Configure IS-IS Routing for IPv4

We’ll start with the simplest possible scenario: we’re adding a new router to an existing IS-IS network and have to configure IS-IS routing on it.

Lab topology

Device Requirements

Starting the Lab

Assuming you already set up your lab infrastructure:

  • Change directory to basic/1-simple-ipv4
  • Execute netlab up
  • Log into your device (RTR) with netlab connect rtr and verify that the IP addresses are configured on all its interfaces.

You’ll get a lab with IPv4 addresses configured on all interfaces and IS-IS routing configured on X1 and X2 but not RTR.

Configuration Tasks

The minimal IS-IS configuration process has three or four steps:

  • Start an IS-IS routing process with a global configuration command similar to router isis. You might have to specify a routing process tag – a string that identifies the IS-IS instance in other configuration commands.
  • Assign a CLNS address (NET – Network Entity Title) to the IS-IS routing process with a routing configuration command similar to net address (more details). The NET assigned to your router is 49.0001.0000.0000.0042.00
  • Enable IS-IS routing for IPv4 on individual interfaces with an interface configuration command similar to ip router isis tag or isis enable tag.
  • Some devices also want you to configure which address families (IPv4 or IPv6) you want to route with IS-IS. The router configuration command to do that is often address-family ipv4.

Tip

If you’re using FRRouting on RTR, you might have to enable the IS-IS daemon first. You’ll find more details in the Configuring IS-IS on FRRouting lab.

Unfortunately, most IS-IS implementations are ancient and use defaults that might have made sense decades ago. You should also configure the following parameters1:

  • Configure the links with two nodes attached to them as point-to-point links with an interface configuration command similar to isis network point-to-point (more details in IS-IS on Point-to-Point Links)
  • Configure your router to be a Level-2-Only router with the router configuration command similar to is-type level-2-only2 (more details in Optimize Simple IS-IS Deployments).
  • On some devices, you’ll have to configure wide metrics with a router configuration command similar to metric-style wide4 (more details in Using IS-IS Metrics).
  • Finally, you might have to configure the standard point-to-point three-way handshake on Cisco IOSv3. The interface configuration command to do that is ‌isis three-way-handshake ietf.

Finally, make sure your configuration includes these settings:

  • Logging of adjacency changes (typically configured with a router configuration command like log-adjacency-changes)
  • Dynamic mapping of NETs to hostnames (usually enabled; sometimes you can’t turn it off)

Validation

You can use the netlab validate command if you’re using netlab release 1.8.4 or later and run FRRouting on the external routers. This is the printout you should get after completing the lab exercise:

You should also do basic manual validation on your device:

  • Inspect IS-IS adjacencies with a command similar to show isis neighbors or show isis adjacency. Your router should have level-2 adjacencies with X1 and X2; this is the printout you should get on FRRouting:

IS-IS neighbors (FRRouting)

rtr# show isis neighbor
Area SomeTag:
  System Id           Interface   L  State        Holdtime SNPA
 x1                  eth1        2  Up            29       2020.2020.2020
 x2                  eth2        2  Up            30       2020.2020.2020
  • If your device has commands similar to show isis topology, show isis network topology, or show isis route, use them to display the shortest paths to other routers and the best IS-IS paths.

IS-IS network topology (first part of the printout) and IPv4 routing table on FRRouting

rtr# show isis route
Area SomeTag:
IS-IS paths to level-2 routers that speak IP
Vertex               Type         Metric Next-Hop             Interface Parent
rtr
10.1.0.0/30          IP internal  0                                     rtr(4)
10.1.0.4/30          IP internal  0                                     rtr(4)
x1                   TE-IS        10     x1                   eth1      rtr(4)
x2                   TE-IS        10     x2                   eth2      rtr(4)
10.0.0.2/32          IP TE        20     x1                   eth1      x1(4)
10.1.0.0/30          IP TE        20     x1                   eth1      x1(4)
10.0.0.3/32          IP TE        20     x2                   eth2      x2(4)
10.1.0.4/30          IP TE        20     x2                   eth2      x2(4)

