When I ping Google my average ping is 29ms ? isn't that kinda of a lot ?

Discussion in 'Networking & Security' started by Subzerok11, Apr 4, 2019.

  1. Subzerok11

    Subzerok11 Gawd

    Messages:
    543
    Joined:
    Aug 13, 2014
    I have spectrum cable and living within 5 miles from Columbus Ohio (kind of a big city) centrally located.

    Also with further testing using the Speed Test from Ookla with a spectrum server located in Columbus I'm getting about 15 ping in the morning at 9:00 AM and in the evening at around 5:30 PM I'm getting like 27 ping. Shouldn't I be getting like around 5-10 ping at the most being so close ?

    I'm thinking about getting a really long cat 5 cable and run it directly from the back of the modem from the basement and see if it's something inside the house or is it spectrum cable.

    What do you guys think is it worth further investigation ?

    Thanks
     
  2. FNtastic

    FNtastic [H]ard|Gawd

    Messages:
    1,340
    Joined:
    Jul 6, 2013
    With cable internet? Nope. Fiber? Maybe.
     
    Cmustang87 likes this.
  3. MMitch

    MMitch Gawd

    Messages:
    640
    Joined:
    Nov 29, 2016
    Tested the default server using Cable internet 400mb and got 9ms from montreal, CAT6 wired (yeah overkill, same price so...) thru ASUS router.
    This may be highly ISP dependent. Mine is VERY reliable and aggregate the street cable to optical lines.

    I would test from another location to confirm the numbers, if they're much lower and connecting directly to modem changes nothing, I would ask that a technician verify the line integrity along with grounds (Those caused me a LOT of trouble in the past. I mean grounds from the actual devices outside the house... blew a few TVs HDMI Inputs from their decoder).
     
  4. Subzerok11

    Subzerok11 Gawd

    Messages:
    543
    Joined:
    Aug 13, 2014
    Just checked Fortnite and it says my game connects at 26ms.

    I guess I should mention my house was wired with cat5 not cat5e. So my router is in the basement with a store bought extension 15 foot new cat5e cable then I got that connected to a old cat5 non e cable with rj45 I put at both ends of this cat5 myself. Then that goes to the second floor to a outlet. Then I got another store bought extension 15 foot cat5e cable going to the back of the PC from the outlet. So basically two store bought cat5e extensions cables and one old cat5 with the rj45 connectors I put myself. I tried to replace/rewire the outlet it's just not possible, plus the original installers used these U type nails that are holding the cat5 super tight against the 2X4 in wall at a few places.

    Not sure if this info is important or not. But I am getting the speeds I'm paying by the way.
     
  5. Cmustang87

    Cmustang87 [H]ardness Supreme

    Messages:
    4,399
    Joined:
    Oct 4, 2007
    20-30ms is perfectly acceptable latency for a cable internet provider. 1-3ms is what you'd expect if you have fiber.
     
  6. mlcarson

    mlcarson Limp Gawd

    Messages:
    289
    Joined:
    Jul 12, 2009
    Come on people, the underlying technology doesn't guarantee lower latency. It's mostly about location and the BGP routing path to what you are trying to get to. This is mostly out of your control unless you're open to moving. Latency variability is probably due to congestion along the way. You're not going to shave much fixed latency off in your own setup unless you're using Wifi or have a very bad router.
     
    mwarps likes this.
  7. FNtastic

    FNtastic [H]ard|Gawd

    Messages:
    1,340
    Joined:
    Jul 6, 2013
    I didn't see anybody say "guarantee". The physical connection certainly makes a difference in latency. With all other variables the same, light pulses over fiber travel with less latency than current over a copper wire. Period. If you have evidence to the contrary, you're well on your way to rewriting the laws of physics.

    "Speed of electricity through copper is about 0.5c. Speed of light through fiber is about 0.7c."

    I'll also add that, while not a blanket statement, typically with cable internet, the customer is affected much more greatly by congestion at high traffic times, further degrading latency.
     
    Last edited: Apr 5, 2019
  8. EniGmA1987

    EniGmA1987 Limp Gawd

    Messages:
    192
    Joined:
    May 2, 2017
    Try pinging 1.1.1.1 and 9.9.9.9. If you get pretty much the same ping then it is your ISP that is the cause of the latency.
     
  9. Nicklebon

    Nicklebon Gawd

    Messages:
    550
    Joined:
    May 22, 2006
    SMH ... nothing but misinformation in these replies. OP if you want to test your link the only address you should be pinging is the default route of your router. This is your first public hop and it and only it will determine the quality of YOUR link. Any broadband service should return single digit pings to the first public hop. Any hops beyond that are a measure your ISP's internal network and their connections to other networks. As a user of consumer class service there is little to nothing you can do to influence an ISP to fix an issue there but, any problems there will be picked up as they will impact all users of the ISP. I'll also add that your testing should be first done from your router and then from your PC. The delta is the latency of your network and your router.
     
    mwarps, ChoGGi and DrLobotomy like this.
  10. mwarps

    mwarps [H]ardness Supreme

    Messages:
    7,001
    Joined:
    Oct 6, 2002
    Big LOLS at thinking Columbus Ohio is a big city, or centrally located. Your packets are likely going to chicago or DC, both about 30ms from you.
     
