06-24-2019 05:31 AM
06-24-2019 06:19 AM
Hello,
watch out: IS-IS does not put interfaces in areas, but has three different types of routers:
IS-IS Level 1 (L1) Router
IS-IS Level 1-2 (L1/L2) Router
IS-IS Level 2 (L2) Router
the Interconnection link between PAN-1 and PAN-2 in IS-IS L1 or OSPF 1 ????
--> interfaces on the left in OSPF area 1
--> both routers need to be L1 routers
the Interconnection link between ABR-1 and ABR-2 in IS-IS L1 Or IS-IS Level 2 ????
--> both routers need to be L1/L2 routers
the Interconnection link between ABR-3 and ABR-4 in IS-IS L1 Or IS-IS Level 2 ????
--> both routers need to be L1/L2 routers
the Interconnection link between PAN-3 and PAN-4 in IS-IS L1 or OSPF 1
--> both routers need to be L1 routers
--> interfaces on the right in OSPF area 1
06-24-2019 07:00 AM - edited 06-24-2019 07:05 AM
Hello Ibrahim,
as explained well by Georg in IS-IS a router can be member of only one IS-IS area but it can operate at IS-IS level 1, IS-IS level2 or IS-IS level 1-2 (the default in IOS).
So the link between PAN1-PAN2 will be in IS-IS level-1. To be noted OSPF can be run over the same link at the same time, but OSPF is used as edge protocol on the left.
The IS-IS backbone is made of a contiguous set of L2 IS-IS links that can be inter-area or intra-area.
The IS-IS backbone cannot be partitioned (Cisco implementation does not support the equivalent of OSPF virtual link called partitioning feature).
So the best thing to do for ABR1 - ABR2 and ABR3-ABR4 links is to run IS-IS level-1-2.
PAN3-PAN4 can be ISIS level 1 and OSPF if desired, interfaces on the right are in OSPF.
Please note, that both IGPs are only used to build first level LDP LSPs and that BGP with labels BGP unicast labelled is used to assign labels to service related IP prefixes.
This is done to achieve the maximum possible scalability in the control plane by adding when required an MPLS label to the stack.
You see in the picture the iBGP sessions with labels.
This is a mobile service provider specific network design.
The unified MPLS stands for that using iBGP with labels over IGP+LDP
the result is that service related prefixes are carried within MPLS using two labels:
the external label is provided by the IGP/LDP used in the specific section
the inner label is provided by BGP with labels.
The label space is only one and BGP with labels cannot allocate a label already used by LDP and also the opposite is true LDP cannot use a label allocated by BGP with labels.
The advantage is to keep the IGP database size to the minimum size possible using them only to provide loopback routing.
The next-hop-self in BGP to the core and to the edge on the ABR-x devices is used to merge/join LSPs and to create and end-to.-end MPLS forwarding path that uses different external LSPs as the BGP next-hop is changed on the path.
This also minimizes the total number of LSPs between loopbacks.
Hope to help
Giuseppe
06-25-2019 01:06 AM
thanks guys
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide