Uh, Ron, the Erlang formulas have nothing to do with uptime; they have to do with queuing systems, and are usually applied to calculate call blocking (erlang B) or call delay (erlang C) for things like trunk groups and ACD agents. To convert uptime into minutes of outage per year, you simply multiply (1-availability) x 525,600 (the number of minutes in a non-leap year).

I hope you figured that out before you went and implemented Erlang's formula in Excel.

Hello, Anonymous from Siemens -

Hehehe... actually, at least this shows that someone out there is paying attention to what I am writing ;-)

You are absolutely right... judging from what I wrote in this blog entry, one could get the erroneous conclusion that I hinted that you need Erlang's formula to calculate just what was on the table (which you obviously don't). But what I did not include in the table were some combinations of the call success and call abortion probabilities that would lead to the availability rates that I had in the table (please refer to the previous blog entry and the paper).

You are absolutely right - figuring out the associated downtime per year given a certain availability is a trivial exercise (for which Erlang is not needed). The reason why I wrote the Erlang code was to see what would be the lambda and the mu combos that would give the call success probabilities that would lead to the "equivalent" call success probabilities (1- p(blocked calls) - just to get a feel for what call arrival rates and mean duration of calls would trigger these different availabilities. I did not add that to my table... blame it on the late hour of my post, and I formatted the table entering HTML code, so I was lazy.

But thanks for your explanations, which are quite useful to the TF readers - I know both formulae quite well (the Erlang C and B) - in fact, if you do a patent search under my name, you will find that I used the Erlang C quite extensively on mine.

I decided not to get too complicated and post some of this stuff, as I want to keep the blog entries simple, but if you want to discuss this further, please send me an e-mail and we will continue the dialogue.