- Interleaving - does it increase line speed?
- Posted by WCZ on February 21st, 2008
I can't seem to find a definate answer to this question. At the moment
interleaving is set to auto and I can't see from my router if its on or off.
If I were to force it to be on, would that offer any increase in sync speed?
Clearly I could just try it but it takes 48 hours for any change to take
effect so I thought I'd ask here first.
Cheers.
WCZ.
- Posted by The Natural Philosopher on February 21st, 2008
WCZ wrote:
at getting more packets trough a noisy environment.
It won't affect sync SPEED..that is a layer below the interleaving.
It PROBABLY slows down throughput a little, but makes it more reliable.
- Posted by Alastair on February 21st, 2008
"WCZ" <Dav123456@willnotwork.com> wrote in message
news:fpjn0s$kui$1@newsfeed.th.ifl.net...
Interleaving allows FEC to work better so a noisy
line will sustain a higher sync speed,
It increases latency somewhat.
- Posted by James Egan on February 21st, 2008
On Thu, 21 Feb 2008 11:52:06 +0000, The Natural Philosopher <a@b.c>
wrote:
BT reckon interleaving reduces the maximum achievable line rate on max
from 8128k to 7616k
Jim.
- Posted by Invalid on February 21st, 2008
In message <fpjn0s$kui$1@newsfeed.th.ifl.net>, WCZ
<Dav123456@willnotwork.com> writes
am sure)
Interleaving makes no difference to the sync speed per-se.
It trades off an increase in the overheads in the protocol in order to
reduce packet loss and improve link reliability. However that
improvement can stabilise the link at slightly higher sync speeds, and
stop the exchange from negotiating a lower speed because of dropouts. In
which case it can cause an increase in sync speeds.
Whether it increases or decreases download speeds depends on the level
of packet loss your line experiences without it. If it is significant,
then having interleaving on will (may) reduce the level of
retransmission caused by packet errors and consequently INCREASE
effective download speeds. OTOH if you get no real packet loss with it
off the protocol overheads will be lower, and download speeds may be
higher because a higher proportion of the link traffic is available.
I suspect you will find it hard to tell the difference in terms of
download speed unless your link is particularly good (off is faster), or
particularly bad (on is faster). The variations in speed you get from
the test site is probably going to swamp any difference caused by
interleaving.
I understand however it has a significant impact on round trip "ping"
times, as a consequence I believe gamers tend to turn it off and live
with the increased packet loss and link instability.
Do let us know the results of your experiments.
--
Invalid
- Posted by Steve B on February 22nd, 2008
"Invalid" <news@wisty.plus.com.invalid> wrote in message
news:gfgkzzBqMcvHFwa4@wisty.plus.com...
I believe that the actual latency with interleaving on depends on the type
of equipment used in the exchange. I get 21ms pinging my ISP with it on so
have never bothered to get it switched off.
- Posted by alexd on February 22nd, 2008
On Fri, 22 Feb 2008 06:29:48 +0000, Steve wrote:
He's talking about the line rate, not the IP throughput.
--
<http://ale.cx/> (AIM:troffasky) (UnSoEsNpEaTm@ale.cx)
08:38:18 up 13 days, 11:41, 2 users, load average: 1.15, 1.13, 1.05
Convergence, n: The act of using separate DSL circuits for voice and data
- Posted by Alex Fraser on February 23rd, 2008
"Steve B" <sbradsPUTTWENTYFOURHERE@ukfsn.org> wrote in message
news:fpm1n9$jd7$1@energise.enta.net...
[snip]
The latency introduced fundamentally depends on the interleave depth, which
could be fixed by the exchange equipment or (more likely I think) its
configuration, or chosen based on assessed line conditions.
FWIW, I also got the same RTT using interleave path, which suggests the same
interleave depth was used on this line as yours. As a comparison, on the
same line using fast path, I get 7-8ms.
Alex
