InformationWeek (IW) just published a February special issue on application performance management (APM). You’ll want to read it.
The highlights will be familiar to anyone who follows IT Ops: system complexity makes APM simultaneously more difficult and more valuable; increasingly-virtualized infrastructure elevates the importance of APM; and at least part of the attention executives gave to APM a few years ago has turned now to such subjects as cloud, mobile and social.
IW has released parts of this material already; just last month, for instance, I praised “A Thoughtful Report on the Future of APM“. For those who prefer a consolidated offering, though, this month’s Special Issue brings all IW‘s research together in a single digestible package.
I summarize the best parts this way: while legacy APM which assumes “manual” configuration and modelling is simply too expensive and unscalable to survive, “smart” APM is more essential than ever. The best modern APM is more tightly-focused than before, on applications where it can make a difference, and knows how to do more on its own both in regard to a variety of infrastructure elements, and in “self-configuration” to adjust to changes in those elements. This kind of APM manages, not just monitors.
Virtualization an apt complementary subject
A useful current complement to the APM “Special Issue” is Dick Csaplar’s “Optimizing Business Performance: Why Every C-level Manager Should Care About Virtualization“. While a bit too high-level at times–real-world virtualization doesn’t “simplify” quite as potently as his rosy description has it, for example–Csaplar is on-target in arguing that we’re past the point where virtualization’s return is too “uncertain” to risk. He also makes the right choice in explaining the principal categories of virtualization now as server, desktop, and storage: software-defined networking remains too new to qualify among the “leading forms of virtualization deployed in datacenters today”.
IW‘s definitive coverage of APM strategics settles that topic for at least a few months, and gives me the opportunity to return to small-grained tactics of devops and performance management. Over the next week, I’ll follow up with tips on database governance, the place of “social”, and the importance of subjective technique.