Standards and the State of NAC

Opinion: Nobody can seem to agree whether network access control's mainstreaming or dying off, but I'm most interested in the standards work.

Hows NAC working for you? Are you even implementing or planning for it?

Depending on whom you talk to, youll hear that network access control is either dead or its being widely studied and implemented.

The good news story says that compliance, SOX in particular, is the main force driving implementations. This makes sense to me from the point of view that compliance is, in part, about getting control of your systems. NAC helps you get that control, or at least to say that youre making an organized and good-faith effort to get that control.

The counterarguments say, as reports such as a recent Forrester report did, that current NAC implementations are proving difficult and impractical. Joseph Tardo of Nevis Networks agrees with this point of view to a degree, saying that the key to NAC is the management of policy and that, to succeed, policy management has to be improved. It cant be the creation of a new island of policy.

/zimages/3/28571.gifOf course bots exist on corporate networks, but how big a problem are they? It could be that nobody knows. Click here to read more.

Theres also Jason Brooks viewpoint that NAC impedes the development and adoption of new technologies outside of the mainstream and cant give a level of confidence sufficient to generate the cost and work it entails.

As to how successful NAC is in the real world, Ill welcome your reports. Unlike Jason, Im rooting for NAC and I dont see why it cant do good enough. His concern over support for new devices is a reasonable one, but if things develop well, a NAC agent could be a simple thing to implement, and mandatory for market acceptance as a device driver.

Im thinking of standards development in the NAC space, and specifically the IETFs NEA (Network Endpoint Assessment) working group. The groups Overview and Requirements document states:

... network operators need a proactive mechanism to assess the state of systems joining or present on the network to determine their status relative to network compliance policies. For example, if a system is determined to be out of compliance because it is lacking proper defensive mechanisms such as firewalls, anti-virus software or the absence of critical security patches, there needs to be a way to safely repair (remediate) the system so that it can be subsequently trusted to join and operate on the network. The NEA technology strives to provide a mechanism to report the configuration of an endpoint for evaluation against network compliance policy. Such a mechanism could offer a useful tool for the network operators arsenal but should be recognized as not being a complete endpoint compliance solution in and of itself. Sounds like every responsible marketing document for a NAC product.

From the first time I heard about NAC, I realized that standards would be critical. Forcing customers to buy into one vendor whole hog was planning to fail. Effective NAC standards would allow anyone to implement NAC client agent software to work with any NAC back-end system. The vendors realize this and are happy to facilitate it, since the back ends are where all the money is.

The chairs of the working group are from Juniper and Cisco. The Requirements document was written by representatives of Symantec, Intel, Avaya, Cisco and Nevis Networks. In the mailing list discussions I saw someone from IBM, Paul Hoffman of the VPN Consortium and a developer of an open-source project. I dont see anyone from Microsoft yet, but theres every reason to believe theyll support any standards that emerge, especially since theyve already worked for interoperability with Ciscos NAC, and Cisco seems enthusiastic about NEA.

In the short term, theres no question that Jason is right that requiring an agent presents a barrier to entry for devices that are either unconventional or difficult to secure (for instance, Ive heard of companies that have eliminated wireless networks in part because of difficulties with NAC). In the longer term, I dont see why standards-compliant agents shouldnt be easy to write, almost template-driven.

The hard part, as Tardo says, is developing policy for them. But responsible security management will develop policy for such devices anyway. NAC helps to implement it, not the other way around.

Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983.

/zimages/3/28571.gifCheck out eWEEK.coms for the latest security news, reviews and analysis. And for insights on security coverage around the Web, take a look at Security Center Editor Larry Seltzers blog Cheap Hack

More from Larry Seltzer