After reading Dave's answer, I realized that I neglected to send this response to the list. I'm doing it now in case someone else might benefit.
Dave ---------- Forwarded message ---------- From: Dave Donovan <[email protected]> Date: Fri, Feb 27, 2009 at 6:50 PM Subject: Re: [on-asterisk] Fwd: trunk utilization To: [email protected] Hi Bob, I think the model you're looking for is called Erlang B. I know that it's usually discussed in terms of call centres but I think it also applies to trunking and blocking situations (as opposed to queuing situations which is where the call centre is particularly interested). Here's an article that gives some background. It's in the context of call centres but the language is pretty accessible. I'm didn't get a math degree so even the wikipedia entry was pretty useless to me. :-) If you flip ahead to the section on Erlang B, you'll get some background on what the inputs are to the model. http://dmi.uib.es/~loren/docencia/webxtel/bibliografia/Erlang%20B%20&%20C.pdf >From the article: _______________________________________ Erlang B: The Easy One The most common traffic engineering problem involves sizing a trunk group — how many trunks are needed to carry your toll-free calls, how many tie trunks between two offi ces, how many ports into your voice mail system, or some similar question. Erlang B handles that relatively easily, in four steps: _______________________________________ Once you've collected some data and you have your BHT (Busy Hour Traffic) then you can just go to an Erlang B calculator like this one: http://www.erlang.com/calculator/erlb/ There are three inputs: BHT, probability of blocking and trunks. Typically you provide your BHT and the acceptable probability of blocking (busy signal) and it will tell you how many trunks are recommended by the model. The question you posed is a bit different. You've given the number of trunks. I'll assume that you don't want a lot of busy signals. And, you're looking for talk time. The model doesn't say how much your people actually talk, just how much they could talk with a given blocking percentage. If we assume that you almost never get busy signals (p =0.001) we can calculate that you can support a BHT (in erlangs or channel hours) of 2. That means your people can chat for a cumulative 120 minutes between them and still expect to get almost no busy signals. Good luck. I hope this helps. Dave On Fri, Feb 27, 2009 at 5:52 PM, <[email protected]> wrote: > Hi Alex, > > 8 trunks, 24 users, legal firm. > > I am looking at models but none seem to exist. Any opinion? > > Bob > > > > On Fri, Feb 27, 2009 at 12:58 PM, Alex Kink <[email protected]> wrote: > >> Bit more info on the company would help. >> >> How many people total, how many sales people, customer support? >> >> >> On Fri, Feb 27, 2009 at 12:35 PM, <[email protected]> wrote: >> >> > Hi there, would anyone care to offer up an opinion on what you would deem >> > to >> > be typical call usage minutes on a trunk? It is for a termination >> > utilization calculation (not a call centre, just a typical enterprise >> > customer) with 8 or 9 x 5 business hours operating on a 5 day week. >> > >> > Curious if anyone could provide an opinion. >> > >> > Origination (In bound): >> > Termination (Out bound): >> > >> > Could be either on a per trunk or per user basis. I am about to make a >> > guess, and I would appreciate a second or third viewpoint. >> > >> > Any input would be appreciated ! >> > >> > Bob Smith >> > [email protected] >> > >> > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
