Very high ping to Jira servers

[CliffViegas]

My team is overseas and for them their ping is great, they can use Jira OnDemand with no problems.

I'm in Brisbane, Australia and unfortunately due to the high ping it's difficult for me to effectively project manage them :(.

Is there anything you can recommend that could help me use Jira without the large delays between actions?

I recall seeing a version of Jira that we could setup on our own servers?

Thank you :)

===== Ping Details ====

PING https://netnow.jira.com (63.246.22.216): 56 data bytes

64 bytes from 63.246.22.216: icmp_seq=0 ttl=54 time=265.818 ms

64 bytes from 63.246.22.216: icmp_seq=1 ttl=54 time=264.178 ms

64 bytes from 63.246.22.216: icmp_seq=2 ttl=54 time=263.709 ms

64 bytes from 63.246.22.216: icmp_seq=3 ttl=54 time=263.565 ms

===== Tracert Details ====

traceroute to https://netnow.jira.com (63.246.22.216), 64 hops max, 52 byte packets

1 10.0.0.1 (10.0.0.1) 1.840 ms 2.044 ms 2.650 ms

2 xx.xx.xx.xx (xx.xx.xx.xx) 79.831 ms 15.666 ms 18.203 ms

3 xx.xx.xx.xx (xx.xx.xx.xx) 20.634 ms 19.304 ms *

4 bri-sot-wic-crt1-gi-2-1-0.tpgi.com.au (202.7.171.217) 20.102 ms 15.807 ms 16.184 ms

5 bri-nxg-alf-crt2-pos-5-1.tpgi.com.au (202.7.171.245) 20.386 ms 18.867 ms 20.130 ms

6 syd-sot-ken-crt4-te-10-0.tpgi.com.au (202.7.171.133) 28.583 ms 28.442 ms 29.873 ms

7 10gigabitethernet3-1.core1.sjc1.he.net (72.52.66.21) 222.836 ms 216.761 ms 214.508 ms

8 10gigabitethernet2-1.core1.sjc2.he.net (72.52.92.118) 211.624 ms 211.400 ms 211.610 ms

9 * 206.111.6.165.ptr.us.xo.net (206.111.6.165) 220.965 ms 236.261 ms

10 207.88.14.225.ptr.us.xo.net (207.88.14.225) 215.450 ms 213.277 ms

207.88.14.233.ptr.us.xo.net (207.88.14.233) 224.495 ms

11 te0-5-4-0.rar3.denver-co.us.xo.net (207.88.12.85) 271.829 ms

te-3-0-0.rar3.denver-co.us.xo.net (207.88.12.57) 284.780 ms

te0-5-4-0.rar3.denver-co.us.xo.net (207.88.12.85) 268.314 ms

12 ae0d0.mcr2.marylandheights-mo.us.xo.net (216.156.0.182) 364.813 ms 282.583 ms 300.611 ms

13 ae1d0.mcr1.marylandheights-mo.us.xo.net (216.156.1.89) 297.755 ms 269.547 ms 299.626 ms

14 * 206.181.23.22 (206.181.23.22) 263.978 ms 263.926 ms

15 63-246-15-30.xiolink.com (63.246.15.30) 265.983 ms 263.712 ms 282.976 ms

[Harry Chan · Accepted solution]

Hi Cliff, the OnDemand platform is hosted @ Contegix @ St Louis. The time it takes from St Louis in the USA to Australia is very long as evident in the trace. I also don't believe their service offers a gigabit port, which may be a problem if you are attaching large files / video content etc.

If this is a problem for you, I'd recommend finding a JIRA hosting solution in either Australia or somewhere in the US with low latency to Australia/Asia and also the US.

If you need such an enterprise solution, drop us an email at sales@queryfoundry.com. Our pings from a server in the US is ~180ms to Sydney.

[Felipe Cuozzo]

Hi Cliff,

Would you be able to provide more details on what type of actions you're seeing high response times?

We use our own OnDemand US based servers for our internal development so we dogfood it in a very similar link conditions (Sydney->US), most of the time the response times are reasonable and we can use it without long waits. It would definetly help to get a bettter understanding of your usage pattern to see if we can improve the software side of things ;)

This is a ping from my machine inside Atlassian's Sydney office:

ping https://netnow.jira.com

PING https://netnow.jira.com (63.246.22.216): 56 data bytes

64 bytes from 63.246.22.216: icmp_seq=0 ttl=49 time=204.036 ms

64 bytes from 63.246.22.216: icmp_seq=1 ttl=49 time=204.210 ms

64 bytes from 63.246.22.216: icmp_seq=2 ttl=49 time=202.991 ms

Getting a local server will only shift the latency times to your overseas peers so not very helpful IMHO.

Glad to help.

Cheers,
Felipe Cuozzo
OnDemand Team

[Harry Chan]

Hi Felipe,

ping https://netnow.jira.com

Ping from PIPE Networks (TPG)

Success rate is 100 percent (5/5), round-trip min/avg/max = 204/331/400 ms

Ping from iinet (which routes through telstra here)

Success rate is 100 percent (5/5), round-trip min/avg/max = 200/225/244 ms

Ping from internode (which uses Southern Cross cable)

Success rate is 100 percent (5/5), round-trip min/avg/max = 200/203/204 ms

The above pings go to the ISP only. You'll need to add 20~40ms going to any user.

Those are the if not only major 3 routes going from Australia to US.

Cliff is on PIPE/TPG and hence the high latency. It looks like Contegix doesn't peer with whoever PIPE peers with in the US.

(We've tested our network with all 3 routes and has been very fast from US as I've stated.)

[Felipe Cuozzo]

Hi Harry,

Thanks for the detailed explanation I didn't know about the fact we only had 3 major routes from Australia to US.

But again if there are 5+ people in the US and 1 in Australia, it makes more sense to me to host the server in the US where more people can benefit from the proximity/low latency.

My questions were around the actual product usage to see if there is something we could improve, like reducing number of requests, improve caching, etc.

[Harry Chan]

There is a 4th route that was in the works, but it got canned. There were quite a lot of backers, but the funding didn't work out.

It make sense to host the server in West Coast of USA so it's fast for people in the US and quite nice for Australia too.

You can definitely improve the speed of web servers, coding, etc, but latency doesn't improve unless you have better network/location.