Are IT managers still looking for MOM?
Management software, like most of high tech, is not only driven by technology, products, marketing, IT requirements and business needs - its also driven by fashion.
Fashion is in turn driven by the media, which require change and conflict, and encourage short-term memory. Its no fun if fashions move at a glacial pace; we tend to expect them to change on a yearly basis, if not more frequently.
The result is that whats fashionable is often not yet real, and whats real is no longer fashionable. For example, the blush on Javas rose faded several years ago - at just about the time it was becoming entrenched as a real differentiator, and meaningful management solutions were beginning to deploy it. Service-level management came and went and is coming back again as "service assurance." The term "QoS" (quality of service) has been rebounding from acronymic limbo for years and is now used to describe something much more generic - although no one has so far been able to explain to me exactly what that is. The "e-" prefix to "business" and "infrastructure" management was applied last year to reflect "all business" and "all infrastructure" and is now selectively being dropped - at a time when, in reality, the need to manage Web-based business is more relevant than ever.
The truth is that unlike cars, and clothing styles for very thin women, management software itself doesnt go through the annual fashion transformations that its marketing might suggest. While many companies come and go in months or very few years, many of the real forces behind user adoption need to be understood in terms of longer periods; in some cases, they may even span more than a decade. This may seem heretical in a market climate in which its hard to see the future even two days out, but it might also be viewed as reassuring.
I was especially struck by this phenomenon recently when in a discussion among 15 or so IT professionals, the term "manager of managers" (MOM) received a fairly positive reception. What does MOM mean? About 10 years ago the term referred to an architecture where a single management system could unite different management systems for centralized control. At that time, the mainframe world was still viewed as holding the keys for MOM deployment. The advantages were superior infrastructure control and business efficiency - the Holy Grail of management. The problems were that the idea was a schema, not a reality.
The term became unfashionable at least five years ago when - from a marketing perspective - no one wanted to call themselves a "manager of managers." The terms "platforms" and "frameworks" took over and then suffered their own twists of fate at the hands of fashion and, frankly, reality. I am in no way suggesting that its time to bring back MOM as banner for tradeshows, headlines and lead sentences in press releases. But I do get the feeling that we are about to enter the period in management software when the advantages of a MOM can begin to be meaningfully achieved.
Why is this? First of all, MOM hasnt gone away. Tivolis Enterprise Console and its Business Systems Manager may not be easy to implement, but they do provide a MOM, in concept. Aprismas Spectrum and RiverSofts OpenRiver reflect new architectural approaches (even if Spectrum is in fact more than a decade old itself) to integrating management applications into a core intelligent base. And single prroducts are evolving towards suites, often with the potential for sharing common databases. Witness Micromuse in its current evolution. Even in the performance management market (or what used to be the performance management market and is now becoming something broader), there is consolidation that has a MOM-like feel. Witness Concords capabilities for real-time management and NetScouts step-by-step integration of NextPoint. Frameworks, such as OpenView from Hewlett-Packard and TND from Computer Associates, are also reengineering themselves to enable MOM-like capabilities.
What do these have in common that the past "manager of managers" vision didnt? Rather than being a strict play for centralization, modern MOMs allow distributed management deployments - in case you want to manage your routers from Chicago and your applications from Gainesville - but still profit from knowledge of the interdependencies that can affect the performance of both. More importantly, modern MOMs will offer real integration on the one hand, and a less grandiose vision on the other. The two suggest a meeting place somewhere in the next few years where users and vendors might meet with some degree of satisfied expectation.