Traffic Prioritisation and 20CN

[#1]

@Kelly,

Yes it does - thank you.

I now infer that fixed IP address and 20CN users use the 'old' network for back haul and RAS termination ...

Now going off topic ... would that mean that Traffic Prioritisation would sill have merit for those users even if they are on 'unlimited' products.

Moderator's note by Mike (Mav): Changed thread title to reflect issue after split

[#2]

That depends on the profile associated with their product in the traffic management platform.   The current unlimited product doesn't have any applied. (As I understand it)

[#3]

Moderator's note:

Split to it's own thread by request.

[#4]

I asked for this to be split from the IP thread so we could deal with the separate questions about Old vs New network.

In Plusnet we don't really recognise the terms "old network" and "new network" any more.  The only real difference now is the kit which the customer connections terminate on.  

Previously "Old Network" would refer to any user who was provisioned on the WBMC Shared Platform which meant their connection would come via L2TP to our old City Lifeline and Telehouse datacentres.   We then added in the use of WBMC Dedicated at our new datacentres and started provisioning customers on this.  This was the "New Network".   This meant there was suddenly different bits of infrastructure that customers would be going over based on if they were "old" or "new", and the distinction was meaningful.

Nowadays, the only difference is down to the point where an IP address is assigned.  Customers on usage limited broadband products, IP Stream (20CN) or who have a static IP have their sessions terminated on our BNG end points.  These customers have effectively an additional piece of kit (the BNG) added into their route out to the internet where the IP address is assigned, but apart from that are going through the same network devices as other customers

All customers (except IP Stream) can go through one of WBMC Shared, WBMC Dedicated or Market B, irrelevant of where there their connection terminates. All customers go via our new datacentres and access the wider network through the same peering links.

Does that make sense?

[#5]

Thanks @Kelly.  Are all customers terminating on Plusnet equipment on BNGs now?  I think there was a thread a week or two ago where someone was still going (as far as they were saying at least) through an AG gateway.

[#6]

Yeah, the E320s aren't in the network anymore.

[#7]

,

Given that business accounts still benefit from traffic prioritisation would I be.correct in inferring they also terminate 'in house'?

Where the termination is in house, are two levels of RAS taking place? One at the PoP which either assigns a dynamic IP and establishes the PPP session or creates the L2TP tunnel to the BNG which assigns the IP address and establishes the PPP session?

Reason for this question is to understand if the allocation of a 172.16.x.x gateway as here - https://community.plus.net/t5/Broadband/Loss-of-connection/m-p/1465583#M316700 - is indicative of a failure in the latter scenario?

[#8]

---

@Townman wrote:
Given that business accounts still benefit from traffic prioritisation would I be.correct in inferring they also terminate 'in house'?

No, that's based on the same rules as above. (Limited product, Static IP, IPStream Connect). Prioritisation can be done where required via ether method.

Where the termination is in house, are two levels of RAS taking place? One at the PoP which either assigns a dynamic IP and establishes the PPP session or creates the L2TP tunnel to the BNG which assigns the IP address and establishes the PPP session?

In the exchange each user goes through a BRAS or MSE; on WBMC dedicated/Market B the PPP can be terminated there from a Plusnet Dynamic IP address Pool, or tunnelled via L2TP to our BNGs which then assign the IP.

Reason for this question is to understand if the allocation of a 172.16.x.x gateway as here - https://community.plus.net/t5/Broadband/Loss-of-connection/m-p/1465583#M316700 - is indicative of a failure in the latter scenario?

---

Could be any number of things wrong here. They need a fault raising.

[#9]

Lots of people will now have a 172.16.x.x IP address for their gateway, it's not a problem.

[#10]

I must be working with out of date perceptions then ... getting one of those addresses used to indicate a BT routing fault.  That user reports the router diagnostics all coming up green aside from...

Connectivity to Gateway (172.16.10.30) failure

[#11]

I think the "BT routing fault" would be when you get a 172.16.0.0 - 172.31.255.255 IP address as your WAN IP address, not the gateway.

I think the router diagnostics are broken and one of the tests will fail even when there's nothing wrong (perhaps due to the gateway not responding to pings).