- This topic has 9 replies, 1 voice, and was last updated 19 years, 11 months ago by Guilherme.
23rd January 2001 at 03:20 #33187Johnny HoGuest
Is there any practice for dimensioning trunk size of PABX system that is particularly for hotel application based on number of guest rooms and/or extensions.
More, any particular parameters, like BHT, Blockings, Grades of Services that should be applied when dimensioning trunk size for hotel application since we normally expected that the system should be 99.99% available, and assume this should be critical application.
While, any formula for the dimensioning?
In normal, I will dimension the trunk size base on 10% of the total extension of the PABX required, would this be enough?
Please give some ideas.29th January 2001 at 21:32 #33188Alex YuenGuest
BHT in average = 60 secs in short
BHT in average = 180 secs in long
Although is 100% available but normally the traffic is very low which compared with local telephone lines.
Therefore we need to dimension in different type of trunks:
1. DID line for hotel,
2. DOD line for extn., direct line.
Different GOS will be used to dimension these two trunk type.
1. for DID – 0.5% GOS
2. for DOD – can be divided into seperate trunk group with overflow facility to ensure no blocking. For initial trunk group – 30% overflow and 0.5% GOS for final route.
Nevertheless, it will depend on the equipment investment from the company.1st February 2001 at 09:51 #33189Johnny HoGuest
Can you give some example to size the trunk group?1st February 2001 at 22:57 #33190RommelGuest
The suggestions Alex provided you are sound much more like suggestions for sizing an enterprise environment rather than a hotel environment.
Although I dont claim to work in that industry, I believe the requirements are different.i.e.
Is the hotel targeted towards business customers or vacationers?
I guarantee that vacationers will use the phone a lot less than business folks so if you are working on a phone system for the “fun and sun” in Florida the type of traffic they will see will be different from the Hyatt downtown Chicago.
Also 99.99% availability has more to do with the system and not the trunks. That type of availability means the system should be down no longer than 53 minutes per year.
IME rules of thumb are quick and easy but its better to get the right information first and the customer will be happier in the long run. You will be able to back all your decisions with data.2nd February 2001 at 21:43 #33191KarenGuest
I have a similar question, but this is more fundamental. Is there a rule of thumb for RESIDENTAL users how many line sides connection to a trunk side connection? For example, in the business world we’ve been successful using a 4:1 ratio when we have no historical BH data or information on call patterns. I’m looking for another quick and dirty rule of thumb, but this time for residential users. If anyone has ideas, please post.
Thanks!6th February 2001 at 06:26 #33192Alex YuenGuest
Take an example for a hotel has 200 rooms, with 10 DID line and 30 DOD lines.
The originating traffic rate & the terminating traffic rate for a hotel is much difference from residential subscriber line.
For normal residential subscriber line, the originating traffic will be only 0.04 to 0.05 erlang. But for business line is quite different, it will be 0.06 to 0.16 and depend the type of business line you have.
You can divide the rooms into say 5 trunk groups, with room no. 1,6,11,16… to be TG 1; room no. 2,7,12,17…. to be TG 2, room no. 3, 8, 13, 18….to be TG 3 and so on….. Each TG will have 5 DODs, the other DOD lines will be used by office.
You also can plan the overflow routes between the TGs. Suppose TG1 overflow to TG3 and then overflow to TG5.
It is only a simple configuration for such small size phone lines.
(The traffic rates of residential lines had been already mentioned above. You also need to know the day to busy hour ratio, there should be have two different busy hour for residential lines, and the day to busy hour ratio will be 10 to 12)
Alex6th February 2001 at 09:04 #33193Chris TaylorGuest
Alex: this seems a strange design tactic. Surely, it’s better to have one fat pipe instead of several thinner pipes?7th February 2001 at 03:10 #33194Alex YuenGuest
If you have one fat pipe, how many trunks for this fat pipe?
But if use several thin pipes, it can save a lot of trunks.
On the other hand, you can use the overflow criteria to ensure the blocking to near zero.
For dimensioning purpose, if you use overflow condition, it is on the save side for GOS and the needs of trunks.
Alex17th February 2001 at 13:28 #33195Chris TaylorGuest
Sorry, but I don’t agree. If you overflow between small groups, then effectively, you’ve ended up with one fat trunk – so what’s the point? You certainly don’t reduce the overall number of trunks you need. You just make the situation more complicated.
And how many trunks for the fat pipe? Well, you have to ask yourself that question whatever way you choose to engineer your trunks.11th May 2001 at 19:31 #33196GuilhermeGuest
Supose you have 200 subs generating 200 call/hour and 10 min of holding time:
L=200 call_h/60 min = 3.33call/min
t= 10 min
A= 3.33*10 = 33E
1% blocking means 40 trunks is sufficiently to this exemple.