BSM is all about managing services. Services are made up of hardware and software. Let’s trace the service backward from the human.
Kathy is the human, she interacts with her computer and the software she is using. For this example, let us use a browser based application. Browser talks to the application server over a network and that server typically talks to a backend, like a database, over a network.
The End User
Let’s look at the typical tools we have to monitor this situation. End user management (EUM) tools acts like Kathy and simulate transactions. This will tell us when Kathy will see a problem.
Server management tools will validate the server(s) are working. If it is a more robust monitoring tool it can also manage the applications as they run on the server. I make this distinction because knowing that a database is running low on a particular resource is different than knowing how that impacts a user.
In small shops the operations team will probably know which servers run which applications. In larger shops, there will typically Configuration Management Database (CMDB) to map the application to its dependencies.
Network Management tools can identify problems in the network. This can be any number of things. Finding a down network node, or more subtle CPU or memory issues in a router. Maybe all the routers are fine and it is a congested link, the tools will find that also.
All is good?
Seems like we have everything covered. Well not really. Networks are more problematic than it seems at first look. My presumption here is a reasonably larger network that will automatically reroute.
Consider this scenario, your EUM software tells you there is an application problem. Your network management software is pointing to a number of problems. How do you determine which network problem is the one affecting the specific application you are troubleshooting?
The problem, most network management tools don’t track paths. If you know where the end user is and you know where the server is you still have to find which of the many links and many nodes are in the active path between the two.
In the simple diagram above if the active path is R1-R3-R4 and then errors on the R1-R2 link don’t mater to your investigation. Something to consider as you look at tools to help with BSM.