Need help with T1 performace testing

[azev]

When doing performace testing for T1 lines, what do you look for ? and what a good number and what is bad ? What I am trying to find out is what would be considered good T1 or a bad T1. We are going to use Firebird 6000A tester to test different mediums for transporting t1. These medium are wireless, fiber, copper, etc. We are going to run overnight test on all mediums and create a matrix off all the result. I have an idea on what to look for, but I am sure someone here had previous experience with what i am about to do, and hopefully willing to chip in some info

Thanks

 

[spidey07]

A good T1 is one that runs with no errors.

A bad T1 has errors.



this is what bert tests are for.

 

[azev]

hehehe no kiddin spidey, but you have a valid point. I am actually trying to find out what is how much error rate is considered acceptable, how much are too much ?

 

[spidey07]

any errors on a bert test is no good IMHO.

And why are you trying to transport T1 over broadband technologies? I can't see the need. A T1 is a TDM technology and uses the appopriate muxes do deliver the bearer and clock.

 

[azev]

well, what we're trying to do is to transport T1 data over wireless network. We are using a propietary box to convert T1 digital packets to udp ethernet packets. From the converter box we are using wireless network to do the final transport to the destination location, and use another propietary box to convert the data back to T1.
Since wireless are not a reliable method of delivery, we're trying to find the treshold of how much errors are acceptable before we're actually loosing data/frames.

 

[jersiq]

Azev

Working for a wireless company I have tested 1000's of T1's in my short career. We use T1 as a transport from a switching office to individual cell sites.

I agree with Spidey that no errors at all is the only acceptable result for a good span (T1).

I would suggest running a very stressful pattern such as 55-octet (or 55DLY) if your test sets support it. Other patterns to run are all 0's (this proves a B8zs span) and of course, QRS on your first 1/2 hour of testing. Will you be running the tests head to head, or to a loop? I think I am confused by your setup. Is the local TELCO providing a T1, or is this an internal network? I just want to check to see if I am giving you some good information as all the T1's I test are delivered to me off a SONET ring.
Also, what type of wireless are you using. If it's unlicensed microwave then there may be some other issues.

 

[azev]

Well, we are using Firebird 6000A to generate T1 traffic and we are using IP tube to convert T1 packets to udp ethernet packets. Then we are using a layer 2 encryption box and connect to wireless keg. The other side uses the same exact device. We are trying to determind using the firebird tester how much errors are acceptable when transmiting T1 packets over wireless media. We did a 24-48hr testing using all the equipment less the wireless and we got zero errors. When we introduce the wireless radio in the equation, we started seing some errors. We need to find some sort of baseline of what is the max error we can have before the T1 data/frame are bad, and need to retransmit. According to the research that I've did, 10 -9 through 10 -12 errors is considered great T1, but ITU recommendation o.150 say 10 -3 through 10 -5 is still acceptable. I need to find more supporting document such as the ITU recommendation o.150, but so far that is all I got. I am hopping somebody here have done the bold attempt we're trying to do ( transporting T1 via wireless ).

 

[w0ss]

I will second the any errors are to many. If you have any errors now what happens when the weather is bad? If your wireless gear is already showing errors then you will see even more errors when it rains/snows. Also are you testing this where it will be deployed? If not make sure you do some type of site survey to make sure there isn't some nice high tension wires overhead in a valley that gets 100" of rain a year.

 

[azev]

Now are u saying wireless are error free ? According to my research testing of a few major wireless equipment, BER/CRC errors on wireless network are normal to a certain degree. Btw we did a site survey and determine the best location to put the AP. Again what I am trying to find out is a treshold number for T1 errors, before the data or packets are corrupt and need to be retransmit (which will reduce available bandwidth).

 

[spidey07]

Originally posted by: azev
Now are u saying wireless are error free ? According to my research testing of a few major wireless equipment, BER/CRC errors on wireless network are normal to a certain degree. Btw we did a site survey and determine the best location to put the AP. Again what I am trying to find out is a treshold number for T1 errors, before the data or packets are corrupt and need to be retransmit (which will reduce available bandwidth).

An error on a T1 will corrupt the packet. from there layer3 or 4 will retransmit.

there should be zero errors on a T1.