To be more precise, I didnt research all of BSM but roughly a third of the broader market that I characterize as BSM service impact (SI). The other major pieces are BSM financial optimization (FO) and BSM service lifecycle (SL). BSM SI targets the more operational side of BSM. With its service level management (SLM) roots it is, at least historically, the heart of BSM.
I took a top-down approach to cross-domain monitoring in terms of assimilation, analytics, change impact and business impact. I wasnt looking for vendors with the most probes or agents, but the best capabilities to assimilate and reconcile multiple sources, including multiple brands, for analysis. The critical requirement was a well defined capability for supporting the many constituencies, or roles, required for optimizing service performance in terms of business outcomes. These include executives and professionals both inside and outside of IT. I was also looking for integrated support for automation, even if the automation wasnt directly provided by the BSM vendor in question.
Here are a few of the more striking BSM SI revelations that emerged from this research:
When it comes to values and benefits, not surprisingly, BSM SI initiatives are especially strong in environments seeking to establish cohesion and consistency across multiple environments. An example is when multiple data centers are being integrated due to mergers and acquisitions. One deployment saved more than a million dollars in costs regarding service availability in the first year. This doesnt mean that all BSM initiatives need to be giant in scope in the initial phases, or only pertain to large companies, but that most are not green-field environments with nothing in place. Too many tool sets can be as much a problem as too few, and BSM can bring choice, focus and value by reconciling existing management investments into a more relevant, service impact context.
BSM SI initiatives like CMDB, typically require strong executive support because they categorically require cooperation across domains. This usually means you, the CIO! As an example, one deployment was limited because the networking organization refused to include its application-aware network monitoring tools, so that both applications and networking would have common ground for diagnostics. This networking group didnt want to share any of their data and claimed, with more than a touch of paranoia, that it would be misinterpreted if anyone looked at it other than themselves. The result was a sub-optimized BSM deployment.
From an operations perspective, many I spoke to for this research believed that a fully implemented BSM SI deployment requires two overarching systems (sometimes brought together in individual databases): An event management system and a system targeted at performance metrics. Each system can employ its own capabilities for analysis and automation across platforms and domains. In several ambitious deployments, these two systems were being brought together under a single unified dashboard, which was often not from a platform vendor, but from a smaller, third party innovator with a strong multi-brand focus.
Just as there seem to be two overarching flavors to real time monitoring emerging in the industry, there are multiple variations on application dependency. Once again, in a growing number of deployments I saw how classic, in-depth configuration dependencies are increasingly being complemented by more real-time, transaction-oriented views of application flows.
There is a lot of interest in bringing CMDB and configuration management systems (CMS) initiatives into more mainstream service impact initiatives. Generally, these more advanced deployments are looking for capabilities that could enable more automated updates to their CMS systems, while also leveraging the very real advantages of linking configuration changes more closely with service performance issues.
Related to this idea, I believe the BSM SI headline will be modeling. Not the kind you can see in reality shows on TV, but more precisely metadata. A surprising (if not quite perfect) unanimity arose among at least eight of the 13 vendors mentioned above that service modeling through metadata was their path to unifying real-time event and performance management systems with more classic CMDBs. And heres another surprise: many BSM software vendors are already at least beginning to do so but are afraid to tell you for fear of sounding arcane! This modeling includes real-time, or run-time updates to CIs and is versatile enough to be war-room ready.
Since this is under the covers, you might ask why you should care. The answer is that, if the industry steps up to the challenge of reconciling multiple sources of information as I believe it must, modeling and metadata will be at the heart of an expanded, federated CMS. This will redefine how you think about management investments and management solution integration. A modeled view of your universe allows you to link logical (e.g., problem owners, customers, etc.) CIs and CI attributes with physical ones, such as servers and software. Doing this with intelligence and planning can, in my experience, multiply the value of your investments in diagnostics and automation as much as tenfold. It also prepares you to move more closely to working with modeled business processed (where they exist) for true business alignment.
OK, but dont hold your breath on this last. With the exception of leading-edge financial services institutions mostly in Northern Europe (for some unknown reason), most organizations are not yet ready for their IT architects and visionaries to sit down with their business analysts and work out a more extended, integrated system. Nor, to be honest, have many vendors pushed very hard in this direction yet. Yes, there have been some automated capabilities, e.g., linking application transactions and reservations closed, as low hanging fruit, but, for the most part, the bigger picture of getting the business and IT to sit down at the same table beyond discussions of cost and rather rudimentary application prioritization and value is just not there yet.
Not surprisingly, most of the even more advanced BSM SI deployments I spoke to were still primarily focused within operations. Even bridging the gap that divides architects, engineers, and service planners from the multiple domains in operations is generally more a future than present reality, albeit one thats becoming increasingly easy to envision.
Nonetheless, I have become more encouraged than disheartened over the last five months. No one I spoke to, at least, was giving up on their strategic directions just to embrace tactical short-term cost cutting. And giving shape to BSM SI through some variation of an event management system and a performance (metrics) management system, combined ideally with integrated analytics and an overarching dashboard is a solid step towards clarity in BSM planning, if not quite a revelation in itself.