Skip to content

TechEd '07: Day 1 Keynote; Applying more Agility to developing Agile IT Management

Not much too report from the keynote. There was an actually entertaining film with “Doc Brown” from Back to the Future about several, uh, failed Microsoft visions of the past, e.g., Hailstorm. The point here was that the keynote was to contain little “vision” and more technology and concrete grapple-hooks.

Well, for me, that was the high-point.

To be fair, if you haven’t been spoon-feed Microsoft’s IT vision over the past year as I have (thanks!), the keynote was a good summary of that vision. It was concise, clearly articulated, yet packed. The Dynamic Systems Initiative, the maturity model, making IT agile so that they can respond quickly to new business needs rather than being locked into technology.

I couldn’t tell if the audience was too enthuesed. Perhaps it was just because, unlike usual, I was sitting up in the very front rows, wrangled into the “analyst” area, but this audience seemed the most disinterested I’ve seen in a while. At one point, during the Silverlight segment, we saw 10 videos playing at once over another video. “Isn’t this cool?” the demo’er said in gestures and words. There was, really, no response from the audience. Perhaps they were trying to figure out how that’d help setup Exchange clusters.

The Microsoft IT Vision

While there were demos of some good stuff in System Center — a dip into the idea of model driven IT management (Zenoss should take a look at all that as competitive background) — and even some point-and-click programming for reports (that sales guy is so fired!), the keynote was largely a summing up of Microsoft’s vision. As I’ve said in the past, there’s nothing wrong with that vision. Indeed, it’s a nicely done culture driven ecosystem.

In the realm of IT management, what I feel is missing here is a clear articulation of what that life looks like. Much of the demos revolve around taking action once the problem is identified rather than how they hell you figure out the problem in the first place. While the vision is nifty, it’s sort of what the vision for IT has been for awhile: using knowing what you have, hammering down processing, business goal-based planning and culture, and interop among all the IT. This isn’t to say that the IT management world has delivered on that vision. But that’s the point: at this point, we need to see the running code, documented practices, and overall understanding of what the culture of IT management should be.

Agile IT Management

Pulling back, Microsoft actually has a “lower level” vision than many other IT people. Their articulation of what it means to be “agile IT” is somewhat novel in that most large IT management vendors have been obsessed with the C-level stories of late (that said, encouragingly, BMC has re-learned the word “sysadmin” recently).

The imperative here is then to show us to code, solutions, and the supporting material. As someone sitting next to me said, Microsoft has “discovered” all sorts of technology theory we’ve had for awhile — SOA, SaaS, IT modeling, oh my! They’re at the level that their IT-land competitors are in understanding what people want. What they need now is to deliver the goods. Otherwise, it’s death by road-map.

The looming irony here is that we’re in year 4 of Microsoft’s 10 year IT management plan, a plan which all about agility. My ears lit up when I heard that part of agility was doing more frequent, smaller releases. I can’t help but think that a 10 year plan to get to an agile IT management culture is sort of, well, weird.


One thing in particular makes for a good example: SML and SDM. I’ve mentioned the SML world many times, and for good reason: we have no good domain model in IT management. The last successful one we have is MIB based SNMPs, which are sort of fun if you like self-flagilation (read: are a programmer) but not really what we’d like to have. While SML and SDM look to be baroque XML, it’s a nobel effort. What’s depressing for me is how slow and how opaque the process seems to be going. Technically, modeling IT shouldn’t be too difficult of a process, esp. if done in an Agile fashion.

Standards bodies have had junky success with developing IT management domain models in the past, so I’m not too sure the traditional standards development practices would actually help here. But, I do feel that for wanting to make an industry wide domain model, Microsoft is being too closed. Where are the wikis? The blog posts? The email lists? The public discussions about it? When I ask IT management vendor architects and developers what they think of SML and SDM the response I usually get is “whu?” to, at best, “has that come to anything yet?”

The point here is: in IT management, creating a widely accepted domain model requires engagement not only with key partners, but everyone involved in the production of IT management software, up and down the corporate hierarchy, across friends and enemies. Dare I even suggest it requires being open source-y?

Ad hoc/Pragmatic Standards Development

To stop throwing rocks a bit, I’d suggest looking at the way OpenID has been evolved and spread over the past 2 years. Even better, check out how microformats have evolved and spread. Both of those are tackling problems that no vendor or clutch of vendors has really solved and having early success. Both are done in the clear, rapidly, and, sure, with all the bickering and snarling that comes with developing any technology. But, they’re out there working right now. Broadening from OpenID to identity in general, Microsoft’s involvement in the “Identity 2.0” space of late is another pool of experience to grab from.

No Shackles Here

When it comes to competing in the wider IT management space, Microsoft doesn’t need to be hampered by the shackles of success. Getting the Big 4 to deliver on The New New Thing in IT management involves all sorts of bureaucratic knife-fighting, freaked out legions of lay-off jittery developers, and otherwise impossibly annoying cultural wars needed to bend the will of those organizations to move from legacy to Systems Management 2.0. The Innovator’s Dilemma they call it.

Microsoft, on the other hand, doesn’t have the comprehensive, decade-old, elephantine revenue pumping legacy-ware that Big 4 vendors do. It can draw on this freshness to actually be agile in developing it’s agile IT management vision. Now, Microsoft is still chained to a desktop-centric view of the world in much of their thinking: they’ve got a big SOA and SaaS pill to swallow at some point that’s gonna be a bitter one for Windows and Office to absorb. But, in the world of IT management, there’s still the chance to work unchained. As ever, the story of PowerShell is illustrative for this: a command-line for Windows? How’d that happen? It happened because the IT management

So far, it doesn’t feel too agile: it feels like the plan and technology are being developed behind closed doors, on long multi-year plans. That’s been Microsoft’s favored way to develop technologies, growing them under the Redmond heat-lamps and then throwing them over the wall to the public. I’m not sure that method will work this time for IT management. It’d be great to see the process leap ahead of the same-old, same-old in IT management, open up, and start delivering at the pace required to be successful.

Why would it be great? I actually like Microsoft’s IT management vision; as James said, their UI always seems weird to me, but I’m no Windows guy. It avoid the burned-out cynicism that comes with being in IT management too long, it’s low-level instead of exclusively focused on “business IT alignment dashboard enablement hoopla,” and sort of knows what the goals of IT are in relation to making money for The Business. My concern here is more around getting that it all done, out the door, than the vision itself.

Below are my raw-notes from the keynote:


Disclaimer: Microsoft is a client and paid travel and expenses for TechEd. Zenoss is a client, as is BMC.

Technorati Tags: , , ,

Categories: Companies, Conferences, Systems Management.