IS-IS L2 IPv4 routing table:

 Prefix       Metric  Interface  Nexthop   Label(s)
 ----------------------------------------------------
 10.0.0.2/32  20      eth1       10.1.0.2  -
 10.0.0.3/32  20      eth2       10.1.0.6  -
 10.1.0.0/30  20      eth1       10.1.0.2  -
 10.1.0.4/30  20      eth2       10.1.0.6  -

IS-IS network topology on Arista EOS

rtr>show isis network topology

IS-IS Instance: SomeTag VRF: default
  IS-IS paths to level-2 routers
    System Id        Metric   IA Metric Next-Hop         Interface                SNPA
    x1               10       0         x1               Ethernet1                P2P
    x2               10       0         x2               Ethernet2                P2P
  • Finally, inspect the IP routing table and verify that you can see three loopback prefixes (10.0.0.1/32 through 10.0.0.3/32). This is the printout you should get on FRRouting

IS-IS routes in the IPv4 routing table (FRRouting)

rtr# show ip route isis
Codes: K - kernel route, C - connected, L - local, S - static,
       R - RIP, O - OSPF, I - IS-IS, B - BGP, E - EIGRP, N - NHRP,
       T - Table, v - VNC, V - VNC-Direct, A - Babel, F - PBR,
       f - OpenFabric, t - Table-Direct,
       > - selected route, * - FIB route, q - queued, r - rejected, b - backup
       t - trapped, o - offload failure

I>* 10.0.0.2/32 [115/20] via 10.1.0.2, eth1, weight 1, 00:00:47
I>* 10.0.0.3/32 [115/20] via 10.1.0.6, eth2, weight 1, 00:00:18
I   10.1.0.0/30 [115/20] via 10.1.0.2, eth1 inactive, weight 1, 00:00:47
I   10.1.0.4/30 [115/20] via 10.1.0.6, eth2 inactive, weight 1, 00:00:18

FRRouting uses the IP routing table as the repository of all routes, not just the best routes, so it displays the inactive IS-IS routes. Other network devices, for example Arista EOS5, display just the active IS-IS routes:

IS-IS routes in IPv4 routing table on Arista EOS

rtr#show ip route isis detail|begin Priority
Priority Codes:
       PL - Priority low, PM - Priority medium, PH - Priority high

 I L2     10.0.0.2/32 [115/20]
           via 10.1.0.2, Ethernet1 rtr -> x1
 I L2     10.0.0.3/32 [115/20]
           via 10.1.0.6, Ethernet2 rtr -> x2

Next: Explore IS-IS Data Structures. That lab uses the same topology and device configurations, so you can keep going without restarting the lab.

Background Information

IS-IS was designed to be the routing protocol for the ISO/OSI protocol stack. Its name is derived from ISO terminology: IS stands for Intermediate System (the device also known as router, gateway or layer-3 switch). Thus, IS-IS is the protocol that runs between Intermediate Systems (in other words, the router-to-router protocol).

While ISO designed IS-IS to be used with CLNP, its data structures (everything is CLV6-based) made it easy to extend, and so IETF quickly specified how to use IS-IS for IPv4 and (later) IPv6 routing.

IS-IS does not run on top of another layer-3 protocol like OSPF or BGP run on top of IP; it is an independent protocol at layer-3 and uses a dedicated 802.1 SAP (more details). You cannot use the standard IP access control mechanisms with IS-IS; on the other hand, it’s impossible to spoof IS-IS messages if the intruder does not have access to the underlying layer-2 infrastructure.

