My Ping and I

[#1]

@jbligh87 wrote:

Legend of Zelda, Ocarina of time was my childhood...ping didn't effect that game *contemplates fishing out the old N64*

Recently plugged my N64 in a few days ago, and started playing Goldeneye again.

Gandalf by day, secret agent by night.

Also, if it was all owned by BT/OR, how can Sky remote in and reset their DLM during a call?

As far as I'm aware Openreach are the supplier for all providers including LLU, and LLU providers have their own back-haul network.

[edit]

out of interest I ran ping plotter with a 2.5 second interval as svrN suggested, it highlighted hop 7 with a low packet loss although still shows spikes in ms in the lower graph. Please see attached.

Thanks for that. Does seem a bit to show some packet loss there, not sure if this shows something not right on the BT network but we'll see what our networks team come back with next week

[#2]

Hiya,

Don't worry too much about the packet loss on the core - they don't prioritize ICMP packets (pings) hence why they can spike and even drop them completely and not effect the end-hop.

I firmly believe the latency is introduced before hop 2 on openreaches network, as your routing from BT Openreach to plusnet is transparent through the cabinet, exchange and via EAD/WBC/WBMC until it hands over to the ISPs PoP - this would account for a number of ISPs having the same issue (BT Retail, Zen, Plusnet, IDNET).

https://kitz.co.uk/adsl/wbc_wbmc.htm

[#3]

For LLU Openreach still manage the phone line from your house, through the PCP (the green cabinet with the number painted on, wider than it is tall) back to the exchange where it connects to the LLU providers equipment.

For VDSL your phoneline "splits" at the larger cabinet next to the PCP cabinet, fibre carries the VDSL part back to the exchange, and analogue voice goes to the PCP - for LLU this then terminates on the LLU equipment (via some vlan routing magic AFAIK).

For non LLU, its Openreach back haul all the way to the nearest point of prescence for the ISP, depending on what the ISP bought (EAD/WBC/WMBC).

Once my new line is in (June 4th) if it doesnt have the same issue then I can rule out the "first mile" and presume its the openreach back haul as it will be in the same cabinet and openreach have already told me my existing line is fine - if it does still have the problem i'll have more weight to get openreach to investigate the cabinet.

I'd be interested to see if you guys having the issue have had a G.Fast pod installed on your PCP - if you good image search G.Fast pod you'll see that its a narrower add-on box to the PCP with vents.

[#4]

Thanks again!

So would being on a dynamic IP or a static IP matter then or not?

(If not, what's the point in Plusnet trying them?)

On a Plusnet connection, a dynamic IP does not always, due to the router (in the network) sometimes being busy, respond to ICMP pings as seen in hops 3 and 4, whereas a static IP is routed differently.

We also see the static IP routing via a traceroute on hop2.

Also, Plusnet host some of the content within their network, things like Netflix for example. Would having this content within their network also be a contributing factor in your opinion?

(As Plusnet do host the content, I'd guess to say BT are doing the same)

Or should all of us be pointing a finger solely at Openreach?

(edit. Totally missed your reply directly above, just read it now though)

[#5]

Static vs dynamic doesnt mean much (it could be used to fix which gateway you connect to by telling openreach which PoP to hand you over to which would alter your first hops routing), unless plusnet have specific higher priority routing for a range of IP addresses they give out as static - routers will optimize the routing path via BGP (border gateway protocol) and OSPF (open shortest path first) etc - but it could be that the static IP allocations are linked to the pro add on? So the routers prioritise the static IP range - but i have no clue what plusnet do.

Hop 2 and 3 are the plusnet core, hops 4, 5, and 6 are the BT Retail core (PlusNet is owned by BT Retail so its not unusual to see them using BT Retail core).

In terms of "mirroring" content, they'll have a data centre connected to their core, as its more efficient and cost effective to keep that data within their network, instead of using their peering agreements to transfer so much data accross multiple peered networks to reach netflix - this again though, to reach you, as you are not LLU will have to go through Openreach WBC etc..

(Edit) they also have far more control of the quality of service being provided, and capacity as its within their network

[#6]

unless plusnet have specific higher priority routing for a range of IP addresses.....

I really don't know if they do higher priority that way, but as you mention they do sell the Pro Addon

I didn't think static and the pro addon were linked, as dynamic users can have this too.

(but I now wonder if static and pro are indeed linked... hmm?)

I keen to find out if your new line does something or not

My connection has had a lot of work done, but my problem is cable fill. In my case someone has got to foot the £10k bill before that gets resolved.

We were going to leave just over a month a go and go back to LLU. As our connection was able to manage 30 days without a drop we decided to stay.

[#7]

10k ouch..

Dont worry i will be reporting back here as soon as its in after testing.

