General Discussion
  >> Fibre Broadband


Register (or login) on our website and you will not see this ad.


  Print Thread
Standard User smurf46
(member) Wed 21-Sep-11 14:19:22
Print Post

The most stupid question ever (probably)


[link to this post]
 
Does your broadband need a regular good run, like a dog?

Since I've had an FTTC service, generally I've driven it to the max on almost daily simultaneous video downloads during my morning off-peaks. Although the local loop is poor, and the few speeedtests I've done are variable (though BT speedtester consistent) its performance seems consistent. Only noticeable change is that according to TBQM ave latency doubled from 9-18, presumably as interleaving introduced (as I'd expect). Could my practice be helping? If so, are speedtests themselves too transient to have a beneficial effect?

No-one was ever convinced by what I said, but by what they understood.
Administrator MrSaffron
(staff) Wed 21-Sep-11 14:35:12
Print Post

Re: The most stupid question ever (probably)


[re: smurf46] [link to this post]
 
No need for a regular run, it should be as good after not being used for a week as it is if you left a HD video streaming for a week.

The TBQM is an end to end measure, not just monitoring the FTTC part of the service, i.e. all the links inbetween and interleaving may have been the change, as could routing changes by the ISP on their own network.

Andrew Ferguson, andrew@thinkbroadband.com
www.thinkbroadband.com - formerly known as ADSLguide.org.uk
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User WWWombat
(committed) Mon 26-Sep-11 15:24:54
Print Post

Re: The most stupid question ever (probably)


[re: MrSaffron] [link to this post]
 
While I agree with Andrew's answer from a logical & technical basis, I have some experience of the opposite...

When my FTTC was first installed, the line was experiencing around 4% packet loss 24/7 (as measured by the BQM graphs), no matter what was going on on the line. This was fixed by DLM after 48 hours, by introducing interleaving, and increasing latency - again visible on the BQM graphs.

However, for 1 hour-long period, the packet loss dropped to almost zero. Was it really a coincidence that I was downloading a Fedora DVD during that hour? I suspect not, bu8t I don't have much of an explanation either!

Guess what time the download was...


Register (or login) on our website and you will not see this ad.

Administrator MrSaffron
(staff) Mon 26-Sep-11 15:44:14
Print Post

Re: The most stupid question ever (probably)


[re: WWWombat] [link to this post]
 
Packet loss or just delayed long enough to timeout on a standard ping?

Some power saving modes might mean on a quiet line (in terms of traffic) a delay to the first packet after a gap in time.

If this is the case, with enough FTTC users on our BQM it should be visible.

Andrew Ferguson, andrew@thinkbroadband.com
www.thinkbroadband.com - formerly known as ADSLguide.org.uk
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
Standard User WWWombat
(committed) Mon 26-Sep-11 16:19:40
Print Post

Re: The most stupid question ever (probably)


[re: MrSaffron] [link to this post]
 
The BQM displayed the packet loss as a solid band of red at the top of the graph, 24x7. It was variable, but always present, almost never going over 10% and averaging around 4%.

When DLM added interleaving (I guess), which increased latency (not a guess), the packet loss disappeared entirely.

So - I would guess it *isn't* a timeout issue where the ping response doesn't quite make it - as adding latency into the issue wouldn't fix the problem. And of course the BQM is going on 24x7 too - which ought to prevent any power-saving modes. Shouldn't it?

BQM after 48 hours - at the time of the addition of interleaving.
  Print Thread

Jump to