open source performance test tool for Oracle apps
Sasank
James Pulley
Wouldn’t you have better luck in getting an answer in a tool agnostic forum such as SQAForums.com performance testing forum?
And yes, there are other tools and they are pretty much all commercial in nature. I suggest you reorder your priorities and appropriately measure the risk of downtime in your environment. Once you know the cost of an hour of downtime for your use population and how many hours you spent in downtime related to performance issues, then you will have a better picture of risk and cost in the universe of open source and commercial.
James Pulley, http://www.loadrunnerbythehour.com/PricingMatrix
--
You received this message because you are subscribed to the Google
"LoadRunner" group.
To post to this group, send email to LR-Loa...@googlegroups.com
To unsubscribe from this group, send email to
LR-LoadRunne...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/LR-LoadRunner?hl=en
Floris Kraak
<loadrunn...@jamespulley.com> wrote:
>
> And yes, there are other tools and they are pretty much all commercial in
> nature. I suggest you reorder your priorities and appropriately measure
> the risk of downtime in your environment. Once you know the cost of an
> hour of downtime for your use population and how many hours you spent in
> downtime related to performance issues, then you will have a better picture
> of risk and cost in the universe of open source and commercial.
>
Choosing the right tool for the job is very important, but you cannot
blame someone for looking for tools that do not require him to be
beholden to the whims of a hugely expensive proprietary license
agreement first.
The argument "but paying us lots of money will save you a lot more
money" is very nice and all, but a good open source performance
testing tool would do the same thing while probably costing a lot less
AND giving you the power to fix the tool if there are bugs. (The
latter of which I consider more important, imho. I really don't want
to know how much time I've lost on some of the nigh unfixable problems
I've run into in the course of the past few years)
The fact that there are no FOSS tools that measure up to the LR
standard right now is no reason to tell people to stop looking for
them.
Regards,
Floris
---
'Apple says "you can't distribute this iPhone app in any form or
through any means unless we explicitly allow you to do so, and you
can't allow others to disobey this rule".
GPL says: "you can distribute this iPhone app in any forms and through
any means you want, and you can't prevent others to do so".
Do I sense a subtle opposition?'
--- Gabriel Morin
James Pulley
source CMS, DB, email, F&P, web, etc.... When I was in the process of
deciding LINUX (and which distribution) versus other for the back office I
did not go to a commercial UNIX forum or Windows forum and ask others to do
my own research. Plus it would have been flat insulting to the users in
that forum, "Hey I would like a Boeing aircraft for all of its features, but
could one of you recommend a far less expensive aircraft with all of the
same features?" An intense curiosity and an ability to research well are
prized skills for performance testers, of which the poster seemed to prize
or demonstrate neither. Regardless of tool selected this lack of curiosity
and demonstrated research skills does not bode well for the success of the
endeavor.
To the second issue, the cost of the tool is only one cost component for any
deployment decision. Witnessed by my own infrastructure, I can assure you
that the care and feeding of the environment in terms of labor and
infrastructure dwarfs what would have been spent on even the most expensive
infrastructure offering I could have chosen for my own enterprise.
Performance testing tools are no different. If you cannot justify your tool
choice to the business based upon its projected or actual return on
investment, then you will forever be looked at as a cost and not as a
benefit in your performance engineering practice. To not understand and be
able to speak to the business side of the house will be a career limiting
decision.
James Pulley, http://www.loadrunnerbythehour.com/PricingMatrix
-----Original Message-----
From: lr-loa...@googlegroups.com [mailto:lr-loa...@googlegroups.com]
--
Sasank
Floris Kraak
<loadrunn...@jamespulley.com> wrote:
> Plus it would have been flat insulting to the users in
> that forum, "Hey I would like a Boeing aircraft for all of its features, but
> could one of you recommend a far less expensive aircraft with all of the
> same features?"
