So the incoming line is going to the Cisco? Which then feeds the UDMP which then feeds the switch? (Line <> Cisco <> UDMP)

Or is it the other way round, line <> UDMP <> Cisco <> switch?

In reply to a post by pluralist:

So the incoming line is going to the Cisco? Which then feeds the UDMP which then feeds the switch? (Line <> Cisco <> UDMP)

Or is it the other way round, line <> UDMP <> Cisco <> switch?

It a fully managed service rather than "wires only" leased line. So the former.

In reply to a post by Pheasant:

In reply to a post by pluralist:

So the incoming line is going to the Cisco? Which then feeds the UDMP which then feeds the switch? (Line <> Cisco <> UDMP)

Or is it the other way round, line <> UDMP <> Cisco <> switch?

It a fully managed service rather than "wires only" leased line. So the former.

This is correct.

(Also @Pheasant)
In which case surely Daisy are referring to the Cisco when they say how to configure the router? That will be the DHCP they are referring to, and the DHCP in the UDMP should be disabled?

In reply to a post by pluralist:

(Also @Pheasant)
In which case surely Daisy are referring to the Cisco when they say how to configure the router? That will be the DHCP they are referring to, and the DHCP in the UDMP should be disabled?

The Cisco wont be end user accessible/manageable. The Cisco effectively presents the service on the respective user ports. That's effectively the network demarcation point for the OP.

Any routing etc on "their" side would need to be checked and confirmed by Daisy.

You're saying the Cisco is bridged then? It did occur to me after posting that is what the OP is asking Daisy to confirm.

(Isn't that then a very expensive piece of kit for the purpose?)

Cisco or Juniper are pretty much the defacto standard fare for a managed leased line service here. Depending on what services are delivered, then it’s a little bit more than a bridge - typically the provider will use it to provision and manage VLAN(s) for the various service. Some customers may wish to have one or more VLANs presented on one or more particular physical ports. This enables the CP to be able to define these and check their operation on the fly. It gives them end to end management of the circuit and the services laid on top.

The difference between managed and unmanaged (or wires only) is that the latter simply present the “circuit” to the customer on one of the Adva ports. The customer must then configure the VLAN termination and any associated routing on their own kit (router) for onward connection to their network.

In reality the costs of providing the actual kit are minimal compared to the overall cost of the circuit and tails etc.

Thanks. That clarifies some thing I didn't know.

From what you say though, these settings:

Gateway IP: 2001:xxx:xxx:2300::1/64

Usable IP's: 2001:xxx:xxx:2300::2 - 2001:xxx:xxx:2300:ffff:ffff:ffff:ffff

Daisy's DNS Servers:

DNS Preferred: 2a04:b2c0:202:3501::35

DNS Alternate: 2001:b98:202:3502::35

Are for the UDMP? In which case this page should have 2300 not 2300, and this one have IPv6 Address also ending in ::1 not the ::2 in the screenshot.

Despite the /56 in this quote:

The IPv6 Range 2001:xxx:xxx:2300::/56 has been routed to port Gi0/1/0 (Vlan100, port currently in use by your IPv4 network) & Gi0/1/1 (Vlan100, port currently free) on the Cisco 1111-8P router

I think the intended meaning is "this specific /56".

Effectively a /64, although ::23ff may well be routed through it. Though we can't be sure 23ff is routed inwards.

I've received a reply from Daisy.
Note, all the x's represent the same numbers/letter as in the initial configuration sent to me (probably obvious to you).

"On 28/05 I added the IPv6 configuration, emailed yourself confirming the /64 assigned to the interface Gi0/1/1 (Vlan100), that we have routed a /56 through our core to site (the /64 is part of this), the gateway IP, the usable IPv6 addresses and our IPv6 DNS servers.

On the 09/06 you advised when you connect a laptop directly into the port and configure one of the available IPv6 addresses etc that you have Internet facing connectivity, so confirmed all is working and this confirms that there’s no issue with the IPv6 configuration/routing applied.

There’s been no contradictory information, we reserve/route a /56 on our core and configure a /64 on your LAN interface (this give you 18,446,744,073,709,551,614 usable IPv6 addresses so should be more than you would ever need). My colleague also advised that we don’t support DHCPv6 and this is down to yourself to manage on your LAN via a device that can act as the IPv6 DHCP server.

Why do you need the LAN changing to a /48, as long as you configure the gateway/subnet correctly etc then there should be no issue with the configuration we have added and provided details on?


If you want to see the config added to your CPE, then please see the below:

int Gi0/0/0
ipv6 address 2001:xxx:xxx:101::35:2/112
!
int Vlan100
ipv6 address 2001:xxx:xxx:2300::1/64
!
ipv6 route 2001:xxx:xxx:2300::/56 Vlan100
!
ipv6 route ::/0 GigabitEthernet0/0/0 2001:xxx:xxx:101::35:1


There’s nothing further we can advise on this, as there’s no issue with the solution we are providing."

Where do I go from here?

Think it would be useful to get an output from the UDM of the IPv6 routes and routing tables. Can you get that from the CLI/shell commands?

Can various devices connected the LAN ping6 each other and the UDM?