As with everything in technology, the needs around DB & Application Performance Management continue to rapidly evolve, which can leave production support teams in the dark if they don’t keep up. To illustrate, I visited a large financial services customer two days ago, who recently recognized that they took their eyes off of performance management tooling for the past couple of years, and consequently now find themselves in a deep hole with respect to their performance management capabilities relative to their competitors. So when we talk about “Back to Basics” subjects around technology performance management, I believe it is important to describe what the “new basic” is.
Not long ago the DB & App platform vendors provided next to nothing in terms of native performance management capabilities. This was true for management at both the basic underlying platform infrastructure level, and the more advanced transactional and code level running inside and through those platforms. In recent years however, some platform vendors like Oracle have improved their native platform performance management capabilities, which is a natural evolution for most platforms; the end result is commoditization of some of the homogeneous platform infrastructure performance management capabilities. Combine this with performance management complexity increases from expanded virtualization of platform infrastructures, increased heterogeneous mix of platforms in production environments, more widely distributed Web 2.0 application architectures, including edge components, and mobility, and suddently the new DB & Applicaiton Performance Management "basic" expectation has now moved considerably upward for top performance management professionals.
It is no longer sufficient to manage performance from a purely infrastructure perspective alone. Application, transaction and end-user experience correlation to that infrastructure is key. Fortunately, modern instrumentation techniques enable unprecedented visibility into these performance details for mainstream platforms. Does anyone on this blog have an outsourcing arrangement, where some large 3rd party is responsible for the performance management of the infrastructure underneath your applicaiton(s), while separately you are responsible for the performance of the application(s) and/or platforms running on that outsourced infrastructure? Clearly this has been a particular challenge for many IT teams that I have worked with. Since IaaS and PaaS are essentially just self-service mechanisims for this same sort of outsourcing, do you think the performance managemnet challenge here will improve or worsen with IaaS/PaaS? I'm curious to hear people's thoughts here.