Register (or login) on our website and you will not see this ad.
|
|
|
OK, some months further on. The good news is that the problem is both less marked in day to day experience and the BQM confirms that - some days have been really excellent in the past few weeks there's been only a few days that were iffy, although sometimes they were very iffy. But the underlying problem hasn't gone away - I've engaged with Aquiss, raised a ticket, tried the things suggested but none of these has made a difference. Martin has looked at things, but his view is that this isn't down to either Entanet (my original suspicion, now disproven), nor to Aquiss. Without any indications of the cause I tried running Pingplotter to both TBB, and to Google when the problem was evident, and what that shows is that one particular server is seeing occasional spikes. As I can't post a screenshot, I'll type below an example of the key data, apologies for the formatting, best I can do without a fixed width font.
Server/hop Avg ms Max
My router 1.0 1.8
78.33.253.10 7.6 17.4
78.33.253.1 8.0 13.1
188.39.127.242 8.1 15.2
172.30.2.53 9.6 71.1
172.30.1.148 7.7 10.6
195.66.236.240 8.6 28.3
80.249.97.84 9.3 41.2
80.249.106.141 8.3 16.3
Now, I know that there's always a bit of leeway on hops and traffic routing is subject to variability, but the persistent thing is that when I observe problems, Pingplotter always picks out 172.30.2.53 as having much higher maximum value, as in this case where in a few minutes of monitoring that hop was up to 71.1 ms.
Thoughts, opinions?
Thanks, Andrew
|
|
|
what that shows is that one particular server is seeing occasional spikes.
That doesn't tell you anything useful. It's not a server, it's a router; and whilst routers forward packets in hardware, but they send response packets from the CPU (and at low priority).
In this case, it's sending an ICMP error "time to live exceeded" in response to a traceroute probe. It means that the CPU on that router is busy from time to time, nothing more. It's definitely not a problem. The fact that packets reach the subsequent hops without any additional delay proves that.
best I can do without a fixed width font.
To get a fixed width font, surround your text with "code" and "/code" in square brackets.
For posting images, use an image hosting service like imgur and post the URL.
|
|
|
OK, some months further on. The good news is that the problem is both less marked in day to day experience and the BQM confirms that - some days have been really excellent in the past few weeks there's been only a few days that were iffy, although sometimes they were very iffy. But the underlying problem hasn't gone away - I've engaged with Aquiss, raised a ticket, tried the things suggested but none of these has made a difference. Martin has looked at things, but his view is that this isn't down to either Entanet (my original suspicion, now disproven), nor to Aquiss. Without any indications of the cause I tried running Pingplotter to both TBB, and to Google when the problem was evident, and what that shows is that one particular server is seeing occasional spikes. As I can't post a screenshot, I'll type below an example of the key data, apologies for the formatting, best I can do without a fixed width font.
Server/hop Avg ms Max
My router 1.0 1.8
78.33.253.10 7.6 17.4
78.33.253.1 8.0 13.1
188.39.127.242 8.1 15.2
172.30.2.53 9.6 71.1
172.30.1.148 7.7 10.6
195.66.236.240 8.6 28.3
80.249.97.84 9.3 41.2
80.249.106.141 8.3 16.3
Now, I know that there's always a bit of leeway on hops and traffic routing is subject to variability, but the persistent thing is that when I observe problems, Pingplotter always picks out 172.30.2.53 as having much higher maximum value, as in this case where in a few minutes of monitoring that hop was up to 71.1 ms.
Thoughts, opinions?
Firstly - one stray packet in middle of a traceroute isn't much to worry about. The average latency isn't too far out. Routers aren't designed to respond TO that traffic to the router - they focus on delivering packets THROUGH the router.. You'd have to look for an increase consistently from a point to draw conclusions. It doesn't rule an issue out but neither does it rule it in.
172.30.2.53 is an internal IP probably on Entanet's network
195.66.236.240 is our LINX LON2 router (route back appears to be via LINX LON1)
188.39.127.242 is Entanet
Looking at your latt 24 hours I don't think there's anything wrong with your BQM.. you have a few spikes in max latency but nothing of major concern. A couple of dropped packets maybe in 24 hours..
seb
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
|
Thanks for that, Seb. I'd agree there's no current problem, indeed the last 24 hours have been very consistent and no cause for complaint, it's just that there's odd days when things aren't so rosy. Although the figures I quoted show the maximum, it's not about that single highest packet, but that particular IP.
So when I am noticing problems I might run a Pingplotter session, and that will show multiple packets with higher latency spikes at that IP, usually not enough to lift the average, but certainly enough to cause break up or inconsistencies in Teams, and inevitably an impact on gaming. So eg 6th or 13 March were disappointing in my view for a new-ish FTTP connection, and the figures quoted were from a short run on the 13th (incidentally, contrast 1 March with 8 March). I can't see anything obvious in terms of time-of-month, but it seems when problems do occur they're weekdays and mostly working hours.
Andrew
|
|
|
Pages in this thread: 
Print Thread
