Security: The Final ISP FrontierBy Larry Seltzer | Posted 2006-09-25 Email Print
Re-Thinking HR: What Every CIO Needs to Know About Tomorrow's Workforce
Opinion: These are the errors of the ISP market. Its 15-year failure to protect its customers. Is a secure ISP network science fiction?The role of ISPs in security is one of the great neglected topics in our industry, and one of my favorite subjects back to the time before I started focusing on security.
Back, I believe, in 1999, I wrote an article predicting (because it made perfect sense) that the future of security for consumers was through the ISP.
Anti-virus, Anti-spam, perhaps even network security like firewalls could be implemented by the ISP.
When you think about it, many cutting-edge enterprise network security features could be applied to an ISP, up to and including NAC.
But this week Trend Micro is releasing its ICSS (InterCloud Security Service), a first step toward helping ISPs and some other large network providers, like universities, to make their networks safer.
ICSS replaces the existing recursive DNS in the network and uses that position to monitor activity looking for suspicious acts, especially those indicative of botnets.
They claim it can detect compromised systems in near-real time, remediate them and remove the infection.
This capability is impressive, but not surprising coming from Trend Micro and its vast experience on corporate networks.
Some of the behaviors they look for are relatively obvious: Any ISP client computer that does a large number of MX lookups on the DNS in a short period of time is probably a spam bot.
Even better, companies like Trend Micro have good maps of the big botnet C&Cs (command and control networks).
If a system on the network makes requests to one of these C&Cs, it has basically dropped its pants and you know it's a bot.
But would you really want your ISP (or your university) remediating your computer?
That's a very complicated question; I'll take the coward's way out and declare it to be a "policy issue." Clearly some users wouldn't mind this at all, just as some would raise cries of "Big Brother."
Personally, I went Republican on this issue quite a while ago and wouldn't mind ISPs blocking client systems that exhibit behaviors that are well-understood to be indicative of bots.
There's a good way to do it and a bad way.
In the good way the ISP a) pre-publishes and notifies customers of the criteria for blocking; b) notifies the customer when they are blocked, including details on what their computers did to get them blocked and instructions on how they can remediate; c) includes a reference # and a support phone number to call.
But ISPs have been miserable failures at fixing compromised systems on their networks, including large botnets.
Part of the problem is that they don't want to do all the work involved; part is that they don't want to offend customers by inconveniencing them just because their computer is a bot sending out spam; and a big part is that they don't have great tools for fixing the problems. This is what Trend says it is trying to address.
For instance, Trend tells me of one ISP in France that has 500,000 compromised client systems on their network.
They are currently plowing through four to five a day, meaning that we'll be colonizing Alpha Centauri by the time they're done. Clearly, the current methods are inadequate.
Being in the position of the DNS, there are some helpful tricks that a service like ICSS can perform: If they detect a command or request to a botnet C&C, ICSS could spoof the result.
They could tell the system to perform an attack on 127.0.0.1 (the loopback address). They could redirect the user's browser to Trend's Housecall site, where they could have the system scanned and cleaned, or some other site with a similar purpose.
It's not hard to see how malware could be designed to work around this specific technique, such as changing the default DNS to an external compromised one.
This isn't the point. The point is to provide tools to give large network operators a practical way to make their networks more secure. I know I'd pay more for that.
Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983. He can be reached at email@example.com.
Check out eWEEK.com's for the latest security news, reviews and analysis. And for insights on security coverage around the Web, take a look at eWEEK.com Security Center Editor Larry Seltzer's Weblog.