If there was such a thing known to the people posting there I am sure
that those who have no financial ties to Boeing would quickly start to
drift away to a different forum ;-)
> An intense curiosity and an ability to research well are
> prized skills for performance testers, of which the poster seemed to prize
> or demonstrate neither. Regardless of tool selected this lack of curiosity
> and demonstrated research skills does not bode well for the success of the
> endeavor.
Unfortunately the vast majority of the population does not have this
inclination to do basic research on their own for some reason or
other.
I feel lucky because I do know a number of people who possess this
trait, but it is far from a given.
>
> To the second issue, the cost of the tool is only one cost component for any
> deployment decision. Witnessed by my own infrastructure, I can assure you
> that the care and feeding of the environment in terms of labor and
> infrastructure dwarfs what would have been spent on even the most expensive
> infrastructure offering I could have chosen for my own enterprise.
> Performance testing tools are no different. If you cannot justify your tool
> choice to the business based upon its projected or actual return on
> investment, then you will forever be looked at as a cost and not as a
> benefit in your performance engineering practice. To not understand and be
> able to speak to the business side of the house will be a career limiting
> decision.
>
In general terms this is true. In the specific case of performance
test tools, I fully agree as well.
However, in the long run I keep evaluating my options because I am
convinced that the balance of this decision is bound to change - the
pendulum will swing towards FOSS at some point in the future, simply
because the inherent cost of proprietary licenses will never go away.
And with inherent costs I mean things like "Not being allowed or given
the ability to fix defects yourself", "supplier lock-in", "not having
the benefit of a diverse ecosystem of suppliers who can add features
or fix defects for you".
Things like that are hard to account for but they are very real.
James Pulley
The world of performance testing tools is a pretty finite one. I would encourage you to use your basic research skills to examine all of the options in the market on the open source front and then develop a working hypothesis as to whether a given tool will support your version of ORACLE for your testing needs. Once you have a hypothesis as to the nature of the support then a test would be in order to examine whether the hypothesis is true or false as to the nature of the support in the tool in question and whether that support matches both your environment needs and team skills, or the dreaded Proof of concept. Be sure and cover all of the aspects you need to examine apart from simple interface support, such as tool capabilities for remote generation (if needed by your group), results collection, reporting and analysis capabilities.
You will likely find that two candidates in the open source world bubble to the top above the others. This is primarily expressed because of their use of Java for interface connectivity, but not for any specific native integration with ORACLE – These tools simply inherit JMS and JDBC interface support from their underlying language support for Java which then allows for a connection to ORACLE. You will have a far greater dependency upon refined development skills in the open source world, a characteristic which is not always present in performance test teams. The same holds true on building reporting and analysis which is a weak link of open source tools.
It is possible to take the raw data out of open source tools and build a “mock” LoadRunner results set that can be loaded and used with the LoadRunner Analysis tool, which can address many of the back end reporting and analysis limitations of open source products, for which you could make use of your existing LR 8.0 Analysis toolkit and methods. Such an integration will wind up being customized with the tool you ultimately select and may involve other tools to post process data from your test tool to format it for use in a LoadRunner Results Set (.lrr file and associated event data). And yes, this custom integration with analysis is something our organization has done a number of times to integrate data from both commercial and open source tools for customers who use both and want common reporting and analysis capabilities.
Btw, I have been looking for a new BBQ smoker. I really like the versions that come on a trailer made out of two 50 gallon drums with a firebox on the side. I particularly like the double grill feature that allows one to flip the meat being smoked while it is contained between two grills. Can anyone recommend anything smaller that I can still fit two pigs in for family barbeque events. Oh, and #1 says it has to be pleasing to the eye and cannot dominate “her backyard.” That last feature is key! Cheaper is better: Any recommendations? ;)
James Pulley, http://www.loadrunnerbythehour.com/PricingMatrix
From: lr-loa...@googlegroups.com [mailto:lr-loa...@googlegroups.com] On Behalf Of Sasank
Sent: Wednesday, November 10, 2010 7:33 AM
To: lr-loa...@googlegroups.com