Contributed by Anup Surendran
In the pursuit of solutions
Performance, Scalability and Reliability are three non functional attributes every Architect deals with when they work on enterprise systems or large audience applications. I had been through a recent contract search and I found that different aspects of these were asked at different times within different interviews.
Let's look at what these mean :
Throughput - When you don't deal with clients directly, it is the amount of work your department gets done in an hours time. Basically this is your productivity measure.
For example, in the Supply Chain Management, the resources are consumed by the operations at a 'node'. However, resources have a limited capacity. For example, if it takes one hour to assemble a car on an assembly line, then you can produce only eight cars in an eight-hour workshift. This can constrain the flow of goods and services through that node in the network. This is called throughput. Constraint based planning is a big thing in this industry.Performance
Performance goals will differ based on the use cases which the system needs to support. If there is an underlying application technology some of the objectives can be measured by the following factors :
Concurrency, Throughput, Server Response time and Render Response time. To read more about these go to http://en.wikipedia.org/wiki/Software_performance_testing .
What Performance measures apply to Business ?
The 4 equivalent measures which applies to Business are Concurrency, Throughput, Business Unit Response Time and Customer Servicing Time.
Concurrency - It is the ability of a business to serve multiple clients at the same time for a certain product / service.
Business Unit Response Time - The business equivalent of Server Response Time. Each Business function's which is performed by the Business Unit should be measured as a black box.
Customer Servicing Time - This is the business equivalent of Render Response Time. A Customer request is serviced by different business units. It is important to measure the combined servicing time for one Customer Service Request.
More on Scalability and Reliability in the next post.
Contributed by Anup Surendran
Works as a Solution Architect in Canada's largest Retail Company.
Loves aggregating topics at Paper.io .
Follower of OnBliss principles.
Theme created for Posterous by Obox
