My department has recently moved offices. The previous office was fine, but in the new office, the ping times from my PC to the 2 data centres usually varies a lot, like this:
Reply from 123.123.123.123: bytes=32 time=52ms TTL=123
Reply from 123.123.123.123: bytes=32 time=11ms TTL=123
Reply from 123.123.123.123: bytes=32 time=16ms TTL=123
Reply from 123.123.123.123: bytes=32 time=9ms TTL=123
Reply from 123.123.123.123: bytes=32 time=9ms TTL=123
Reply from 123.123.123.123: bytes=32 time=13ms TTL=123
Reply from 123.123.123.123: bytes=32 time=20ms TTL=123
Reply from 123.123.123.123: bytes=32 time=28ms TTL=123
Reply from 123.123.123.123: bytes=32 time=10ms TTL=123
Reply from 123.123.123.123: bytes=32 time=19ms TTL=123
Reply from 123.123.123.123: bytes=32 time=9ms TTL=123
Reply from 123.123.123.123: bytes=32 time=33ms TTL=123
Reply from 123.123.123.123: bytes=32 time=17ms TTL=123
Reply from 123.123.123.123: bytes=32 time=23ms TTL=123
Reply from 123.123.123.123: bytes=32 time=31ms TTL=123
Reply from 123.123.123.123: bytes=32 time=8ms TTL=123
Reply from 123.123.123.123: bytes=32 time=20ms TTL=123
Reply from 123.123.123.123: bytes=32 time=10ms TTL=123
Reply from 123.123.123.123: bytes=32 time=11ms TTL=123
Reply from 123.123.123.123: bytes=32 time=20ms TTL=123
Reply from 123.123.123.123: bytes=32 time=9ms TTL=123
Reply from 123.123.123.123: bytes=32 time=9ms TTL=123
Reply from 123.123.123.123: bytes=32 time=16ms TTL=123
Reply from 123.123.123.123: bytes=32 time=23ms TTL=123
Reply from 123.123.123.123: bytes=32 time=10ms TTL=123
Reply from 123.123.123.123: bytes=32 time=15ms TTL=123
Reply from 123.123.123.123: bytes=32 time=22ms TTL=123
Note that I have changed the IP address. The server I am pinging has a physical IP address and is not being used. I get similar results when I ping other servers and at any time of the day.
We are not fully utilising our bandwidth.
If I ping another server (located in a 3rd data centre), the ping times are higher (around 100ms) but far more consistent (to within about 3ms).
So does this indicate that there is a potential issue between my new office and the 2 data centres?
ping time can vary so much according to factors outside your control that it's worth aggregating results to eliminate inconsistencies - there is a useful linux tool called mtr which is included in many distributions and aggregates traceroute and ping into a single utility. It will give you the best and worst ping times, as well as average and standard deviation. While not conclusive, it will offer a more balanced view.
Variance in ping times can be indicative of several things, but the fact that you get the same behaviour regardless of which server in that data centre you ping seems to indicate that there may be an issue with a link or device along the path.
It's worth doing repeated traceroutes to the sever(s) in question, to see if you can identify a particular network device where latency is varying; or alternatively, it could be that there are multiple paths to this location, one of which experiences generally higher latency.
Are you seeing any packet loss when you perform extended pings? If so, this may be indicative of a congested link.
Could you provide a bit more detail about the connectivity between your office and your data centres? Are you using leased lines? Internet VPN? Have you spoken to your company's NOC/network administrators? They may be aware of this issue already.
Possibly, but not certainly. How is your office linked to the data centers? Pings over a site to site VPN can vary quite a bit considering they travel over the internet.
Are you seeing any other performance problems or issues with the connection besides the pings? Those response times are still all under 100ms, which could just be normal for the new links you are on.
Just another thought, if one of the links you are going through is near full capacity, pings are usually given a lower priority, which could explain the variation.
But like others have said, there are so many factors at play, it could mean something but also it may not mean anything.
What does traceroute show? it will give you pings to each hop.
I typically see a lot of variation on wireless connections, which I consider normal.
I think, this potentional issue is called load-balancing router. It routes your request sometimes to the 1st interface and sometimes to the 2nd. They have different latency.
For example, we have optics on wan1 and ADSL on wan2. For one particular server ping through wan1 was 17 ms and through ADSL - 112 ms.
Do you have some sort of firewall between the data centers doing any sort of filtering? Have you done a tracert between the offices and seen where in the transaction the "hold ups" are happening? This might help you better figure out where your problem lies.