If it is causing you super pain you can mitigate it heavily (its still spikey but not as bad) by limiting the media streaming devices to 2mbps down and 0.5mbps up, i currently use the ax11000 to limit the client bandwidth which then through the wan port connects to an ubiquiti edge router 3 lite which runs SQM (uses fq_codel to queue packets) and that then connects to the modem. So my spikes are from 15ms to 50ms (more manageable). The ER3 Lite forwards all ports to the ax11000 so it isn't double NAT'd.

[#8]

I said a bit more than ouch!!

I really hope it goes well

Oh, I thought the BTSH6 did all that!!

I've never gone down the route of a non ISP router myself, though I guess at some point I really should

(but will more than likely wait until we move back to LLU)

[#9]

Finally got some time to shift my setup and run some tests today. All over the place. 

1x Twitch stream running.

1 <1 ms <1 ms <1 ms ARCHER_VR900 [192.168.1.1]
2 56 ms 66 ms 37 ms 172.16.17.64
3 * * 9 ms 129.hiper04.sheff.dial.plus.net.uk [195.166.143.129]
4 73 ms 76 ms 77 ms 128.hiper04.sheff.dial.plus.net.uk [195.166.143.128]
5 9 ms 9 ms 9 ms 195.99.125.136
6 161 ms 159 ms 169 ms peer8-et-0-1-4.telehouse.ukcore.bt.net [109.159.252.106]
7 * * * Request timed out.
8 10 ms 12 ms 9 ms po4-31.core-rs2.thdo.ncuk.net [80.249.97.85]
9 9 ms 9 ms 9 ms pingbox1.thinkbroadband.com [80.249.99.164]

1 <1 ms <1 ms <1 ms ARCHER_VR900 [192.168.1.1]
2 4 ms 4 ms 4 ms 172.16.17.64
3 * * * Request timed out.
4 10 ms 10 ms 10 ms 128.hiper04.sheff.dial.plus.net.uk [195.166.143.128]
5 9 ms 9 ms 9 ms 195.99.125.136
6 150 ms 151 ms 94 ms peer8-et-0-1-4.telehouse.ukcore.bt.net [109.159.252.106]
7 * * * Request timed out.
8 11 ms 13 ms 13 ms po4-31.core-rs2.thdo.ncuk.net [80.249.97.85]
9 55 ms 9 ms 9 ms pingbox1.thinkbroadband.com [80.249.99.164]

Tracing route to pingbox1.thinkbroadband.com [80.249.99.164]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms ARCHER_VR900 [192.168.1.1]
2 18 ms 4 ms 5 ms 172.16.17.64
3 * * * Request timed out.
4 10 ms 10 ms 10 ms 128.hiper04.sheff.dial.plus.net.uk [195.166.143.128]
5 33 ms 45 ms 45 ms 195.99.125.136
6 10 ms 10 ms 10 ms peer8-et-0-1-4.telehouse.ukcore.bt.net [109.159.252.106]
7 * * * Request timed out.
8 12 ms 12 ms 10 ms po4-31.core-rs2.thdo.ncuk.net [80.249.97.85]
9 40 ms 44 ms 42 ms pingbox1.thinkbroadband.com [80.249.99.164]

Trace complete.

C:\Users\becke>tracert pingbox1.thinkbroadband.com

Tracing route to pingbox1.thinkbroadband.com [80.249.99.164]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms ARCHER_VR900 [192.168.1.1]
2 4 ms 4 ms 4 ms 172.16.17.64
3 * * * Request timed out.
4 10 ms 10 ms 10 ms 128.hiper04.sheff.dial.plus.net.uk [195.166.143.128]
5 9 ms 9 ms 9 ms 195.99.125.136
6 10 ms 9 ms 10 ms peer8-et-0-1-4.telehouse.ukcore.bt.net [109.159.252.106]
7 * * * Request timed out.
8 10 ms 14 ms 12 ms po4-31.core-rs2.thdo.ncuk.net [80.249.97.85]
9 10 ms 9 ms 9 ms pingbox1.thinkbroadband.com [80.249.99.164]

[#10]

Cheers for the detail @svrN, Keep us posted on how your new line goes it’ll be interesting to know if that fixes the issue. 

With regards to dynamic IPs and static IPs as far as I’m aware there’s no difference in priority, and we also don’t have any traffic prioritisation or management policies for unlimited residential packages apart from the option to apply the Pro Addon upon request although that just priorities traffic on a home network. 

[#11]

https://gyazo.com/0aff599dc5e8661963ce46a2899abc3b

some solid jumps for the bank holiday

[#12]

@Gandalf Ran the BT wholesale test earlier this morning if you wouldn't mind having a look. 

[#13]

@gdaviez, we can't open media/file sharing sites internally I'm afraid.

Can you add the content to a post?

@ElGappo I've tested your line and no performance result has come through. Did you complete the further diagnostics at the end of the BT speedtester and did it provide an IP profile? That'll be an indication as to whether it's worked

[#14]

@Gandalf Yeah I did the further diagnostics test. Running it again now

[#15]

@Gandalf Is there any update from the marvel team...?