[asterisk-users] measuring network quality in the field

[Tim Panton]

On 28 Jun 2008, at 18:36, Tzafrir Cohen wrote:

> On Sat, Jun 28, 2008 at 12:16:53PM -0400, Steve Totaro wrote:
>> On Sat, Jun 28, 2008 at 12:06 PM, Tzafrir Cohen
>> <tzafrir.cohen at xorcom.com> wrote:
>>> On Sat, Jun 28, 2008 at 11:25:44AM -0400, Steve Totaro wrote:
>>>> On Sat, Jun 28, 2008 at 10:11 AM, Tzafrir Cohen
>>>> <tzafrir.cohen at xorcom.com> wrote:
>>>>> On Fri, Jun 27, 2008 at 05:03:01PM -0400, Simon P. Ditner wrote:
>>>>>> What open source tools are people using to quantitatively  
>>>>>> measure how
>>>>>> well QoS/traffic shaping is performing out in the field, and  
>>>>>> what call
>>>>>> quality people are experiencing in terms of jitter and packet  
>>>>>> loss?
>>>>>
>>>>> Maybe ntop?
>>>>
>>>> Maybe SNMP?
>>>
>>> To monitor what data, exactly? SNMP is a potential way to get data  
>>> from
>>> various hosts. But how do the hosts actually get the data?
>>>
>>
>> I always enjoy how you snip my posts.
>
> Yes, specifically when you quote my signature.
>
> I snipped methods for gathring information that is available through
> SNMP already. But is this information useful?
>
> Specifically, I don't care about fragmantion and packet loss of all
> packets. Just of those in the relevant streams.

RMON should be able deliver this sort of filtering for SNMP, many  
routers
support RMON, but it takes some setting up to get useful results.

Likewise there are some vendor specific MIBs that may be useful:
"Cisco IOS provides QoS statistics in its CISCO-CLASS-BASED-QOS-MIB"

There is something to be said for monitoring QoS on the device that is  
enforcing  it.
In theory at least you should get a more accurate picture of the  
situation and the
reasons behind it.

Conversely ntop running  on a separate PC may give a more  
'independent' version of
events.

Tim.