Standard Commercial Workload Model for J2EE

team bbl


Currently, there is only one accepted industry-wide standard benchmark application for J2EE-compliant application servers. The benchmark, SPECjAppServer, is from the SPEC organization and is the successor to ECPerf, which was the premier benchmark application in 2000. The SPECjAppServer application emulates a heavyweight manufacturing, supply chain management, and order/inventory system. Because the goal of SPECjAppServer is to benchmark the EJB container, it uses a thin Java client to send requests via RMI/IIOP instead of HTTP requests to the web server.

In 2001, SPECjAppServer2001 was developed in a concerted effort by companies such as IBM, BEA, and Oracle. SPECjAppServer2001 is based on the J2EE 1.2 specifications. Currently, a new benchmark application called SPECappPlatform is being developed that is designed to measure the scalability and performance of enterprise platforms that are based on J2EE and .NET with web services. SPECappPlatform is a superset of SPECjAppServer. SPECjAppServer2002, which was released in September 2002, is similar to SPECjAppServer2001 except that the SPECjAppServer2002 is based on the J2EE 1.3 specification. More specifically, SPECjAppServer2002 uses EJB 2.0 instead of EJB 1.1, which was used in J2EE 1.2.

The SPECjAppServer2002 benchmark consists of order and manufacturing applications. The throughput of this benchmark is directly related to the load (injection rate) used by these applications. Hence, the injection rate needs to be increased to scale up the throughput. The metric Total Operations Per Second (TOPS) is the average number of successful total operations per second completed during the measurement interval.

TOPS is composed of the total number of business transactions completed in the customer domain added to the total number of work orders completed in the manufacturing domain, normalized per second. The benchmark also requires a number of rows to be populated in the various tables. The cardinality of the customer orders and supplier tables scale based on the load (which is influenced by the injection rate). Therefore, as the injection rate is increased, the database grows larger.

A number of submissions from different vendors showcase the performance of their application servers using the SPECjAppServer standard workload. There are several categories, depending on how the system was configured (for example, dual node and distributed). Information about SPECjAppServer and submission results is available publicly at http://www.spec.org/jAppServer2002.

    team bbl



    Performance Tuning for Linux Servers
    Performance Tuning for Linux Servers
    ISBN: 0137136285
    EAN: 2147483647
    Year: 2006
    Pages: 254

    flylib.com © 2008-2017.
    If you may any questions please contact us: flylib@qtcs.net