Packet Drops every 2 hours

Anyone aware of an issue with BQM? I see packet drops every 2 hours like clockwork, then the pattern breaks, and starts again.

See my graph

I know it's not my connection as my ISP (Andrews & Arnold) also use a Firebrick to produce similar graphs, that one shows no packet loss.

Just thought I'd report it as the product is still in beta.

Thanks,

Ellis

Random plusnet customer
http://www.thinkbroadband.com/ping/share-thumb/7fdfa...

No sign of a generic issue.

and cannot spot anything on the thumbnails for the others at http://craigswebsites.co.uk/ping/

Thanks for the reply, interesting nobody else is seeing this.

Here is the graph from my ISP for yesterday (ignore the drops between 3-4 am, my FTTC is still going through it's 10 day training period). It shows no packet loss.....

It's a mystery, but I'll sleep at night.

Thanks,

Ellis

Could it be anything to do with you using an IPV6 address for the monitor? (Assumed that from the monitor name)

If you perform a tracert to the ping monitor, it could be down to any of the hops that are external (between the 2 ) It looks like a switch somewhere that aint working to good, but it's unlikely to affect the performance of the serviceYou usually only see this type of thing occasionally (IPV4),

Good thought, I'll setup an IPv4 monitor to be sure.......

Just a thought- when you set up the IPv6 monitor, did you simply accept the one that tbb showed in the IP box?

Because that one belongs to the computer, not the router. If you want the BQM to ping the router you have to enter the router's IP address manually.

If that's what happened, tbb and AAISP may not be pinging the same thing.

In reply to a post by billford:

Just a thought- when you set up the IPv6 monitor, did you simply accept the one that tbb showed in the IP box?

Thanks for the reply. Yes, I am monitoring the routers WAN ip, so I know that AAISP and TBB are monitoring the same thing.

UPDATE : IPv4 monitor has the same issue as the IPv6 one.

Thanks,

Ellis

In reply to a post by EllisD_UK:

UPDATE : IPv4 monitor has the same issue as the IPv6 one.

I thought it would, mine usually look broadly similar.

I'm stuck... I can only think that either:

a) tbb and AAISP pings are being handled differently by your router (seems pretty unlikely!) or

b) there's something odd on the part of the route to you that tbb uses and AAISP doesn't. Which seems equally unlikely, so there's probably a third possibility that I haven't thought of

How often do AAISP ping your router? Is it possible that the brief periods of packet loss are "sneaking between" the pings?

In reply to a post by billford:

How often do AAISP ping your router? Is it possible that the brief periods of packet loss are "sneaking between" the pings?

AAISP don't use normal ICMP pings for this, they use LCP echo between their gateway FireBrick and the end user router: http://aaisp.net.uk/kb-broadband-cqm.html

Same pings as our pingbox uses
http://www.firebrick.co.uk/products_6000.php

In reply to a post by MrSaffron:

Same pings as our pingbox uses
http://www.firebrick.co.uk/products_6000.php

I didn't think your pingbox could do LCP ECHO as you weren't the ISP handling the L2TP session? Surely you're doing standard ICMP ECHO ?

http://www.thinkbroadband.com/ping/share/e9b12ba304d...

I've been getting little tiny packet loss spikes on my graph for a while now (at least over a week) if you look above. they are tiny but consistent throughout the whole day but I haven't noticed any effect from them.

Ignore the spikes about 3pm and 1am, that was me restarting my modem.

My bad

In reply to a post by jchamier:

AAISP don't use normal ICMP pings for this, they use LCP echo between their gateway FireBrick and the end user router: http://aaisp.net.uk/kb-broadband-cqm.html

I'm sure that must imply something relevant to the OP's problem but I'm out of my depth, so I'll leave it to those who know what they're talking about!

Bear this in mind: A packet loss is where the time to respond is greater than a defined parameter; A delayed packet is one where the response time is a lot greater than normal. There has to be a limit where a delayed packet becomes a lost packet.

I had a 40/2 connection for around 18 months and for ALL of that time I had a latency spike, from my normal average of 15ms to 150ms, every 45 minutes. Remember, the point plotted is the average of 100 pings.

My 40/2 Broadband Ping

Tried various suggestions from here and had BTs second line support working on it too - they could not find the reason or how to resolve it. We tried different modems (Huawei and ECI), various routers and more but it remained. Had the latency been a little greater I would have seen packet losses.

I moved from 40/2 to 80/20 and it just disappeared ... same modem, same router ...

My 80/20 Broadband Ping

Ultimately, does it cause you any problems? If not, it is one of those strange quirks you will have to live with.

In reply to a post by billford:

I'm sure that must imply something relevant to the OP's problem but I'm out of my depth, so I'll leave it to those who know what they're talking about!

I'm hardly an expert, but I understand that LCP is link control protocol, and your router and ISP communicate using this protocol and the PPPoA session is within this. Your TCP/IP communications (including ping) are then encapsulated within the PPP (point-to-point protocol).

So an LCP ECHO is well outside the TCP/IP network and only something an ISP can do, and definately not something ThinkBroadband as a third party website can do.