff of the eBay X.commerce company, who says security-event information comes in from the networking tier, applications tier, and desktops. "We look at 10K events per second."
Safai of investment firm Jeffries & Co. said the firm even uses its own versions of honeypots — "PCs we know that are easy to get to — and if we detect someone has hit that PC, we try to figure out how." He said the firm makes use of the FireEye and Solera security technologies for packet capture. He adds that expanding business into countries where there's thought to be even more risk has made the security team even more concerned about internal traffic.
Praveen Money from DataShield said the main questions that get asked are how is an attack spread, what was compromised and is there an advanced persistent threat related to stealthy attack, and what was the exploit.
None of the CISOs on the panel, which was moderated by Richard Stiennon, analyst from consultancy IT-Harvest, seemed to express much enthusiasm for the concept of Big Data Security now being put forward by IBM, HP and RSA that calls for analysis of gigantic amounts of data both traditional security and business, such as from human resources (H.R.).
"It's not about volumes of data," Tosheff remarked, though he noted that eBay's approach in the X.commerce division is to "use some H.R. and physical security data. We know that a person is currently out of the country," and geo-location data helps identify potential attacks. He adds a lot of defensive techniques his team has come up with are relatively home-grown and there are "beacons" in the network to monitor the presence of an invading attack to watch what happens before making a calculated response.
For now, it's not obvious why Big Data Security should be getting a whole lot bigger.