Connect, Don't Reject.
ICW solutions make for effective networking in the health care sector.
Benchmarks - Glossary
Benchmark
A benchmark is an action done by a team of performance experts which has the aim to test out different system configurations under production conditions. It is usually done at a performance center at big hardware vendors. Its aim is to generate hardware scaling information in order to enable sales to scale system configurations appropriate to user demands. The second aim is to compare different hardware products of individual hardware vendors.
Benchmark script
Software which contains a sequence of commands for a load generation tool like Apache JMeter. It reflects that what an idealized user does on the system.
Benchmark suite
A collection of benchmark scripts which are needed to complete a benchmark.
Benchmark run
The action of taking the system under test, setting it under load and determining the maximal load when the system still stays stable. The maximal load is expressed in terms of virtual users concurrently on the system.
Benchmark result
A figure given in virtual users which is the maximum number of virtual users the system can support simultaneously on the system. Load driver A machine/server which is used to generate load of a system under test.In our case it contains a number of JMeter instances which in turn run the benchmark scripts.
Application server
A machine/server which contains a number of Tomcat instances which run the PHR application. The number of Tomcats per node depends on the hardware capabilities of the underlying machine.
Database server
A machine/server which runs the database application. Oracle was used for this with the Real Application Server (RAC) option. Response time Response time is the average time in seconds the system under test needs to complete a given HTTPS request. The response time is measured in the Apache httpd logs.
Virtual user
The load driver software JMeter can be configured for a given number of threads which run simultaneously in parallel. 100 virtual users then are 100 test scripts which run in parallel on JMeter. What a virtual user does is roughly equivalent to that what a real user would do.
System configuration
The way the test system is set up, as a result of the individual settings of all the components included (Load drivers, Application servers, Database servers and Load balancers).
System under test
The complete system seen as a whole during the test.
