Linode network ping
So I've been using Linode for a long time, especially since here in Missouri, I'm able to get a 12ms ping to Linode servers in Dallas, TX.
Well, that was until today…
I've tested many Linode servers in Dallas, TX today and tried pinging them and now I'm getting a 29ms ping. Very frustrating, especailly with what I'm doing.
Did Linode change their network or something recently? Is anyone experiencing the same thing?
I'm using AT&T U-Verse internet 1gbps with a LAN connection. I've always gotten 12-13ms connection to linode, and it's 29ms now. Very very consistent, but 29ms.
There are a few potential reasons as to why your ping times are now longer than they previously were. We haven't recently made any changes to our network that would affect ping times to Dallas. The most common reason for latency to occur is a slowdown along the route traffic takes to and from your Linode. You can run a set of MTR reports to see each network hop and how long it takes to reach them. You also may want to reach out to your ISP, especially if there's a chance they may be throttling traffic. We'd also be happy to take a look at your MTR reports and your specific Linode if you open a Support ticket. We can't guarantee that we'll be able to make an impact on issues outside of our network, but we can help provide some insight into where any potential issues are occurring.
ping/ping6 response times are not a very good measures of network performance. A much better tool is traceroute IMHO.
I'm using AT&T U-Verse internet 1gbps with a LAN connection.
My guess is that ATT has
- throttled your traffic as @jyoo suggests;
- introduced new filters for your traffic for their own marketing, network-management or security purposes; or
- introduced new hops in the path between you and the Dallas datacenter.
All of the above could be true for any network ATT hands your traffic off to for routing to the Dallas datacenter and back. Also, most ISPs (large and small) give a very low priority to handling ICMP echo (ping) requests/responses.
If you have IPv6 set up, try doing your measurements with ping6… I can almost guarantee you'll get better results ;-)
traceroute will show you the performance of each hop in the route between your ISPs presence node and your Linode. You can do the same measurements for IPv6 with traceroute6. traceroute normally uses TCP but you can force UDP by using traceroute -U.