|
|
Optimize with a SATA RAID Storage Solution
Range of capacities as low as $1250 per TB. Ideal if you currently rely on servers/disks/JBODs
Page 2 of 2
From my perspective at least, many extensions to the framework readily suggest themselves. Top of the list for me would be to implement the design to run on the .Net Common Language Runtime (probably using C#, although any supported language could be used). This port of the framework would be very contentious—now we would be testing not only the hardware, but also how efficiently our chosen technology stack can drive that hardware.
The scope of pluggable test cases could be widened to include the most commonly encountered scenarios. Some obvious candidates include a Spring action and a database-intensive application. Or what about measuring the effect of AJAX-enabling your application?
Finally, I would like to add an online component to the framework, allowing developers to post results obtained on their hardware to a central repository for communal viewing and analysis.
To make your own comments about the framework, please use the discussion thread at the end of this article.
Benchmarking code is straightforward. Constructing and certifying an objective and fair benchmark is more difficult. Running a benchmark that is applicable to your application is even more difficult! Keeping a benchmark current and relevant requires ongoing work. In this article, I have presented a simple benchmark framework that separates two logically distinct requirements: how I run a benchmark versus what the benchmark actually tests. The source code for the benchmark framework is freely available from Resources along with guidelines on how to extend it.
Read more about Tools & Methods in JavaWorld's Tools & Methods section.