20.4 Typical performance characteristics of a WebSphere Commerce site
Performance characteristics of a typical WebSphere Commerce site.
At a high level, the trends depicted here can arguably apply to most well-performing eCommerce sites, although the scales of these graphs and some other details will differ.
For the sake of our discussion, we assume the workload to be WebSphere Commerce application tier intensive and that the other tiers are able to handle it with relative ease. For example, we assume that the database server gets only a fraction of the overall workload due to a high Dynamic Cache hit ratio and that it is able to handle increasing workload without consuming system resources, such as CPU, with as much intensity as WebSphere Commerce tier would. For such a system, the typical performance behavior of a site should be as depicted in Figure 20-3 will be referred throughout the remaining portion of this chapter.
WebSphere Commerce benefits from the robust underlying scalability provided by WebSphere Application Server. In our testing we find that once the site reaches its maximum throughput (TM), it is able to maintain that throughput even with many-folds increase in workload, with only slight degradation to throughput. That is, the sites held well under stress. However, eventually, at a certain breaking-point, TX, it is not able to maintain maximum throughput, and then throughput comes down and suddenly we start seeing all sorts of concurrency and time-out errors.
Although maximum throughput (TM) may be maintained for a much larger volume of the workload, maximum business capacity (TB) may not. A system's maximum business capacity is defined by its business requirements. Specifically, in Figure 20-3, any excess capacity (for example, throughput) that may be available past the limit defined by maximum business requirements (for example, maximum response time allowed, which is RB in our example) cannot be counted towards business capacity available.