Point to point network speed test tools

[tarquel]

Hi all

Just been wondering if anyone uses such a thing and if so, what do you use and how reliable is it?

What I'm after is sort of a ping-style tool that on one machine, you start it in server mode, and then on the other pc, you can start the tool in client mode and it will record how fast it takes information to travel between the two, bandwidth, faults on the way, etc.

Also a hardware style tool - one to act as the reciever and one as the sender to record a similar thing.

Any ideas on this as it'll help out no end on a few matters on the network?

(hope it all makes sense)

Cheers
Nath

[Geoff]

for bandwidth testing, fear the power of bing!

terror ~ # bing -s 5 terror ipcop
BING    terror.internallan.org (10.0.0.5) and ipcop (10.0.0.1)
        5 and 108 data bytes (1648 bits)

--- terror.internallan.org statistics ---
bytes   out    in   dup  loss   rtt (ms): min       avg       max   std dev
    5 42624 42624          0%           0.055     0.061    17.438     0.097
  108 42624 42624          0%           0.052     0.059    18.815     0.163

--- ipcop statistics ---
bytes   out    in   dup  loss   rtt (ms): min       avg       max   std dev
    5 42624 42624          0%           0.160     0.192    29.058     0.303
  108 42624 42624          0%           0.198     0.226    27.116     0.269

--- estimated link characteristics ---
host                              bandwidth       ms

ipcop                               43.368Mbps       0.105

http://www.freenix.fr/freenix/logiciels/bing.html

ping and traceroute should be sufficent for latency testing. If you need more than millisecond resolution though you will have to ditch windows and use a real OS. eg, here's what my Linux box manages:

terror etc # ping ipcop -c 5
PING ipcop (10.0.0.1) 56(84) bytes of data.
64 bytes from ipcop (10.0.0.1): icmp_seq=1 ttl=64 time=0.303 ms
64 bytes from ipcop (10.0.0.1): icmp_seq=2 ttl=64 time=0.276 ms
64 bytes from ipcop (10.0.0.1): icmp_seq=3 ttl=64 time=0.277 ms
64 bytes from ipcop (10.0.0.1): icmp_seq=4 ttl=64 time=0.332 ms
64 bytes from ipcop (10.0.0.1): icmp_seq=5 ttl=64 time=0.325 ms

--- ipcop ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4004ms
rtt min/avg/max/mdev = 0.276/0.302/0.332/0.030 ms

For faults, you can get a bit of mileage out of a PC running as a layer 2 bridge. For example here at Carr Hill we have a Linux box running ntop which regularly picks up stupidity on the network.

http://bridge.sourceforge.net/
http://www.ntop.org/

If you want anything more, buy dedicated hardware testing gear. Expect to have a large dent in your budget as a result.

[tarquel]

lol now I wonder who is a linux fan (baring the obvious avatars )...

I best explain why I need it then *sighs*

The network, for the most part, is fairly quick (well, quick enuff for now) but one link to the technology block is slower than a snail transfering the packets to and fro

See this post:
http://www.edugeek.net/index.php?nam...ic&p=1296#1296

All I want to do is get a tool that'll sit on my pc in the server room and run the tool on the machine in the tech block, then run a looooong (but not too long ) Cat5 cable from one switch to the other, and try the tools again to see the difference.

This'll then confirm the fibre / converters to be the problem.

So as you can see, just a little Windows tool that I can click "Recieve" on one end and "Send" on another would be well nice

I'll try your bing 'thing' tho Geoff Thanx

Cheers,
Nath

[Geoff]

All I want to do is get a tool that'll sit on my pc in the server room and run the tool on the machine in the tech block, then run a looooong (but not too long Wink) Cat5 cable from one switch to the other, and try the tools again to see the difference.

Its called ICMP. Its built into TCP/IP. Ping is sufficent for this.

[tarquel]

bah! - you're no fun lol :P

I want to be lazy here and have fancy readouts, buttons to press, combo lists and checkbox things :P

*sighs*

Nath

[Geoff]

You may wish to review some of the posts in the Cable Testers Thread then.

[mark]

Our local council have a piece of software running on a box with a 22" monitor with cable runs mapped out as lines. If there's a problem on a line it turns red/ flashes etc. No idea what it was - anybody know? - tho' I expect it's another bank breaker!

Ethereal looks excellent for this job tho' - although not graphical - I found some interesting stuff on my network today - like my Samsung printers with IPX enabled sending loadsa useless traffic.

Why do you need readouts at both ends Nath? - I use ping xxx.xxx.xxx.xxx -t and monitor that - though it's not fine enough for local traffic - better over distances. Geoffs suggestions sound perfect!

[Geoff]

Our local council have a piece of software running on a box with a 22" monitor with cable runs mapped out as lines. If there's a problem on a line it turns red/ flashes etc. No idea what it was - anybody know? - tho' I expect it's another bank breaker!

Sounds like HP OpenView. Opensource equivelent would be Nagios (Screenshots).

[mark]

Nagios looks really good Geoff. The requirements look a little imposing "if you don't intend to learn everything about it, don't bother" sort of thing. But the end result is what Nathan is after I think. Maybe overkill for a small site but very desirable all the same!

[Geoff]

Well with Nagios the best approach I've found is to start simple. Just montior your switches perhaps. Once you have the basics, extending it to cover your other systems isn't nearly as daunting as it looks.

Just remeber to backup those configs.

[spc-rocket]

Nath,

Try this, it will do from one endpoint to the other. Its free and runs on Windows.

QCheck -> http://www.ixiacom.com/products/perf...key=pa_q_check

You don't need to buy the other product icowarrior or something just use the Qcheck and will show a lot more than the usual ping utility.

Ashok.

[mark]

Looks intresting Ashok - one for the toolkit!

[ianniow]

Another vote for Qcheck from iaxia

[bossman]

@tarquel:

Surely with a fibre backbone you run layer 3 switches i presume then why not use the 2 switches which connect the Technology block to determine packet loss as these are the main connections to and from this would give you bonny graphs and things without having to spend money.

This is what I do for our network using Cisco assistant and can determine if there are any problems with the actual connection ports and components.