That should be an easy enough question to answer. Yet I suspect no matter how I define it, you or one of my colleagues will define it another way.
To me capacity is the ability to do “x” amount of work. That’s fine but…what is work? There is no standard “work” unit that means anything to every business. This leads to capacity being measured in MHz or MB. But then that tells the business (who don’t know if a MHz is a new games console, or what you get when you use a nail gun on your knee), next to nothing. Somehow we need to get from the resources we can measure, to a description that the business can understand.
To be frank, that’s not always made easy by either side.
ITIL describes three layers of Capacity Management: Resource, Service and Business. That’s not a bad place to start. We might even think of that as MHz, Transactions, Customers.
Measuring the resource utilisation, and trending it in to the future can be done with the most basic of tool sets. So resource measurement can be purchased off the shelf. Now, how do we turn that into Service/Transactions? This is where we step away from the resources and start to monitor the applications. And boy do some application vendors make that hard work!
At this point no doubt someone reading this is thinking that their database reports transactions in a nice clear easy format. Yes it does. But those are database transactions. What we are looking for here are more “Service” transactions, e.g. I transfer money between my accounts while logged in to my bank’s website. To do this I’ll have interfaced with several services, and generated tens of database transactions before I even get to transferring the money. Yet as a customer I will feel I’ve only done two things. Logged on, and transferred cash.
It can be hard identifying these transactions in a complex environment where one transaction may be spread across multiple servers and databases.
It’s not impossible, and with a little effort these types of transactions can be recorded and plotted against resource utilisations of the relevant servers. However it relies on accurate service maps and someone maintaining the relationship.
This is where our new partner Correlsense and their Sharepath software come in to play. Sharepath identifies each user interaction (without any code changes to the applications), and automatically records where the business transaction generates work in the environment. Metron’s Athene software will then provide the resource utilisations and automatically generate reports combining both sets of data - reports that contain information the business can understand. Now we are talking about the maximum number of logons, money transfers, or movie downloads, or shopping carts that can be processed, and if the trend says they’ll hit that number, it’ll tell them when.
What’s your definition of capacity?
http://www.metron-athene.com/documents/factsheets/published_paper/correlsense_and_metron_transaction-based_capacity_planning.pdf
Phil Bell
Pre-sales Consultant
No comments:
Post a Comment