How to measure leased line speed using ping output

To All Geeks out there!

I have got a 64K leased line and I would want to verify that the line is really giving me 64K but I don't have money to buy equipment or software to use for measuring the line speed.

I want to use the output from a simple ping to compute or calculate the speed of the line.

How do I get about doing it?

Please remember that I am not a network guru and would appreciate full interpretation of the ping output.

My output from a DEC Unix box looks more or less like:

PING ( 56 data bytes
64 bytes from icmp_seq=0 ttl=62 time=38 ms
64 bytes from icmp_seq=1 ttl=62 time=18 ms
64 bytes from icmp_seq=2 ttl=62 time=18 ms
64 bytes from icmp_seq=3 ttl=62 time=19 ms
64 bytes from icmp_seq=4 ttl=62 time=18 ms
64 bytes from icmp_seq=5 ttl=62 time=18 ms
64 bytes from icmp_seq=6 ttl=62 time=19 ms
64 bytes from icmp_seq=7 ttl=62 time=21 ms
64 bytes from icmp_seq=8 ttl=62 time=19 ms
64 bytes from icmp_seq=9 ttl=62 time=20 ms
Who is Participating?
scraig84Connect With a Mentor Commented:
Well a ping is only a relative measuring tool.  In other words, you can use it to comparitively test agaist other ping output you've done in the past, but it gives you no hard answers as to the amount of bandwidth, congestion, delay, etc that may have been used or experienced along the way.  Typically, you can use output from the routers to see how much traffic is actually moving accross the interfaces.  If you have access to the routers and know how to get interface statistics from it, you could send a large quantity of data across the line and look at the stats to see if it is around 80% of the total speed (serial lines are saturated at around 80%).

One thing I'm curious about - why would you question the throughput of a leased line?  Leased line service is the most trustworthy service you can get and all of the bandwidth is specifically dedicated to your use.  Usually, people want to do the testing I was mentioning above on services such as Frame Relay.

Hope that is of some help!
tmutsambwaAuthor Commented:

I have previously taken the line speed to be 64K but I don't seem to see the difference between my 64K and one of my friends' 32K.

I am beginning to think that I am looking at a dishonest service provider who is charging me for 64K bandwith when I am only getting around 21K.

If I do a ping at my friend's place, the time taken for 64Bytes is much less than I am getting and yet we are in the same town and our HQs are both in a towm 275km from where we are.

So, I still insist on getting how to use ping output. I don't want to measure how much is passing through the line but rather the maximum that the line can handle.
So is this an Internet T1?  I am assuming if you are comparing ping tests, it is to the same destination.

You can compare if you want, but there are so many variables that a ping test won't work.  Let's take the example of two different people (you and your friend) at different locations pinging the same destination.  At any given time there will be a difference in:

The amount of traffic on the local segment
The total amount of traffic flowing to each ISP
The amount of traffic on each ISP's core network
The amount of traffic moving from the ISP's core to the Internet backbone
The path chosen by the ISP to a given destination
The amount of traffic on the specific core provider's network(s).
This goes on and on until it hits the destination.  Then you get the same variables on the way back.

So, what I am saying is that the amount of variables that you would have to eliminate before a ping test would be a valid testing tool is so large that it is unrealistic to expect anyone to take the data seriously.  

Now, you may be right that an ISP may be oversubscribed and even though you may be getting full use of the leased line to the ISP, they might be too busy to handle the entire amount of traffic of your line.  I dealt with an ISP once that was giving great deals on 1Mb DSL.  Problem was that their connection to the DSL provider was a single T1 serving over 40 customers meaning nobody got full usage of their DSL line's capabilities.

Unfortunately, if you are indeed dealing with the scruples of an ISP, there are very few tools you CAN use due to the high amount of variables on the Internet.  If an ISP wants to do so, they can and will screw you.  Therefore, it is typically in your best interest to go with the ISP that comes highly recommended, rather than the one that offers up the best deals.
Improve Your Query Performance Tuning

In this FREE six-day email course, you'll learn from Janis Griffin, Database Performance Evangelist. She'll teach 12 steps that you can use to optimize your queries as much as possible and see measurable results in your work. Get started today!

samriConnect With a Mentor Commented:

There are time I had that kind of curiosity too.  But somehow I will go with scraig84 - most telcos are pretty much OK.

However, if you insist on checking whether your link can *really* max to 64, perhaps you could try MRTG.

To test, just schedule a number of simultaneous download, and measure the bandwith.  Normally the link might not be able to hit 100% capacity (64K), but you might be looking at 90% mark.  If you are getting 50K++, it should looks good to me.

It's just a suggestion though.

The ping output gives you the round trip time between your computer and the target machine.
In the networking when you talk about the delays we can generally talk about the propagation delay(the delay that passes as your data moves on the physical wire) and the transmission delay(the delay that passes as you put your data onto the physical wire).

In order to have an estimate for your case you may try to ping the fist machine on the other side.(e.g. yoyur default gateway) and also the distance between your machine and the HQ.

Also, this estimate will not be a perfect one, of course...


stevenlewisConnect With a Mentor Commented:
As pointed out above, speed is dependent on many things, but here are a couple of websites to check your speed, then check on your freinds (for comparison's sake)
It's futile using a ping to test line speed (as opposed to end to end delay in which case it is valid).

Assuming you have a link from A to B and you send a ping from A to B which it then returns to A.

Most ping packets run around 50 bytes although this can be changed from the command line.

So, a 50 byte packet traversing a 64kbps link will induce a delay of around 0.0078 seconds, therefore you would see a network component for the round trip of around 0.0156 seconds.

The same packet on a 32kbps link will induce a delay of 0.0156 seconds with a RTT (excluding processing by the remote node) of 0.0312 seconds

I suspect your ping program is going to be hard pressed to tell the difference!

As samri says, mrtg is your best bet but it relies on SNMP so it depends on whether you have SNMP available on your system and the communications stack.

Cheers - Gavin
64 bytes from icmp_seq=0 ttl=62 time=38 ms
you sent 64 bytes of info (the first of 10) from
the time to live was 62 ( the maximum routers it can pass thru before being discarded)
and the response was 38 milliseconds
As stated above, the ping utility is a diagnostic tool to test connectivity, not really a speed testing tool
This old question needs to be finalized -- accept an answer, split points, or get a refund.  For information on your options, please click here-> http:/help/closing.jsp#1 
Post your closing recommendations!  No comment means you don't care.
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.