  11. tedych

    tedych Limp Gawd

    Messages:
    372
    Joined:
    Jan 18, 2013
    Not entirely and always true.
    I have a 1Gbps fiber to the home. The ISP have configured their GWs to respond to pings with lower priority and as a result many customers were "complaining" thinking they have problems with their connection. Pings to these machines (at least from customers) can vary greatly depending on load. BUT if you ping another (close) hop or some local website/server you always get constant 2-3 ms.

    Also the "delta" of his internal network shouldn't be higher than 0.5 ms to 1ms and wouldn't impact the observations especially if in a modem situation.
     
    FNtastic likes this.
  12. scrappymouse

    scrappymouse n00b

    Messages:
    32
    Joined:
    Mar 18, 2016
    As someone who works at a Tier I ISP, listen to this guy.

    Also if you want to take it a step further, do a traceroute to Google, take note of the second hop...first hop will be your router's IP, the very next hop will be the gateway into the ISP's network. Now ping that second hop and see what latency you get....if it's good...you can then be sure that the latency isn't a cause of your internal network setup, Just realized the quote above said default route of your router, he is correct. I just said it a different way :)
     
    Last edited: Apr 7, 2019
  13. Nicklebon

    Nicklebon Gawd

    Messages:
    550
    Joined:
    May 22, 2006
    Prioritization is only an issue if the router is busy. If the router is busy then your packets through the router will suffer even more than your packets to the router. Looking at jitter is a reasonable way to gauge router health. Frankly, you should not be pinging anything that is truly local as those are other users of the service. The ISP is not running any local webservers on your side of the access router. As for your asinine comments on local latency, your assuming everything is working correctly which is not a safe assumption since the OP is likely concerned about something. OP one thing you should realize is that it is almost certain your ISP has oversubscribed the circuits feeding your service. It is also a safe assumption that your neighbourhood is is oversubscribed on the DOCSIS side as well. This means during times of high use your latency will certainly increase and it is also very likely your peak throughput will drop as well.

    upload_2019-4-7_14-58-1.png

    OP this is a screen shot of my firewall. The interface wan1 is actually Sprectrum, wan2 is a fiber connection to another ISP, FEX1 is an LTE link. The lab section is good to look at because my lab circuit is provided by the same the ISP as my fiber connection but located in a different city. Still, notice how much lower the ping and the jitter are vs the Spectrum link which has to go through a peering point to reach my lab. The Google section you'll note is much more similar though still favours the fiber link. In general I really would not get hung up on ping times.
     
  14. tedych

    tedych Limp Gawd

    Messages:
    372
    Joined:
    Jan 18, 2013
    I didn't say everything IS Ok on his local side but if working correctly it should be Ok. This he can gauge by pinging for example his local router before the modem, and he should see <1ms. I just said that pinging the first hop in his connection to the ISP could not be very indicative in principle. He may have several issues along the way.
    If I ping google, it is 2-3ms because google has servers all over the world and I have one locally. If the OP is farther from the closest google server, 29 ms being on modem (arghhh) may be normal. He should ask other customers of his ISP or someone in his neighbourhood using another ISP, to be sure whether its jist his ISP routing packets this way, no matter how far the server is physically. traceroute could help.
    I wonder when this Docsis and cable modems crap will go in museums at least in develop(ed | ing) world...
    I'm on LAN or fiber since 2006 and it's unusual to find an ISP with modems already.
     
  15. mlcarson

    mlcarson Limp Gawd

    Messages:
    289
    Joined:
    Jul 12, 2009
    And the difference in .5c and .7c is 29% and is only relevant over extreme distance. Cable congestion is normally at the local node. Congestion not at the local node is generally at the peer to peer connections of the various ISP's. Your router processing at each hop is doing much more to add latency than any difference in fiber/copper and even the cable companies are using fiber to the node. And of course distance and number of hops are the biggest contributors of latency that are not congestion based. Nicklebon seems to know what he's talking about. The only other thing that I'd add with Cable Internet is the possibility of signal ingress issues leading to packet loss or just bad signal levels that are varying based on time of day/weather issues. There's nothing in the OP's question though that would lead me to believe that this is anything but local node congestion leading to the variability in latency at peak usage times like 17:30 combined with an unrealistic expectation on what latency should be at non-peak times.