IM (instant messaging) for the enterprise is being touted as something new. Don't you believe it. I recall using ICQ Corporate at an ISP in Cleveland in 1998. Is it news because Wall Street is starting to take it seriously? You bet!

Today, serious broadband speeds are omnipresent in the enterprise. Broadband delivers "instant," and everybody loves instant, not just the enterprise (anyone notice the markets that sprouted from TV dinners?). Instantaneous data delivery facilitates the most desirable instant of all ­ instant communication, or IM.

Instant communication ­ everywhere, all the time ­ is a definite advantage for an ever more mobile workforce. But then comes the rub, er, hack. Peer-to-peer applications like IM are prone to security issues. Veteran IM provider Yahoo! recently discovered a buffer overflow vulnerability in both its IM and chat programs. We've seen holes by the same name patched repeatedly in other software only to see them reappear. Suddenly we realize that we're in for the same ride with IM client software.

Swimming Upstream
Intrusion detection experience with IM is hard to come by. State-of-the-art security measures provide a foundation of protection. Enterprises implementing IM need to build on that foundation by targeting IM's port 5190 with intrusion detection so that they can quickly gain that experience.

Like 802.11 hotspots, IM "proliferates in homes and offices faster than it can be secured," says Gary Morse, president, Razorpoint Security Technologies. When you leave port 5190 closed you don't expect trouble (see sidebar). When you open it to IM traffic, unless you analyze and secure this traffic, you have opened a door to hackers, whose computers are often set up to perform automated scans of IP address ranges in search of ports to exploit. Until we have experience analyzing port 5190 like we do port 80 (Web surfing), we will be learning as we go, perhaps the hard way, upon being the victims of intrusions ourselves.

Crippling Clients
IM is exploited by taking advantage of holes in proprietary IM client software, examples being AOL IM, MSN IM, and Yahoo! IM. If there are holes in the software, such as Yahoo!'s recently reported buffer overflow, these can be exploited at length until they are discovered and patched. "There is also a DoS attack on the network component," says Morse. "If there is a DoS on those Internet IM servers, no one using the IM service can contact anyone else. If a company is relying on that IM service, I just took down the ability for that company to use it."

As with other holes, once access is attained, hackers can reach other systems, set up accounts, and ransack the company's data. How do you protect IM traffic? "Application intelligence products [intrusion detection] look deeper into the traffic before it goes through the specified port numbers to see if it really is instant messaging," says Morse.

Other Recommendations
Razorpoint recommends "comprehensive attack penetration tests" for networks. This will immediately determine success or failure in gaining unauthorized access through IM or anywhere else.

"One of the things we recommend is to either limit the use of IM with a product that keeps most of the traffic inside your network [good!], or to install and maintain one of the application intelligence systems that looks more closely at the packets going through to ensure it is only IM traffic and not next quarter's financial reports," says Morse.

Razorpoint also recommends security design reviews before rolling out massive new systems like IM. This includes assistance with the actual architectural design of the system.

Proliferation Speed = Exploitation Ease
IM is easier to set up than 802.11x. This makes users instant targets. "One of the attacks that we've heard about is where an attacker can run software on your machine from outside. They can do it even if your IM client isn't running. So if you installed it but it's just sitting there, not connected, they can send you a carefully crafted e-mail, and when you get that e-mail, that e-mail will invoke a component in your IM client and make you vulnerable without you knowing," says Morse.

Wireless IM
"If you're going to have people do remote access, then you're talking about firewall protection for the remote user and a VPN-type solution," says Morse. The products are here today to afford cellphones VPN and firewall protection. But..."You can give me the greatest firewall and the greatest network today, and three weeks from now there could be new vulnerabilities and you're going to have to re-secure it," says Morse.

What Would Razorpoint Have Done for Yahoo!?
"I think it could be as simple as testing. There are tools and techniques you can use to look for these buffer overflows, back doors, and DoS attacks. I know Yahoo! is doing that. But even if you have a large budget and this is your core competency in the company, you can still miss things [case in point]," says Morse.

Discussion
The traffic analysis is going to be the same for IM as for any other port or service. It's just the traffic that may be different because you're dealing with a different port and different traffic, which we don't have as much experience monitoring. "You're also now going from IM for your friends to IM for business purposes. This has to be looked at as a business solution and maintained and managed as such. Just because someone in an office can download free software and start using it doesn't mean the company doesn't need to have a policy like they do for everything else about technology ­ a policy for IM and how they monitor it," says Morse. If it's on their phones, companies are going to have to own the phones or otherwise have control over them even though employees may be taking them home with them every day (a whole other question).

IM Security ­ Direct, Simple Solutions
I have been talking with Ira Winkler, chief security strategist at Hewlett-Packard, about security issues of late. Here are Ira's thoughts on IM security:

"While IM is a great convenience, it can be a security nightmare. At the very least, companies have to make sure that IM doesn't go in or out of the firewall. IM allows for everything security efforts try to plug up. It allows file transfers in and out. It compromises personal privacy by letting people know where they are in theory.

Companies interested in using IM should purchase a more secure version of the software if they believe the benefits of IM outweigh the security issues. If companies don't want to pay for a commercial version, then they clearly don't think there will be a strong enough benefit."

If you don't want it bad enough to pay for it, you don't really want it. That makes sense. Even more quick, slick, and sensible is the solution that doesn't poke holes in your firewall. Pay for secure IM and keep it on the network, inside the firewall, and you don't need to open ports, perform intrusion detection on those ports, study the traffic, or anything like that. ­ DG