Hello there. I've been with Vodafone about 3 months, and lately particularly at weekends or certain evenings I see a latency spike on my second hop on the network.
That is absolutely fine - it's a non-problem.
To expand: routers have two parts, a Data Plane which forwards packets, and a Control Plane which handles management.
Traceroute works by manipulating the TTL on outbound packets. When your client sends packets with a TTL of 2, it will be decremented to 1 by the first hop router, and decremented to 0 by the second, at which point the packet needs to be discarded and an ICMP error "Time-To-Live Exceeded" sent back to the sender. That's what traceroute prints; the source address of the packet tells you the router's IP address.
The decrementing of the TTL is performed in hardware (Data Plane), but the sending of the ICMP error message is done by the control plane.
The control plane of the router is just a CPU. It could be doing other things, like processing SNMP messages, or handling routing protocols like IS-IS or BGP. Sending ICMP error messages is very low in its list of priorities. So if it's busy doing other things, the ICMP message will be delayed slightly.
Some routers intentially have rate-limiting on ICMP processing, as an anti-DoS measure. It's possible that when you do this two times in quick succession, the router is intentionally delaying the second ICMP response. Some routers will drop the ICMP response entirely.
However, this has absolutely zero impact on forwarding of packets. If you had been pinging 1.1.1.2 with a TTL higher than 2, then the packet would have passed through the hop 2 router with no additional latency, because it wouldn't touch the control plane at all.
The variable latency when pinging 84.65.128.1 is exactly the same issue. Asking a router to *respond* to an incoming echo request is requiring work of the CPU, and is a very much a low priority task for it. If it's in the middle of handling a complex SNMP query, say, then the CPU scheduler will very likely delay the response. Again, the fact that the CPU is busy from time to time still has no impact on forwarding packets *through* the router, and therefore has zero latency or other impact on reaching any other destination on the Internet.
I've tried to log a case with Voda, and also a complaint but it has been absolutely horrendous. Just repeatedly coming back to things in the house, when everything is actually fine.
And they are entirely correct (well, not that there's a problem in the house, but that this traceroute and ping demonstrates no problem at all). Please don't try to escalate this specific report any further; you'll just be wasting their time and yours.
If you have a problem with gameplay, then report the actual problem with gameplay. You've already showed stable latency to 8.8.8.8. But there may be a problem between Vodafone's network and the network where the gaming server is, or the person you're playing against. That's where they'd need to look.
I wouldn't expect much movement on this from a low-cost provider like Vodafone though.
Edited by candlerb (Mon 27-Apr-26 14:29:10)
Print Thread
lt567
