The W2K problem

Yes, what you don’t know about Microsoft’s Net OS can hurt you.

Convergence, managed services, broadband remote access. The last thing today’s communications managers need to do is get involved with an enterprise OS deci-
sion, right?

Wrong.

Windows 2000 is more than just a new OS. It’s more than just a new NOS. It’s the foundation of Microsoft’s strategy to remain at the center of networked e-commerce technology. And it may ultimately have more of an impact on the network than it does on traditional core IT disciplines, such as systems administration and application design.

To start with, according to my research-and-benchmarking friends at The Tolly Group, W2K implements over 60 Internet Engineering Task Force RFCs. These RFCs involve everything from QoS controls to dynamic domain name services. Even something as apparently simple as W2K’s mechanisms for responding to “dead” TCP gateways can have an impact on how your applications behave over the network. All kinds of timing and retransmission issues can pop up as W2K tries its best to route around what it senses as network dead-ends. Is that really the kind of thing you want to have catch you by surprise?

W2K also features native support for IPsec. This potentially gives your end-users the ability to start throwing triple-DES encrypted traffic around your network. Still think W2K is just the sysadmin team’s baby?

THE YEAR 1 A.D.

Okay, now for the biggie: Active Directory. AD represents a fundamental change in how enterprises (especially committed Windows shops) manage their environments. Today, enterprise management under Windows NT is very fragmented. That’s because, for years, NT forced IT staff to manage their resources on a domain-by-domain basis. With AD, Microsoft is finally offering customers what Novell’s NDS solution has provided since the mid-90s—a “holistic” tree-based method for classifying and administering resources across the entire enterprise. This is a great productivity-enhancer for server and PC administrators, who can now apply policies about configuration and application access across large numbers of like users.

AD is also going to be a lot of work. It requires IT departments to rethink how they define and classify every component of their environments: users, PCs, servers, applications, etc. This investment in rethinking—and the subsequent design and programming of AD—is something that they’re going to want to amortize over every aspect of information infrastructure. That includes the network. Everything is going to work off of AD. Network address spaces. E-mail directories. Application access. VPNs and virtual LANs. Everything.

I assume you’re getting the picture.

Naturally, my advice to you is not to get blindsided by the migration. In an ideal organization, your top IT manager will make sure that communications people are on the W2K eval-and-planning team. If this isn’t the case, make a big stink right now. Even if it is, don’t be complacent. It’s a mistake to ride along on the process and just pipe up with a networking question or two as things progress. That makes your interests look secondary. It also gives the systems folk a chance to look smarter than you. Don’t give it to them! Instead, go to Microsoft’s site right now and check on topics like TCP implementation and that “dead gateway” issue I mentioned before. Hit the systems people hard with those questions right away and put them on the defensive before they have a chance to do their homework. That way, they won’t be able to minimize your concerns. They’ll also be less able to minimize the potential drain on your resources that a W2K migration will create.

PAIN AND GAIN

There’s more at stake here than just accommodating a new operating system. The very fact that a PC operating system warrants discussion in a communications publication says a lot about what’s really going on in our industry. For years, I’ve been ranting and raving about the dissolving boundaries between systems, data network, telecom, and application administration. W2K is a major milestone in the progress of that dissolution. As a result, it should serve as another wake-up call to communications professionals who still believe that ownership of the PBX or the firewall somehow translates into job security or an optimal career path.

It doesn’t. Networkers who don’t actively press for control of the total computing-and-communications environment are going to lose out to systems and applications tribes that are closer to the business units and their management. The real convergence, as I continue to preach, isn’t between data and voice; it’s between application development and security and storage and telecom and Web site provisioning.

Of course, this organizational/disciplinary convergence isn’t actually called “convergence.” It’s called “e-commerce.” Networking teams should own this space, but they don’t. They’re being over-run by Web developers and—even worse—outside service providers like Web hosters. W2K represents one more battleground where you’re either going to sit at the table with the strategists or get treated as mere custodians of the plumbing.

Don’t get blindsided by W2K. Do your research and develop an agenda for any migration discussion. Initiate that discussion if you can. W2K is going to be a big pain no matter what your company does. The only question is whether you’re going to have anything to show for your trouble at the end of the game.