IS-IS addressing follows the OSI addressing principles. Network addresses (Network Service Access Points – NSAPs) are assigned to nodes, not interfaces. OSI protocol suite used variable-length addresses, and according to the (poorly formatted copy of) IS-IS specifications, the NET is split into three parts (you’ll find more details in RFC 1237):

+--------+-----------+----+
|   Area | System-ID | NS |
+--------+-----------+----+
  • The NSAP Selector (the last byte) specifies the software entity on the node that should receive the packet (similar to the TCP/UDP/OSPF protocol type within the IPv4 header). IS-IS uses NSAP selector value zero and calls the resulting address Network Entity Title (NET). Thus, the last byte of NET always has to be zero.
  • The System ID (node address within an area, similar to the subnet part of an IP address) is one to eight bytes long. Subsequent parts of IS-IS specification assume a 6-byte System ID (12 hexadecimal characters), and that’s what all implementations use.
  • The Area is whatever is left on the left-hand side of the NET after you take away the rightmost seven bytes.

The NET is usually formatted as a hexadecimal string (two hexadecimal digits per byte) with two-byte portions (four hexadecimal digits) separated by a dot. The placement of dots is a bit counterintuitive:

  • The rightmost dot has to be between the System ID and NSAP Selector.
  • Subsequent dots have to be four characters left of the previous dot, resulting in two dots within the System ID and a dot between the node System ID and the IS-IS area.

For example, the NET we used in this lab (49000100000000004200)has to be written as 49.0001.0000.0000.0042.00.

The area address could have any length7 as long as it’s specified as a whole number of bytes (an even number of digits). IS-IS doesn’t care about the contents of the area address. If two routers have the value in that part of the NET, they belong to the same area; if not, they don’t.

Finally, other ISO standards define the meaning of the leftmost part of a network address (NSAP). The first (leftmost) byte is the Authority and Format Identifier (AFI), and the value 49 indicates private address space (similar to RFC 1918 IPv4 address space) (see RFC 1629 for more details). Using two bytes as the area ID usually makes sense, and that’s why most NETs you find in online examples start with 49.0001 ;)8

Reference Information

Lab Wiring

Origin Device Origin Port Destination Device Destination Port
rtr Ethernet1 x1 eth1
rtr Ethernet2 x2 eth1

Note: The interface names depend on the devices you use in the lab. The printout was generated with Arista EOS running on RTR and FRRouting running on X1 and X2.

Lab Addressing

Node/Interface IPv4 Address IPv6 Address Description
rtr 10.0.0.1/32 Loopback
Ethernet1 10.1.0.1/30 rtr -> x1
Ethernet2 10.1.0.5/30 rtr -> x2
x1 10.0.0.2/32 Loopback
eth1 10.1.0.2/30 x1 -> rtr
x2 10.0.0.3/32 Loopback
eth1 10.1.0.6/30 x2 -> rtr

  1. Trust me for the moment; we’ll discuss them in great detail in subsequent lab exercises. 

  2. You don’t have to configure the interface circuit-type once you configured the device to be a level-2-only router. A level-2-only router will never send level-1 hellos. 

  3. The IOSv image on which I had to do that was obviously ancient but still 13 years younger than the RFC 3373 that standardized the three-way handshake. 

  4. As crazy as it sounds, some devices still use the ancient narrow metrics as the default setting 20 years after the wide metrics were defined in RFC 3784

  5. The detail option of the show ip route command includes the interface description, and the |begin Priority output filter removes the legend. 

  6. Code/Length/Value: every bit of information starts with a code (one byte), followed by length (one byte) and the actual value. This encoding approach is more commonly called TLV (Type/Length/Value). 

  7. There must be some upper limit. 

  8. In the early days of IS-IS, when people still believed that the OSI protocol stack (with CLNS at layer-3) might win the protocol wars, service providers used official NSAP prefixes allocated by government agencies (see RFC 1237 for details). CLNS is long gone, and we only need NETs to address IS-IS routers. Area 49.0001 is good enough for that. You can also use Area 51 if you’re so inclined (yes, it works; IS-IS does not care about the contents of the area field).