Hi All,
What would you recommend is the best tool to test the bandwidth and traffic on an MPLS network from one site to the Datacenter, we have an issue with one of our supplier where their application seems to be the problem in transferring data traffic to the datacenter as i believe that it cannot work across multiple subnets (some routing deficiency).
They have now brought the application server to the local site and it works perfectly fine as it is on the same LAN as the application server and clients are now on the same subnet. They used a script using LAN Speed Test , i told them that this was an incorrect test to do as the LAN speed test is only used on P2P networks not on an MPLS WAN network,
they came back with the below excuse from their findings:
The test was setup in a way to simulate a file transfer through the network. The file size was set to 2.6MB which equals the size of an image. This file has been sent every 15sec to the server and we have measured the response time in msec which you can than convert to network speed. This test doesn’t measure the entire network performance/speed but the response time of a file transfer
|
Date
|
Write min
|
Write max
|
Read min
|
Read max
|
|
20-Jun-2017
|
60ms (333Mbit/s)
|
250ms (48Mbit/s)
|
60ms (333Mbit/s)
|
180ms (111Mbit/s)
|
|
21-Jun-2017
|
60ms (333Mbit/s)
|
450ms (44Mbit/s)
|
60ms (333Mbit/s)
|
190ms (105Mbit/s)
|
|
22-Jun-2017
|
60ms (333Mbit/s)
|
850ms (23Mbit/s)
|
60ms (333Mbit/s)
|
190ms (105Mbit/s)
|
|
23-Jun-2017
|
60ms (333Mbit/s)
|
430ms (46Mbit/s)
|
60ms (333Mbit/s)
|
160ms (125Mbit/s)
|
|
25-Jun-2017
|
60ms (333Mbit/s)
|
250ms (48Mbit/s)
|
60ms (333Mbit/s)
|
180ms (111Mbit/s)
|
|
26-Jun-2017
|
60ms (333Mbit/s)
|
860ms (23Mbit/s)
|
60ms (333Mbit/s)
|
270ms (74Mbit/s)
|
|
27-Jun-2017
|
60ms (333Mbit/s)
|
1070ms (18Mbit/s)
|
60ms (333Mbit/s)
|
180ms (111Mbit/s)
|
|
28-Jun-2017
|
100ms (200Mbit/s)
|
110ms (181Mbit/s)
|
70ms (285Mbit/s)
|
100ms (111Mbit/s)
|
|
29-Jun-2017
|
60ms (333Mbit/s)
|
990ms (20Mbit/s)
|
60ms (333Mbit/s)
|
160ms (125Mbit/s)
|
|
30-Jun-2017
|
60ms (333Mbit/s)
|
850ms (23Mbit/s)
|
60ms (333Mbit/s)
|
170ms (117Mbit/s)
|
|
01-Jul-2017
|
60ms (333Mbit/s)
|
440ms (45Mbit/s)
|
60ms (333Mbit/s)
|
150ms (133Mbit/s)
|
As you can see above the min response time was 60ms (read/write). Max peaks as above.
Overall the network operated within 70-90ms response time (read/write) with the exception 27-28 June and 01 Jul (afternoon). During this time the response time was significantly lower than usual (> 110ms).
21-Jun highest network performance at ~ 8am-10am
22-Jun highest network performance at ~ 2am-10am
27-Jun from 8am –
28-Jun to 11pm low network performance
01-Jul from 2pm on low network performance
Supplier recommendation for the network speed is 1 Gbit/s. At no time the network achieved this performance. Typically it is working on a max 30% level of the recommended network speed.
Our Latency at site to the Datacenter is 3ms, how do we disprove this inaccurate test? Any Network Tool which can use i.e Iperf
Thanks
