Latency problems to various sites

Options
Zhain
Zhain Posts: 19 Participant
edited March 26 in Connectivity

Hello!

I live in Upstate NY (Albany area) and lately since this past Tuesday, I've been experiencing spiky ping and latency issues with different services. Mainly this affects me playing the game FFXIV, where unsteady ping can make things near unplayable.

At first I thought it might have been me, so I went through all the troubleshooting steps including restarting modem and router, as well as connecting directly to the modem, etc. Nothing seems to help.

I ran a few traceroutes and downloaded a program called PingPlotter to try and figure out what was happening. From the best I can tell (I'm not an expert), there seems to be a local Spectrum node causing issues. Please see attached pictures for plots of ping tracing to twitter.com and the FFXIV servers. In both cases it is the third hop (lag-62.hlmnnych01h.netops.charter.com).


Is there a way I can ask about node status/routing? I may try a VPN next to see if the routing is truly the issue.

Comments

  • RAIST515O
    RAIST515O Posts: 98 Contributor
    Options

    Would note they moved hosting for the Americas from Montreal to California years ago... and latency/jitter from the east coast went to crap when it happened, especially during "prime time" and other similar high traffic windows. Took about a year of testing/ tracking /reporting to get routing to the Eidos-Montreal nodes down into the 60's or less from central east coast... and then SE moved them on us when they restructured things because of all the EU issues (they were originally hitting Montreal also).

    Such wonky latency/jitter is a common problem getting over there, and not just for Southern Cali either--the Seattle exchange can be a nightmare too, not to mention some paths through Texas. Level3/CenturyLink has struggled with the southern and northen paths for decades... ALTER.NET/Verizon from the north east is no stranger either... even Zayo can run into issues going through the heartland and more northern routes that drop down. The whole area can become a clusterf-- sometimes... can just spike with too much traffic and the route announcements can take a while to adjust the routes. Throw in some funky environmental events like the recent massive solar flare, extreme temps/weather... and things can become wildly unreliable at the drop of a hat.

    Just saying... sometimes there literally isn't much that your last mile ISP can do directly. Barring sending technicians on site to confirm signal/line/equipment quality and such... best they can do is send notice to upstream peers in case they are not already aware of the congestion or other stability issues occurring along certain segments.

    Should also note that the ICMP ECHO requests are routinely ignored by many nodes to conserve resources--especially in times of high utilization. They are primarily configured to focus on forwarding packets, not responding back to the sender. That is why even though a hop in a route may show high latency in replies (or even timeouts), the endpoint can still return considerably lower numbers. It does not necessarily point to a "problem" greater than that node experiencing an increased load that may be causing some abnormal queuing periods.

    Notice the exchange point into SE shows a timeout... while running the range from the 80's to timeouts with 66% packet loss... but the lobby right after that is 80's/90's with NO packet loss reported.


    To be showing a consistent response under 100ms to Southern Cali from the east coast is actually pretty good for some locations, all things considered.

  • RAIST515O
    RAIST515O Posts: 98 Contributor
    edited February 24
    Options

    Original post explaining things in more detail stopped showing when I tried to edit it for clarification... could have been fat fingers on a small phone screen. ¯\_(ツ)_/¯


    The TLDR:

    10 minute sample to the lobby server in Southern California from the east coast... pings at the endpoint in the 80's/90s with zero packet loss tracked.


    Does not appear to be a "problem" in route beyond nodes programmed to focus on forwarding packets and to potentially ignore ICMP ECHO requests in times of high utilization.

  • RAIST515O
    RAIST515O Posts: 98 Contributor
    edited February 24
    Options

    For contrast, let the router run a continuous ping for a few minutes from here in SC... averaged right at 94ms, 25ms jitter (ran mid 80s up to 112ms spikes).

    Pretty common to that area from here.

    Was sooo much nicer back when it was hosted in Montreal... especially after the routes got cleaned up.

    Here's a link to "live" data you can bookmark for comparisons going forward:

    https://wondernetwork.com/pings/New%20York/Los%20Angeles

  • Jaleesa_F
    Jaleesa_F Posts: 377 ✅ Verified Employee Moderator
    Options

    Hello @Zhain

    Welcome to Community Forums! Sorry to hear that you're experiencing connectivity issues, we are happy to look into this for you. Are you using the Spectrum wireless router? Or do you have your own router connected? Can you go to speedtest.net, run the speed test then send a screenshot of the results? 

     

  • Zhain
    Zhain Posts: 19 Participant
    Options

    @RAIST515O Thank you for taking a look, I appreciate it. I'm wondering what has been causing my recent issues, then, in a game instance I can see ping spike up to ~1000ms sometimes, and no one else is having these problems. I myself have been fine before like Monday/Tuesday. My snapshot may not have captured it accurately.

    @Jaleesa_F I am using the Spectrum modem connected with my own router. Speed doesn't seem to be the problem, it's more latency to specific services, but I'll post a speedtest result here anyway.

This discussion has been closed.