While those of us in server and network infrastructure may sometimes look down on our developer colleagues because, after all, without us their applications wouldn’t have anything to run on, in truth it’s really all about the app. Without applications, there wouldn’t be a need for our servers, desktops, or networks.
Alas, we continue to monitor our networks and server loads without taking much time to consider application performance. It’s not our fault -- getting insight into the performance of custom applications isn’t particularly easy.
However, with System Center Operations Manager 2012, we can now get that deep view into the applications that run on top of our servers and the underlying infrastructure. We can now monitor .NET applications on Windows servers as well as Java applications on WebSphere, WebLogic, JBOSS and Tomcat and have real insight into what’s happening with the application and where problems might be. Application monitoring management packs allow us to monitor distributed applications, allowing us to see availability and performance throughout the corporate environment and across cloud-based resources.
Really – who doesn’t love the call from an end user that complains: “The capex program is running slow!”. Is it the web frontend that’s loading slow, or is the SQL server queries that are the problem? Instead of tossing the problem over the cubicle wall to the capex application development team to figure out, we can work with them to resolve the problem faster and more efficiently because we have insight into both the road and the car running on it.
Below is a presentation delivered last year at TechEd. It’s based on the beta of Operations Manager 2012 as System Center hadn’t been released yet but the information is still quite relevant. If you want to jump ahead, they start talking specifically about application management at around 15 minutes into it.
I think the shift to app-centric infrastructure is just beginning. The promise of virtualization is starting to yield some value, and now we can start shifting focus up the stack. As C-levels start to focus more and more on ROA, we will eventually end up in a meeting room with slides showing infrastructure and platform efficiencies at all time highs, and we'll turn collectively to the developers for more efficient apps. Tooling like SCOM will definitely help us get to that point, but I think what's baking my noodle is - does the final stage of evolution have to happen as a consequence of natural evolution, or is there a possibility that we can jumpstart that and get infrastructure and applications people working together now to build the next generation data centre?