Microsoft's three-product approach to systems and network management—Systems Management Server (SMS), Application Center 2000, and Microsoft Operations Manager (MOM)—represents the company's first attempt at providing an enterprise systems management solution. The products take a soup-to-nuts approach to providing the tools for a comprehensive strategy that targets large-scale networks.
If your business plans include an enterprise-class management tool, make sure you compare the Microsoft options with the big three of systems management: IBM's Tivoli management products, Hewlett-Packard's (HP's) OpenView, and Computer Associates' (CA's) Unicenter. You might even toss BMC Software's PATROL into the fray as a reasonable contender. These products are all comprehensive, cross-platform management environments that can handle your worldwide management needs. Not only that, but they all suffer from the same drawback as well: Their history and development are straight from the UNIX world.
The Stage Is Set ...
A UNIX background is no handicap in network management. I find it likely that hardware (e.g., bridges, routers, switches) that uses some form of UNIX as its OS controls more than 99 percent of the Internet. TCP/IP comes from the UNIX world and was a relative latecomer to PC-based networks, needing first to displace Novell's IPX/SPX transport as king of the hill. But systems management requires an in-depth knowledge of the OS, particularly in an enterprise configuration, and that's where the Win32 folks really shine. The big management players will do their best, but somewhere in the back of your mind you're going to wonder whether Microsoft doesn't know more about managing computers with a Windows OS than a third-party vendor does. You can be certain that Microsoft will do its best to keep that idea in your mind; the company has already claimed that its goal is not to compete with the big systems management players but rather to provide the best possible Windows management tools that can integrate with the big consoles.
That claim might have some validity if it came anywhere but from Microsoft. The company didn't get where it is by providing pieces of other companies' solutions. Microsoft's practice has been to attempt to dominate whatever market it's elected to play in, and that business strategy has been successful for Redmond. The enterprise management business is a perfect target for Microsoft; no one will be able to accuse the company of being a monopoly in a market that includes active players the size of IBM, HP, and CA.
The Players Are Assembled ...
Curiously enough, none of the Microsoft management solutions are completely homegrown—all have been appropriated, to varying degrees, from third-party products. SMS is still at version 2.0 (the next version, code-named Topaz, is on deck), making it the senior statesman of the management line and the furthest from its third-party roots. Application Center 2000 has a lot of homegrown code, but the much touted load-balancing services were acquired from Valence Research's Convoy within the last few years. And the first version of MOM has changed little, although it's significantly cleaner, from its third-party roots (Microsoft based MOM in large part on the licensed version of NetIQ's Operations Manager).
When I spoke to Microsoft about the management solutions' base products, the company representative told me that the product team's first step is to make the products look like Microsoft applications. The next step is to build a significant Microsoft personality into the applications. I believe that eventually we'll see significant integration between the management products so that they're more clearly parts of a whole, rather than a grab bag of tools with a distinct Microsoft cast.
What this situation means in the short term is that enterprise management consoles from other vendors are likely to have more tightly integrated components than the Microsoft products have. If your management needs spread across multiple platforms, you're probably already using one of the non-Microsoft consoles, and convincing yourself or your managers that the current Microsoft solution is the right one for your enterprise could be difficult.
... And Microsoft Has a Long Way to Go
I recently visited a reader site that has a very large worldwide Windows 2000/Windows NT network. This company is in the process of migrating from NT 4.0 to Win2K and from Microsoft Exchange Server 5.5 to Exchange 2000 Server. To manage this global network with its hundreds of servers (close to 100 Exchange servers alone), the company depends on a variety of midrange systems management tools. The IT department plans to upgrade from this mix of tools to a comprehensive enterprise management solution, and the Microsoft products aren't on the list of possibilities.
I asked these folks about that omission because I knew they had considered NetIQ in the past. The response was that a complete, integrated, and—most important—battle-tested solution was what they needed. The department believed that the current incarnation of Microsoft's management solution wasn't up to the task, even though parts of the solution (e.g., SMS) were in wide deployment throughout the company. (Aside from networking infrastructure hardware, the company is almost a 100 percent Microsoft shop.) The last I heard from this company, a Tivoli product and Unicenter were in a dead heat for the enterprise management product of choice, with senior management pushing for Tivoli.
Until Microsoft achieves the integration that I know it's working toward between all its systems management products, the scenario I've just described is one I can imagine playing out on a regular basis. IT departments base their case to management for adopting enterprise management on cost reduction and improved service level agreement (SLA) support. Such cost control and improved support are usually effected through using one product—not by supporting multiple management consoles, as the current versions of the Microsoft management tools do.
But don't count Microsoft out. I would expect significant improvement in the company's management tools within the time frame during which the .NET My Services technologies begin to ship, over the next 18 months. If service-based computing is to succeed, an excellent set of tools for managing services and service providers is necessary. You can bet that Microsoft will be there to fill that need as soon as it appears.
