IDS in Mid-Morph

Intrusion detection systems (IDS) technology isn't dead -- it's just gradually being retooled, according to an IDS/IPS expert who will present his findings at an upcoming conference.

Gene Schultz, author of Intrusion Detection and Prevention, and CISO and CTO of High Tower Software, will outline the future of intrusion detection technology, including intrusion prevention systems (IPS), on May 22 at Interop, a tradeshow and conference owned by CMP Technology, which also publishes Dark Reading.

Although he doesn't expect IDS to ever shake its beleaguered signature-based technology altogether, he sees several trends in how the technology is being augmented to better help organizations protect their networks and data.

"The product isn't really changing, but its use is," Schultz says, although it will look different in the future.

IDS/IPS technology has been behind the eight ball for a long time, unable to keep pace with new forms of attack. Some experts have even dismissed it as irrelevant, or stranded investment. (See IPS: Still Playing Catch Up and IDS/IPS: Too Many Holes?)

But Schultz says there are already signs of new life. For one thing, IDS data is being used as part of intelligence-collection for forensics, he says. "People are gathering a wide range of data about behavior in machines, the state of memory, etc… and combining it to find patterns of attacks. Intrusion detection is one rendition of going more toward the route of intelligence-collection. Instead of focusing on micro-details like packet dumps, [security analysts] are looking at patterns of activity through intensive system and network analysis on a global scale, to determine what the potential threats are."

Schultz attributes this to a new breed of intrusion detection analyst, "more like an intelligence analyst, especially in the government."

Another trend in how IDS is being used is in more sophisticated data "fusion," he says, as in event correlation. "You get valuable output from firewalls, IPSs, routers, that can fill in the pieces of knowledge about patterns of attacks."

Pattern-matching events from different devices can help compensate for shortcomings in individual tools. "The best single source of intrusion detection [data] is the firewall, but there's such an overwhelming problem to inspect firewall data. And firewall logs don't lie."

With input from IDS, IPS, firewalls, DNS servers, and Web server events, "you can start to get a good picture of patterns of activity in your network. A combination of these is the intrusion detection system of tomorrow."

ArcSight, Cisco, and Schultz's company, High Tower, all offer these types of event correlation tools. And Sourcefire's Snort has a strong event correlation engine as well, he notes.

Schultz says he expects IDS/IPS products to be packaged with honeynet technology as well, which will let you record what attackers are trying to do. "Some vendors have had honeypot capabilities built into their products for years now," including NFR Security Inc., he notes.

"It lets you recognize motives, methods, and what machines they are targeting -- all very useful in staving off new waves of attacks."

Neural networks and data mining are two other areas Schultz expects intrusion detection to expand into -- neural network technology for finding patterns in events that are detected and having the intelligence to stop an attack; and data mining for things like spotting slow and gradual attacks. "One of the current limitations of intrusion detection is it can only tell you what's happening now. I expect intrusion detection of the future to not only do real-time correlation of events, but also do sophisticated data mining" to detect stealthy, slower attacks.

"A lot of attacks begin with some kind of vulnerability scan and reconnaissance. Then comes some kind of attack directed to exploit the vulnerabilities."

When these attacks are spread out over weeks, for example, it's tougher to spot them and connect the dots, he says. That's where data mining would come in.

And Schultz says he thinks a name change for IDS/IPS may be in order in the future, especially since the tools have historically gotten such a bad rap: "IDS should probably be 'attack analysis and detection.' "

The good news, he feels, is that many IDS systems are already becoming less reliant on signatures, and using rule-based engines instead. "This generation of IDSs is getting better."

— Kelly Jackson Higgins, Senior Editor, Dark Reading