In the land of IT, guesswork is bad. If you are guessing at what is going on with your IT application portfolio, those applications may punish you by randomly breaking down. A new APM (Application Performance Monitoring) industry has grown up to help figure out what is running where, how much traffic it generates, how much resource it consumes and how long you’ve got before you run out of IT space.
What’s happening in the computer room?
The migration from on-premise to cloud is picking up the pace as more enterprise leaders get fed up with running computer rooms. On-premise hardware and software are expensive to maintain in terms of time, money and expertise. Before you can migrate to the cloud, you have to get to grips with what is happening on-premise.
Picture this completely fictitious data center (there is no set of “on-premise IT” average figures that I know of). A company runs twenty racks full of machines, helping with all aspects of the enterprise. These machines host hundreds of apps handling everything from VoIP to storage. The network traffic shifts trivial customer brochures and business critical workflow. This collection of applications has built up over many years and there is no one person with a solid grasp of what is going on in there.
You are the person picked to get a grip on what is going on in there. What do you do? How do you get a handle on a data center like this? Now imagine you are part of a globally dispersed organization and every building has a computer room like this. Now you must multiply the number of applications and throw in an extra dollop of network complexity. Good luck.
APM (Application Performance Monitoring) is good for discovery, monitoring, SLAs, operations, etc., etc., etc.
An APM (Application Performance Monitoring) product can figure out what is going on. It copies all the network traffic, analyses it to figure out what applications are in there, what the conversations look like, and how it performs on an average day. All sorts of monitoring devices are used to make these measurements, from agents installed on customer systems to appliances plugged into network switches.
An APM continuously monitors performance. The best way to measure application performance is to measure the end-user perception for B2C apps and to measure SLA targets for B2B apps.
Once the APM application knows what an average day looks like, it can spot unusual behaviour and raise the alarm. APM products are often bundled with dashboards for operations staff to view. If you have ever seen a control panel containing rows and rows of real-time graphs, it’s not necessarily a banker’s stock-market dealings – it could be an APM application shining a light on enterprise application performance.
As you can imagine from an application that must keep track of all applications in all locations of an enterprise, an APM product can be a huge and complex beast that only experts can drive.
APM is a new TLA, but it’s not "the" TLA
APM has been around for a few years and it’s picking up steam, but APM is not an analyst's darling. As with all new trends, product vendors are doing their best to create what they believe are APM applications because there is no widespread industry standard of what APM is. If a monitoring tool has something to do with applications and performance, you can bet the marketing guys have stuck an APM sticker on it.
Interop Las Vegas 2013, the big IT conference, holds Best of Interop competitions in various categories. SDN was the big TLA (Three Letter Acronym) contender at the show this year, but APM was also there. Thirty APM vendors entered their products in the Management and Monitoring category and the winner was ScienceLogic.
Nick Hardiman builds and maintains the infrastructure required to run Internet services. Nick deals with the lower layers of the Internet - the machines, networks, operating systems, and applications. Nick's job stops there, and he hands over to the designers and developers who build the top layer that customers use.