About testing from the cloud
Posted: Mon Oct 31, 2011 2:34 pmI see all these articles about testing from the cloud. Does that really make any difference and if so, what difference?
In most cases the use of a "cloud" resource will produce better load testing results, and equally good monitoring test results, compared with doing the same work on your "in house" laboratory computers.leonelj wrote:I see all these articles about testing from the cloud. Does that really make any difference and if so, what difference?
For load testing the main reaons is bandwidth. Typical machines [machine images] that you can get from cloud service providers have very wide input/output (I/O) capacty -- usually 1 Gbps and possibly more. That means that even if you have 1,000's of browser users (BUs) running the chance that you saturate the I/O capacity is very small. You definitely want -- or should want -- low I/O utilization relative to capacity because that means that your tests are more realistic.
It's about a break-even for monitoring. You have the high bandwidth factor just mentioned, but for typical monitoring situation you are not going to be running multiple browser users (BUs), so you have to balance that with the overall difficulty of dealing with an "abstract machine" compared to one in your lab that you can actually put your hands on!
eValid Performance Team