Why would anybody buy LoadRunner when there is JMeter? Say the users are a little technical (actually they are developers), so they can use the JMeter to test their web application - right?
Why would anybody buy LoadRunner when there is JMeter? Say the users are a little technical (actually they are developers), so they can use the JMeter to test their web application - right?
There is an efficiency question at play here and it assumes that JMETER is optimally efficient as compared to the protocol interfaces in LoadRunner. This may not be true for all cases. There is also an assumption here that JMETER's monitoring and analysis components are equal to that of commercial tools, which in turn implies that both types of tools are equally efficient in the gathering of monitor statistics and the analysis of all of the data. once again, I think this assumption may not be true, not only of Loadrunner but of all commercial tools.
Purchasing Software, and this applies to testing tools as well, is very much like purchasing an air conditioning system. What you do not want to wind up with is the functional equivalent of the six ton chiller outside of a log cabin (Hyperefficient, overpaying for utility) or window based air conditioning units in every window of the Sears tower (easy to use, low initial cost but the ongoing operational cost and maintenance is a huge financial burden). There are going to be cases where JMETER is a better match for a particular problem than is any commercial tool, likewise there are goin to be cases where the commercial tools are going to be more efficient that any open source tool.
Working for an organization which delivers services for both commercial and open source tools for various customers, where we have found the open source tools to be most inefficient as compared to their commercial counterparts is in the monitoring of particular infrastructures, the analysis of the results and also in the development of particular protocols. Keep in mond, LoadRunner in particular supports 30++ different interfaces plus four different template scripts (VBSCRIPT, Javascript, Java, C) to rool your own for what is not supported. It's very difficult for the opensource community to keep up with that level of interface breadth. This efficiency, not to mention the backing of support, the VAR community, a robust user community and a track record of having been there and done that already for the commercial tools has a value and in turn leads to lower risk in the use of the toolset versus an opensource tool in many cases. This shift in efficiency combined with the lowered risk in use (as compared with risk of your application failing in production) determines the value of the tool.
If a client dictates that an open source tool has to be used, but there can be no difference in output an analysis, this then has the effect of taking a balloon and squeezing the one end which represents tool budget and just passing the cost into the services side to make up for any inefficiencies the tool may have compared to their commercial counterparts. I had a potential client recently who had spent millions of dollars on an infrastructure involving fantastic hardware and software for thousands of concurrent users. All commercial products, no Open source. The client readily admitted that downtime was not allowable and that even an hour of downtime would approach 100,000 USD. Yet, their approach to performance testing was not aligned with risk. They sought out vendors only using free tools or tools that they could "borrow." Yep, there testing budget was way out of line with accepted industry practices of between 25%-33%. There performance testing budget? "How low can you go?"
Quote:
Why would anybody buy LoadRunner when there is JMeter? Say the users are a little technical (actually they are developers), so they can use the JMeter to test their web application - right?
If JMeter meets your needs, go for it. Many Open Source tools are very useful and will be sufficient in certain situations.
Also.. we are seeing open source tools continue to grow and mature in the test tools space. In the perf world, we have JMeter, OpenSTA, WebLOAD OS, Grinder, etc, etc. This is already starting to shake up the proprietary vendor's plans. This will be very interesting to watch over the next few years.
You must log in to post.
Video embedded using Easy Video Embed plugin