The Elsmar Cove Wiki More Free Files The Elsmar Cove Forums Discussion Thread Index Post Attachments Listing Failure Modes Services and Solutions to Problems Elsmar cove Forums Main Page Elsmar Cove Home Page
Google
  Web Elsmar.com
*Please be aware that SOME RECENT forum threads may not yet be indexed by Google.

View Full Version : Software Reliability Measurement - Telecordia standard GR-282-CORE


Govind
30th April 2004, 01:58 AM
We are planning to set up a formal Software Reliability function in our organization.

We develop subsytem level telecom software and module level firmware.
SW Reliability seems to be a fairly recent development in the Software Engineering field.Hence references and experienced professionals also seem to be very limited.

Iam trying to find out if any one in this forum can share some ideas, experience for setting up of Software reliability function, information on useful models for growth, etc.

We are currently reviewing a Telecordia standard called GR-282-CORE.

Regards,
Govind.

Marc
3rd May 2004, 03:38 AM
Well, I hate to see NO RESPONSE to a post, but as discussed in the other thread, not many software folks visit here.

I've played with software in different application environments. I was involved in a project on the old Mitubishi twin turbo with the auto-magic hardtop store project some years back. That was essentially automotive. Timing was everything and there was virtually continuous firmware 'tuning' as the Master Buck progressed. Little things, like tolerances and such, often changed once or twice a week.

I was involved in military firmware back some years including development of frequency hopping RF devices and IR imaging (focus then was reconnaissance) devices. I was always on the device (LRU) QA side so while I wasn't directly involved in the software (firmware) development, I was responsible for the operation of the devices. Obviously, stable firmware was required, but I never did see a firmware failure in military applications.

That said, this stuff was not 10 million lines of code. That is NOT to say the software wasn't extremely complex.

I was thinking of your other thread and spotted something on SlashDot http://slashdot.org/ you might want to read about debugging. The direct link is: http://slashdot.org/articles/04/05/02/1346209.shtml?tid=126&tid=156 (hope this works). It talks about people still using print statements to debug programs. Ummmm, punchcard era or what?

I've read a bit about CMM (Capability Maturity Model for Software - aka the CMM and SW-CMM) http://www.sei.cmu.edu/ (impressive) and Tick-It http://www.tickit.org/ as well as the old ISO 9000-3, but have no idea what's being used by what sector.

Some of this has been discussed here - See: http://elsmar.com/Forums/showthread.php?t=4511 for example.

An interesting topic might be what quality assurance is 'applied' to debugging software...

Just a few thoughts...

Govind
3rd May 2004, 02:31 PM
Marc,
Thanks for those pointers and your reply. I have visited those links earlier. There is actually a more appropriate link on this subject which I review periodically.

http://members.aol.com/JohnDMusa/ARTweb.htm

Author Dr.Musa is an authority in the subject.

Also, there is another person in Ottawa Canada, Dr.Lionel Briand who has also published some papers and have some implementations.
http://www.sce.carleton.ca/faculty/briand/index.html

What I was expecting from our forum members through my posting was their experience in implementing Software Reliability. I will continue use this thread in the future to post my experience of implementation. We will see if any more members respond to this thread.

Govind.

SteveR
19th May 2004, 02:17 PM
Govind,

I saw your post and thought I may be able to shed a little light on this subject. I have used software reliability growth modeling techniques for many years in order to improve reliability of embedded real-time software - primarily in medical devices. The techniques for modeling and measuring software reliability growth are not that complicated - having someone with a background in statistics helps. More importantly, having people who understand how your organization develops software and who can chose a software reliability model that most closely matches what you do is critical.

I would recommend the following resources to help you learn more...

Software Reliability Engineering, by John Musa, McGraw-Hill, 1999

and my book

Software Verification and Validation for Practitioners and Managers, Artech House, 2001. My book has a chapter that explains how to select an appropriate model, collect and analyze data, and use the data to make business decisions.

I would also suggest that you visit my website. I offer workshops on risk management and reliability growth that may be helpful

http://www.swqual.com/training/risk2.html

Regards,

-Steve Rakitin-

Govind
19th May 2004, 02:45 PM
Steve,
Welcome to the forum. Thanks for your suggestions. I will review your webpage and "private message" any specific questions. Elsmar forum require Software professionals like you to support and answer software Quality postings. Iam glad that you registered.
Thanks,
Govind.