[Asterisk-Users] Server Criteria

Scott Laird scott at sigkill.org
Fri Feb 4 14:15:57 MST 2005


On Feb 4, 2005, at 11:21 AM, Spencer Nassar wrote:
> I've been doing a lot of background reading/searching of this list, 
> voip-info.org, and Google, looking to define a good candidate for a 
> server platform.  I'm very interested in thoughts from others!  So 
> here goes...
>
> Axiom 1:  if you are not doing doing much transcoding (converting 
> between codecs), the bottleneck for supporting high volumes of 
> simultaneous calls is system bus speed, not CPU power
> ---> points to a 64 bit AMD Opteron system, and maybe just one of the 
> two processor slots populated.  Bus is twice as wide as a 32 bit 
> system, and operates at 1.8GHz (a lot faster than a 64 bit Zeon 
> system).  Then add the second processor to the board if you see you 
> need it.

Er, not really.  First, the closest thing to a "system bus" in the 
Opteron system is really only 16 bits wide at its widest, and it 
probably only 8 bits wide by the point that it talks to your PCI 
bridge.  But, measuring the system's throughput on the basis of how 
many wires its internal plumbing uses is kind of nutty--the Opteron can 
move more data over an 8-bit Hypertransport link then a Pentium III 
could over its 64-bit bus.  And, in *either* case, the system bus isn't 
going to be a problem for any sort of telephony application.  I mean, 
even an 8-bit 800 MHz HT link is good for 12.8 Gbps each direction.  
That's a lot of T1s.

The real problem seems to be either CPU performance for transcoding or 
(more frequently) PCI bus bandwidth and latency.  Low-end systems with 
a single 32-bit PCI bus are going to have problems with more then one 
Digium card.  Systems with multiple busses *should* be able to scale 
further, but I haven't seen any sort of testing to back this up.

> Axiom 2:  Get lots of memory
> ---> I haven't seen this quantified, and plan to do some testing.  
> I'll post results here, but can anyone share any insights?  I'm 
> planning to start at 2GB, and go up from there if I see swap getting 
> used.
>    - what would an alaw to alaw connection consume (if it didn't hand 
> off)?
>    - what about a 5 call alaw meetme bridge (and how much memory per 
> incremental caller)

Again, not really.  Asterisk doesn't use a whole lot of RAM.  Make sure 
that it's not swapping, but even 256 MB is probably enough most of the 
time.

> Axiom 3: Don't allow any disk IO
> ---> I'm assuming this is related to #2 - get lots of memory to avoid 
> swap to disk.  Other issues or thoughts?

Well, if you have a really nasty IDE bus with DMA and interrupts 
disabled, then disk I/O could probably be a problem.  Other then that, 
it shouldn't matter.  As others have said, don't run a big DB server on 
the same box, but a bit of disk I/O isn't a problem.

> Axoim 4: Come codecs will take advantage of the faster floating point 
> of a 64 bit system
> ---> unknown... has anyone seen this?  Will Asterisk, compiled in a 64 
> bit Linux environment, reap these or other benefits from being on a 64 
> bit system (other than the system bus speed)?

Dunno.  I suspect that some codecs *could* benefit from 64-bit math, 
but I doubt that any of the current codecs are tuned for 64-bit CPUs.


Scott




More information about the asterisk-users mailing list