The MPLS-OPS Archive

Cell Relay Retreat>MPLS-OPS Archive>month:2003-Mar> msg00046



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

Re: Re: Urgent Help Needed...

  • From: Rik Wade <rik@tcpip.fsnet.co.uk>
  • Date: Tue, 18 Mar 2003 15:41:59 +0000
  • Cc: Muzammil Ahmad Khan <mukhan@ssuet.edu.pk>, mpls-ops@mplsrc.com
  • Resent-Date: Tue, 18 Mar 2003 11:35:08 -0500
  • To: john smith <johnsmith0302@hotmail.com>
  • User-Agent: Mutt/1.4i

On Tue, 18 Mar 2003, john smith wrote:

> Throughput = average number of bits/time (if u dont mind using L2 overheads)

I'd be wary of getting in to using an "average number of bits" in this 
calculation. By definition, an "average" is already calculated over a 
period of time. We are then dividing it by a further component of time.

I'd keep it simple by just saying:

Throughput (bps) = total number of bits transferred / time in seconds

As John says though, including large quiet periods in the selected
time period may adversely affect the result you're after.

> what do most people/customers define as througput?
> i think its *WIRE* capacity.....the concept of a SDH circuits/Lease lines

Customers generally accept the definition as being the raw bandwidth
capacity of a link, unless otherwise specified. If a customer has specific
application-layer throughput requirements then this is normally specified.
--
rik wade

-------
The MPLS-OPS Mailing List
Subscribe/Unsubscribe:  http://www.mplsrc.com/mplsops.shtml
Archive: http://www.mplsrc.com/mpls-ops_archive.shtml