FTTC latency on new TTB service.

The difference between dynamic IP and static IP.

Dynamic
Tracing route to bbc.co.uk [151.101.128.81]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms ttrouter [192.168.1.1]
2 8 ms 8 ms 8 ms ae49-ner001.irl.as13285.net [78.144.1.41]
3 9 ms 8 ms 8 ms 78.144.1.40
4 14 ms 14 ms 14 ms 199.27.73.52
5 9 ms 10 ms 8 ms 151.101.128.81

Trace complete.

Static
Tracing route to bbc.co.uk [151.101.64.81]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms ttrouter [192.168.1.1]
2 15 ms 15 ms 15 ms gig-14-3-4004-rtr001.thn.opaltelecom.net [62.24.254.202]
3 16 ms 15 ms 15 ms host-78-144-1-63.as13285.net [78.144.1.63]
4 16 ms 15 ms 15 ms host-78-144-1-62.as13285.net [78.144.1.62]
5 * * * Request timed out.
6 16 ms 16 ms 16 ms 151.101.64.81

Nothing to write home about I suppose but why would the static IP have higher latency than the dynamic.

TTB have swapped me back to Dynamic for 24 hour monitoring and the above Dynamic trace is the result. I have asked if possible to make my current IP static, I await ticket response.

Try a tracert to the same ip rather than the name.

tracert 151.101.128.81

Tracing route to 151.101.128.81 over a maximum of 30 hops

1 <1 ms <1 ms <1 ms ttrouter [192.168.1.1]
2 8 ms 8 ms 8 ms ae49-ner001.irl.as13285.net [78.144.1.41]
3 8 ms 8 ms 10 ms ae49-scr002.msp.as13285.net [78.144.1.40]
4 14 ms 14 ms 14 ms 199.27.73.52
5 8 ms 8 ms 8 ms 151.101.128.81

Trace complete.

tracert 151.101.64.81

Tracing route to 151.101.64.81 over a maximum of 30 hops

1 <1 ms <1 ms <1 ms ttrouter [192.168.1.1]
2 8 ms 9 ms 9 ms ae49-ner001.irl.as13285.net [78.144.1.41]
3 8 ms 8 ms 8 ms ae49-scr002.msp.as13285.net [78.144.1.40]
4 14 ms 14 ms 14 ms 199.27.73.52
5 8 ms 8 ms 8 ms 151.101.64.81

Trace complete.

On Dynamic IP.

Edited by Banger (Thu 26-Nov-20 20:19:34)

In reply to a post by Banger:

Nothing to write home about I suppose but why would the static IP have higher latency than the dynamic.

Dynamic IP can give you an IP address from a local pool on the BRAS. If you are connected to a BRAS in Manchester, say, then you'll get an IP address from that pool.

Static IP is normally implemented by "tunnelling" your session (via L2TP) to a central BRAS farm, to which the IP block which contains your static IP address is routed. Judging by your traceroute, I'd say that is xxx-rtr001.thn.opaltelecom.net

That means your traffic goes

you -> local BRAS -> static IP BRAS -> destination

which may give a sub-optimal geographic path.

Thank you for that explanation so I am better off with a dynamic IP for lower latency? Incidentally I have worked out how to get TBB BQM working with NOIP on the TT router so I can keep the BQM alive when the address changes which it did overnight.

If low latency is important i.e. a gamer or other interactive activity then a decision to make there is.

A lot hinges on why you want static IP.

I do a lot of video streaming from iPlayer so I thought lower latency is better to the BBC. The reason my preference is for a static IP was for BQM but I think I have found a way to monitor the line with NOIP.

8ms of extra latency won't make one iota of difference for iPlayer streaming. You could use iPlayer streaming quite happily over a satellite link with hundreds of ms of latency. Buffering in your player takes care of all of that.

For interactive logins (e.g. ssh), I've never noticed a difference for anything below 20ms, and I often have to work on longer latencies than that. 150ms to West Coast USA is noticeable but still usable. Maybe it's because 300 baud dialup modems were what I started on

I'm not a gamer, which is where small amounts of latency might make a difference. However I've also heard that game servers add their own compensating delays to low-latency connections, just to even out the players.

Edited by candlerb (Fri 27-Nov-20 10:49:32)

That difference in latency will make no difference to streaming.

Given use of static was just BQM I'd say just stick with the simpler dynamic IP setup.

In reply to a post by candlerb:

For interactive logins (e.g. ssh), I've never noticed a difference for anything below 20ms, and I often have to work on longer latencies than that. 150ms to West Coast USA is noticeable but still usable. Maybe it's because 300 baud dialup modems were what I started on

I don't notice the difference between the 28ms I get at the moment from home to the system at work from the <1ms when I am in the office. It's all just fine. Graphical sessions are somewhat laggy, but I put that down to bandwidth rather than latency. Remote X11 sessions are impossible off campus for example as it's a super chatty protocol. It doesn't help having a HiDPI/Retina display