Scalability of DOORS vs. DOORS Next Gen
Given all that, I read something on the online documentation that concerned me somewhat, and I'd like some clarification on what it means. At http://pic.dhe.ibm.com/infocenter/doorshlp/v9r5/index.jsp, it states: "the current version of Rational DOORS Next Generation does not support the extreme scale of Rational DOORS, where millions of objects can exist in a project's modules."
Does this imply that there are certain limitation to DOORS Next Gen in terms of the number of items in the project? In theory, if we had a project that wanted to use DOORS Next Gen but expected to have many, many requirements, should we tell them no? Please clarify what the limitation of DOORS Next Gen are in the context of the quoted statement above.
Thanks,
Joe Gariano
Accepted answer
>>>IBM is continuing to invest in making Jazz based applications more robust and scalable, and therefore future versions of DOORS Next Gen should only get better in this area.
Yes exactly. In fact the work we are doing for the 5.0 release shows really good improvements, and recent tests using solid state disk and more RAM showed even more scalability. We intend to publish these results in an updated performance report.
>>>IBM is not discouraging the use of DOORS Next Gen on large projects, but given the newness of the tool, they made the statement I quoted above as disclaimer that DOORS Next Gen has not yet been proven at the levels of scale that original DOORS has.
We know of DOORS databases with more than a terabyte of data. These DOORS databases started small and grew over more than a decade of enterprise-wide use. I don't think DOORS NG is ready for that size today. But we do intend to keep increasing the scalability so that as the number teams and number of artifacts grow, DOORS NG will be ready for that.
>>>there is no well defined threshold at which the original DOORS shows its superiority in terms of scalability over DOORS Next Gen.
DOORS NG does some things faster / more scalable than DOORS 9 today, for example, working across a wide area network, faster module load time, and scripting extensions.
Best would be to engage IBM or a knowledgeable business partner to discuss your needs and suggest some performance benchmarks that would help you assess DOORS NG in your particular environment and projected scale.
One other answer
you may want to have a look at the RRC/DNG perf. reports:
https://jazz.net/wiki/bin/view/Deployment/CollaborativeLifecycleManagementPerformanceReportRRC406Release
and compare the order of magnitude (& response times) between the # of assets manipulated there and your plans for your particular project. That's one approach... not answering your exact question though.
Regards,
Stéphane
Comments
Found this related previous question & answer (encompassing a large scope though: RRC/DNG plus other CLM apps): https://jazz.net/forum/questions/104489/clm-project-setup-recommendations
Thanks for the reply, Stephane. There certainly is a lot of information in those links, but I was looking for a simpler answer regarding the scalability difference between original DOORS and DOORS Next Gen -- something which wasn't apparent to me in any of those links. Specifically, I'm wondering if there is some inherent/architectural limitation in DOORS Next Gen that will always make it worse suited for projects with a large number of artifacts. If such a limitation exists, I'd like to understand it better. On the other hand, if its more that DOORS Next Gen is new and hasn't yet been proven with very large projects like original DOORS has, but its scalability is getting better and better with each version, then that's something I can accept. We are not at the point right now of using DOORS Next Gen on an enormous project, but I'd like to know if I should tell a large project not to use it should one ever express interest in it.
Joe, I'm receptive to the kind of questions you're asking. In this blog post, I'm trying to collect what are the known limits or educated guesses on where you should start monitoring your Jazz server. It's covering all CLM products (RTC/RQM/RRC-DNG) and there is a section dedicated to RRC-DNG here. I believe it's still not the level of details you're expecting but... you may want to have a look at this piece of information (for your awareness). It's "light" on RRC-DNG for now but hopefully you'll find it useful. Hope you can get add'al feedback in that area, something matching your expectations in a more direct manner. Please let me know in such case as that could be of a help for enhancing the resource I just mentioned... Cheers.
Thanks Stephane. The firewall here at my work is blocking access to those links. But from what I'm gathering, there is no well defined threshold at which the original DOORS shows its superiority in terms of scalability over DOORS Next Gen. Because DOORS Next Gen is new, and because it is based on the Jazz platform, the scalability of DOORS Next Gen is tied the scalability of the Jazz server, whereas original DOORS is a more mature, standalone application that has been proved over the years on very large projects. IBM is continuing to invest in making Jazz based applications more robust and scalable, and therefore future versions of DOORS Next Gen should only get better in this area. IBM is not discouraging the use of DOORS Next Gen on large projects, but given the newness of the tool, they made the statement I quoted above as disclaimer that DOORS Next Gen has not yet been proven at the levels of scale that original DOORS has.
Is that an accurate summary of the issue?