The Request Log Analyzer measures the performance actually achieved by an MDEX Engine according to that engine's request logs. This "achieved performance" is completely dependent on the amount and nature of traffic sent to the MDEX Engine, and does not measure the capacity or upper bounds of performance that the MDEX Engine is actually capable of - the "potential performance".

For instance, consider an MDEX Engine running the reference sample wine dataset on a modern server and the latest release of Oracle Commerce Guided Search. This MDEX Engine is capable of handling well over 100 ops/second throughput at sub-second response times. Now consider that a single user leisurely clicks through the dataset, stopping to read about the descriptions and flavors of the featured wines. This user will generate a total of 45 requests over a 15 minute time span. When the Request Log Analyzer analyzes the logfile for this single user, it will report an achieved throughput of 0.05 operations per second (45 requests / (15 minutes * 60 seconds)), compared to the known potential throughput of 100+ operations per second.

Because of the possible large differences in achieved vs. potential performance, the Request Log Analyzer, as a standalone tool, is more suited to forensic analysis and behavior profiling than load testing and capacity planning. However, the Request Log Analyzer does work well in concert with other load testing tools such as Eneperf to analyze the performance characteristics of an MDEX Engine under load.


Copyright © Legal Notices