Sizing - PRI and Dedicated Long Distanance Carrier

[mjpearson]

My company is new to dealing with the telephone world. We initially installed two T-1 circuits. One T-1 was configured for local "switched" service (PRI) for inbound/outbound to allow for toll calls. The second T-1 was configured for outbound traffic for long distance.

It is contract-renewal time. My LEC is suggesting that we're oversized and suggests that we can get by with just one T-1 (the PRI) to do both inbound/outbound. He suggests that we can do away with the dedicated T-1 for long distance.

Are there any good rules-of-thumb to assist me in determining my need?

Here are a few of the things that I do know:

* 100 DID numbers; Roughly 100 of non-DID numbers. Only about 60 of the DID numbers are manned during the work day.

* All 200 numbers are enabled to call within the local area code. About 50 are enabled to make toll and long distance calls.

* I've made an attempt to read my phone bills and I think we're doing about 1400 minutes/month of toll call during the day and about 10,000 minutes/month of long distance calling.

Anyone got a feeling for this stuff? Does it make sense to scale back to a single T-1?


mike

 

[senk1s]

for 60 people system 1 PRI is GENERALLY good enough, unless it a call center or a sales/marketing office.

you can get the traffic reports analyzed in the Meridian itself

option2:if both the circuits are with the same carrier, ...
ask tehm on what bases they are suggesting u are under utilized?

Alltimebusy reports/
max busy reports

ask for peak and historical for say a 3 month or more duration ...

 

[mjpearson]

No. Just a small factory with a slightly higher than normal administrative staff and some machinery operators with very inflated egos that think the phone is for chatting with their wives, friends and neighbors.

Obviously, the LEC has access to my local/toll call stats. I ran through the long distance bills can did the math and came up with the 10,000 minutes/month LD calls. I gave him the LD usage numbers and he did a little rule-of-thumb stuff and concluded that we'd be better served with a single PRI circuit.

I haven't figured out how to use the report function in the Meridian but I did some quick math assuming some worse case assumptions:

Assume:

* LD and toll calls are about the same so, a total of 20,000 min/month.

* All calls are made during the normal typical workday (8 hours/day and 20 days/month).

* 23 circuits in a single T-1.

So, 8 hours/day X 60 min/hr X 20 days/month = 9600 minutes of available capacity per circuit. Take this and times it by 23 available circuits.

Full capacity = 9600 minutes/month X 23 circuits = 220,800 circuit min/month.

But, if I'm only using 20,000 min/month, it means that I'm only using about 9% of the full capacity of the T-1.

I'm starting to think that it might be worth a try. Drop the LD T-1 circuit and go back to the single PRI. I can give it a month or two and see how it works out. If I have too many problems, I could obtain the second T-1 circuit.

Are there any good documents that explain how to use the Meridian analysis functions?



mike

 

[ArchibaldJCox]

My company has a traffic calculator that may be able to help you. (We also conduct traffic studies if you would like) But here's the link:

http://www.infoplusonline.com/trafficcalc/index-v.html

Now look through for your highest hour of usage (Maximum busy hour) If you total up the minutes that trunks were in use for a given hour, then you can convert the minutes to CCS (C (one hundred) Call Seconds)

1 CCS = 100 seconds

To covery minutes to CCS, multiply minutes * 100 / 60

e.g. 300 minutes of usage (accross all trunks) in an hour * 100 / 60 = 500 CCS

Just put in your CCS value in and your equipped trunks (46 or 48) and you'll get recommendations for P.01, P.02 and P.03 grades of service.

We use Erlang B, Erlang C and Poisson methods to calculate these values. In order for this to provide any value to you at all, you need to be able to accurately figure your absolute MAXIMUM BUSIEST hour of traffic.

P.01 means that there's a 1% chance of blockage on your trunks.
P.02 means there's a 2% chance, and P.05 means there's an 8% chance of blockage.

Hopefully this helps, if not, let me know. Our traffic studies are VERY affordable, and if you were to eliminate an entire T-1, then it would be paid for within the month!

Archie

P.S. Make sure you account for Incoming and Outgoing traffic in CCS when you put the numbers in, otherwise the value will mean NOTHING!

 

[ArchTelecom]

http://WWW.Erlang.Com

also has some nice traffic tools.

 

[jgideon]

It may be that you only need the 'local' PRI's worth of trunks but what is the difference between your switched and dedicated LD rates?

 

[hawks]

The toll calls on the local PRI does not include any local free calls. So make sure you include any minutes there too. Like jgideon said also look close at the switched rates and compare the 2. One question what toll calls do you have going out the local span?

 

[mjpearson]

Sorry, I haven't been very responsive. It's been a long, long couple of weeks. Two hard disk failures in our server; my laptop went dead; the spare laptop went dead; the main time clock went dead; our Ethernet switch went dead; the wireless Ethernet connection to our guard shack is picking up radio noise. If it weren't for bad luck, I wouldn't have any luck at all.

Back to the problem:

Again, I'm relatively new to this stuff but here is what I do know: The PRI (switched T-1) handles DID (100 numbers); The routing table is configured to accept our local calls (within our own area code) and toll calls to the "intra-lata?" The LD (deidcated T-1) accepts anything beyond our intra-lata.

When trying to do the preliminary estimates, I looked at the intra-lata calls and came up with 5000 minutes so, I doubled it to account for local calls. I'm wondering if I should have tripled it?

Now I need to figure out how to get the reporting out of the switch. I was given MAT software to do it but...sure enough, it's on my laptop (the dead one mentioned above). I'll have to break out the books and see if I can do it via command line.


mike