[asterisk-users] Cisco vs Asterisk

[Alex Balashov]

Al Baker wrote:

> "Yet amazingly (if this is, indeed, a source of amazement for you), CCM 
> and other Cisco software can be just as buggy as anything OSS, if not 
> worse. "
> 
> This is simply NOT TRUE and shows a complete lack of understanding of modern software development.
> CISCO software is developed in a CMM environment.
> It has a formal test methodology and uses Automated Testing on EACH new release to ensure that 100% of the software that functioned in the Last Release, actually works in this release.
> Further, there is mandatory "soak-testing"  for all new software.
> Sorry, anyone who wants to compare Professional TELCO GRADE software development with Open Source is just Completely and Totally freakin clueless.

Uh, no.  That juxtaposition is simply not accurate and shows a complete 
lack of experience with actually-existing reality.

Yes, formal development methodologies, QA, regression testing, etc. are 
certainly very helpful and probably help to eliminate certain categories 
of bugs, but to hear it from you, it is as if serious open-source 
projects (let alone ones with official corporate maintainers and 
sponsors) don't do any of these things, or that these methodologies are 
a panacea that produces bug-free products.

The bugs are still there, in innumerable quantities.  Automated testing 
is relatively ineffective at finding most of the things that beleaguer 
Cisco gear in production environments.  I don't think we need to get 
into a lengthy discussion of the kinds of bugs that we routinely chance 
to encounter, but it suffices to say that many of them do boil down to 
the difference between Marketing's claims and actual backplane/DSP/bus 
capacity and/or throughput.  In the case of "Professional TELCO GRADE" 
software and hardware, that problem is much greater and accounts for a 
much larger share of problems.

Also, an inherent limitation upon the QA and feedback process of 
commercial vendors is the small number of installations.  Sure, Cisco 
may be one of the most universal varieties of anything in the computing 
world, but the number of deployments - let alone ones with fully-fledged 
support agreements - is small.  And how many of those adopters are going 
to push Cisco gear to the limits where it starts to fail so spectacularly?

It's a very, very small number of installations compared to the terrific 
number of open-source deployments, not to mention the pairs of eyes that 
lay on the code in an open-source situation.  If you were to consider 
that with any kind of intimate detail, you would see that the average 
quality and range of user feedback that someone like Digium or MySQL AB 
gets is much, much higher and more discerning than the things that come 
into the Cisco TAC.

In short, there is absolutely not a damn thing that makes commercial 
software superior here ipso facto from an engineering perspective;  not 
a single smattering of an iota of a thing.  As with all projects, the 
intellectual coherence, sophistication, and skill of the implementors 
and other elements of its polity vary immensely.  Some commercial 
products - including very expensive, well-supported ones - are an 
absolutely abysmal, apocalyptic pile of dross.  Some open-source 
projects are extremely well-managed and mature, on the whole (MySQL). 
Most lie somewhere in between.  But there is nothing about the corporate 
method of software development that produces higher quality work;  if 
anything, it is bound to be somewhat lower.

-- Alex


-- 
Alex Balashov
Evariste Systems
Web    : http://www.evaristesys.com/
Tel    : (+1) (678) 954-0670
Direct : (+1) (678) 954-0671
Mobile : (+1) (706) 338-8599