The changing face of network management

With the announced acquisition of Micromuse by IBM in December, I was inspired to go back to a five-year-old report by EMA on the network-centric root cause and topology market. In it, EMA examined 18 companies with products designed to support topology, discovery and root cause analysis.

EMA believed then that this market held significant value as a potential integration point for other solutions based on its ability to capture, and in many cases actually model, infrastructure relationships, as well as the more obvious capabilities for analyzing faults. Many of these vendors also addressed requirements well beyond fault management to support inventory management, service impact analysis, and in some cases basic performance information. And some, through partnership, or through deliberate internal design, supported fault resolution beyond the first three layers of the network stack – to address application-to-network interdependencies.

It shouldn’t be hard to remember that, not all that long ago in the early 1990s, the “platform wars” were explicitly about just this capability – albeit in originally a far more primitive form. Layer-3 auto-discovery with some level of rules-based or other type of correlation was then the battleground for platform pre-eminence. Since then, it shifted away towards systems management, and then service management, and now, arguably, is moving towards positioning around a core configuration management database (CMDB). These last two trends feed back to the original base logic, albeit with huge amounts of added innovation (happily) and complexity (sadly) facing IT buyers.

So what happened to this group of 18 vendors? Well, quite surprisingly, only two have vanished (one after an acquisition and the other without being acquired). Four are still around as small companies – some larger and some smaller than they were then. Four were and are the frameworks that we all recognize today. And seven have been acquired. Among these seven, there are those that have been acquired multiple times (e.g. RiverSoft to Micromuse to IBM and Aprisma to Concord to CA).

Since then, there have been new entrants, although very few purely in the classic model of topology-based root cause. Most of the new entrants are targeted at either new types of analytics, and or full infrastructure vs. network-centric management, or combine engineering and traffic analysis with fault management. One recent acquisition – too new to be included in the original report but in many respects faithful to this model – was Sheer Networks acquired by Cisco last year, potentially to become an integration platform much in the manner envisioned in EMA’s 2000 report.

And while these results were interesting to me – no doubt the even more interesting question is why did this happen and what does it mean for network management, and for the IT management marketplace overall? Well, perhaps the most obvious (or superficial) thing to say is that it represents the implosion of a once great market, one that EMA felt at the time would be a significant catalyst in shaping the IT management marketplace overall. But a more careful analysis would be to look at this as a natural arc for maturation in a still significant and vital area. The gobbling up of major network management players in this vein is arguably as much a testament to their success and relevance as it is to their failures to become empires in their own right. By 2000, systems management had long established itself as the pre-eminent platform focus, so even then EMA viewed the significance of this market as “foundational” and “catalytic” more than the beginning of so many new frameworks in the making.

Most importantly, what does this mean for IT management and network management in the present and future? A couple of points come to mind:

* The potential “integration platform” like the role of these solutions hasn’t vanished, and in one case (Cisco) remains highly visible. For the most part, however, these solutions are being fed into larger management portfolios as a series of foundational technologies ranging from discovery and inventory building, to relationship modeling, to correlative engines that in some cases are being applied to infrastructure analytics well beyond the network space. At the same time, it would be wrong not to acknowledge that many of these acquisitions represent clear stakes in the ground to say, “Yes, we have world-class network management,” and in some cases, “Yes, we are now viable going after the telecommunications marketplace.”

* Does this herald the end of network management as a separate market? Hardly. The huge volume of innovators focused on network management from a wide variety of perspectives – from flow-based optimization and performance management, to configuration management, to service mapping, to effective, modular engines for inventory and resource management, as well as fault – show on balance market vitality ahead of IT management overall. But the shape of the market has clearly changed away from its original form towards a much more complex matrix of solution sets that go beyond traditional definitions. Cisco’s entrance into this market will also change its complexion, and is more likely to make it richer and more vital than narrow existing options.

* On the other hand, it is probably not an exaggeration to say that “network management” as traditionally defined ceased from being a meaningful category several years ago. This is not because network management in itself was losing in relevance, but because market innovators have stretched the boundaries of network management towards a more complete view of the infrastructure and its application services. Network management is now really a series of design points for managing a changing infrastructure to support the effective delivery of application services. This is an area rich in innovation, and one with many years left for growth.