[Asterisk-Dev] Re: Asterisk-Dev Digest, Vol 4, Issue 64

Nick Bachmann asterisk at not-real.org
Thu Nov 25 17:05:24 MST 2004


Darren Sessions wrote:

>>May I point you to http://www.catb.org/~esr/faqs/smart-questions.html
>>
>>"Much of what looks like rudeness in hacker circles is not intended to
>>give offence. Rather, it's the product of the direct,
>>cut-through-the-bullshit communications style that is natural to people
>>who are more concerned about solving problems than making others feel
>>warm and fuzzy."
>>    
>>
>While I see your point - it didn't take much for you to have normal
>conversation. 
>

While I might be more diplomatic than Steve, he knows more about 
Asterisk... take your pick.

>I had a legitimate question - may not have stated it as
>intelligently as I could - but it was legitimate. I wasn't looking for warm
>fuzzies or to get flamed - just a simple yes or no if it could be done
>  
>

But you didn't ask a "yes or no" question... it was like asking "Can I 
shock myself with a defibrillator to remove my chest hair?" Yes you can, 
but a defibrillator is the wrong tool for the job.

>(possibly the simplest form of cut-through-the-bullshit there is).
>  
>
>>http://www.networkcomputing.com/netdesign/nfs4.html
>>http://www.coda.cs.cmu.edu/
>>
>>I have to agree with Steve here; NFS (or AFS, or CODA) is a lot better
>>option than BLOBs.  If you do a little research, you can find that it's
>>very easy to set up replicated NFS/CODA servers.
>>    
>>
>Why exactly is it a better option? Throughput?
>  
>

Think of it this way: relational databases are meant to store many 
fields of relatively short data; that's what they're optimized for, and 
that's what the database protocols (SQL, etc.) are designed to work 
with. They aren't meant to work with binary data and they aren't all 
that great at dealing with several megabyte fields. If you force them to 
do that kind of work, they're not going to do it super fast.

Network filesystems, on the other hand, were meant for doing those 
things, which is why Asterisk gets its sounds out of filesystems by 
default, and not databases. If you put your wavs on a NFS filesystem, 
you don't have to do any custom coding to get them to work, you can just 
use a background() or playback() to play them. You'll usually make more 
problems than solutions by writing custom modules or EAGIs to do audio 
processing work that Asterisk can do internally.

So, not only do network filesystems present better throughput for files, 
but they also make your job easier and reduce the chance of bugs being 
introduced into your Asterisk codebase that weren't there before.

Nick



More information about the asterisk-dev mailing list