IS-IS is the dynamic link-state routing protocol for the OSI protocol stack. It distributes routing information for routing CLNP data for the ISO CLNS environment.

IS-IS operates similarly to OSPF. IS-IS allows the routing domain to be partitioned into areas. IS-IS routers establish adjacencies using a Hello protocol and exchange link-state information, using link-state packets (LSPs), throughout an area to build the LSDB. Each router then runs Dijkstra’s shortest path first (SPF) algorithm against its LSDB to pick the best paths. There is a minimal amount of information communicated between areas, which reduces the burden on routers supporting the protocol.

IS-IS routing takes place at two levels within an AS: Level 1 (L1) and Level 2 (L2).

L1 routing occurs within an IS-IS area and is responsible for routing to end systems (ESs) and ISs inside an area. All devices in an L1 routing area have the same area address. Routing within an area is accomplished by looking at the locally significant address portion, known as the system ID, and choosing the lowest-cost path.

L2 routing occurs between IS-IS areas. L2 routers learn the locations of L1 routing areas and build an interarea routing table. L2 routers use the destination area address to route traffic using the lowest-cost path.

The three types of IS-IS routers are shown in Figure 6-1.

Figure 6-1. Three Types of IS-IS Routers

The path of connected L2 and L1/L2 routers is called the backbone. All areas and the backbone must be contiguous.

OSI routing begins with ES-IS, when the ESs discover the nearest IS by listening to ISH packets.

When an ES needs to send a packet to another ES, it sends the packet to an IS on an attached network. This process is known as L0 routing.

Each ES and IS resides in a particular area. To pass traffic, the router looks up the destination address and forwards the packet along the best route. If the destination is on the same subnetwork, the IS knows the location (from listening to the ESH) and forwards the packet appropriately. The IS can also provide a redirect message back to the source that tells it that a more direct route is available. If the destination is on a different subnetwork but within the same area, the router identifies the best path, using the system ID, and forwards the traffic appropriately.

If a destination address is in another area, the L1 IS sends the packet to the nearest L1/L2 IS. Packet forwarding continues through L2 and L1/L2 ISs, using the area address, until the packet reaches an L1/L2 IS in the destination area. This process is called L2 routing. Within the destination area, ISs forward the packet along the best path, based on system ID, until the packet reaches the destination.

Routing between separate domains is called L3 routing. L3 routing is comparable to Border Gateway Protocol (BGP) interdomain routing in IP. L3 routing passes traffic between different autonomous systems; these autonomous systems might have different routing logic, so metrics cannot be directly compared. L3 OSI routing is not implemented on Cisco routers but is specified as being accomplished through the Interdomain Routing Protocol (IDRP).

Most debates of the merits of these protocols are colored by their mutual history; different groups with different cultures developed them.

Digital Equipment Corporation (DEC) originally developed IS-IS for DECnet Phase V. In 1987, the American National Standards Institute (ANSI) chose it to be the OSI IGP. At that time it could route only CLNP.

The ISO process is an international standards development process. According to an account given by Christian Huitema in his book Routing in the Internet, groups within ISO and outside the United States did not approve of TCP/IP because of its origin (it was also called the U.S. Department of Defense [DoD] protocol).

From the perspective of ISO, IP development was chaotic and imprecise, based on the famous maxim of “loose consensus and running code.” From the perspective of the early Internet engineers, the ISO process was slow, irritating, and disenfranchising.

In 1988, the U.S. National Science Foundation Network (NSFnet) was created. The IGP used was based on an early draft of IS-IS. The extensions to IS-IS for handling IP were developed in 1988. OSPF development began during this time; it was loosely based on IS-IS.

In 1989, OSPF Version 1 (OSPFv1) was published, and conflict ensued between the proponents of IS-IS and OSPF. The IETF eventually supported both, although it continued to favor OSPF. With the unofficial endorsement of the IETF, OSPF eventually became more popular.

By the mid-1990s, large ISPs in need of an IGP selected IS-IS for two reasons. IS-IS supported IP and CLNS (and therefore solved two problems at once), and OSPF was seen as immature at the time.

IS-IS and OSPF are more similar than dissimilar. Both routing protocols have the following characteristics:

The differences between OSPF and Integrate IS-IS are small, but they do exist; these differences are explored in this section.

Area Design

With OSPF, network design is constrained because OSPF is based on a central backbone, area 0, with all other areas being physically attached to area 0, as shown in Figure 6-6.

The border between OSPF areas is inside the ABRs; each link is in only one area. When you use this type of hierarchical model, you need a consistent IP addressing structure to summarize addresses into the backbone. Summarization also reduces the amount of information carried in the backbone and advertised across the network.

Figure 6-6. OSPF Area Design

In comparison, IS-IS has a hierarchy of L1, L2/L1, and L2 routers, and the area borders lie on links, as shown in Figure 6-7. IS-IS permits a more flexible approach to extending the backbone. The backbone can be extended by simply adding more L2/L1 or L2 routers, a less-complex process than with OSPF.

Figure 6-7. Integrated IS-IS Area Design

Advantages of Integrated IS-IS

OSPF produces many small LSAs, whereas a router groups IS-IS updates and sends them as one LSP. Thus, as network complexity increases, the number of IS-IS updates is not an issue. Each packet must be routed though, and routing takes network resources, so more packets represent a larger impact on the network. Because IS-IS uses significantly fewer LSPs, more routers, at least 1,000, can reside in a single area, making IS-IS more scalable than OSPF.

IS-IS is also more efficient than OSPF in the use of CPU resources and in how it processes routing updates, as shown in Figure 6-8. For one thing, there are fewer LSPs to process (LSAs in OSPF terminology). Also, the mechanism by which IS-IS installs and withdraws prefixes is less resource intensive because it uses network entity title (NET) addresses, which are already summarized. Recall that OSPF runs on top of IP, whereas IS-IS runs through CLNS.

Figure 6-8. Comparing IS-IS and OSPF Routing Updates

Both OSPF and IS-IS are link-state protocols and thus provide fast convergence. The convergence time depends on a number of factors, such as timers, number of nodes, and type of router. Based on the default timers, IS-IS detects a failure faster than OSPF; therefore, convergence occurs more rapidly. If there are many neighboring routers and adjacencies, the convergence time also might depend on the router’s processing power; IS-IS is less CPU intensive than OSPF.

New ideas are not easily expressed in OSPF packets; they require the creation of a new LSA. The OSPF description schema is difficult to extend, because of compatibility issues, and because it was developed exclusively for IPv4. IS-IS is easy to extend through the Type, Length, and Value (TLV) mechanism. TLV strings, called tuples, encode all IS-IS updates. IS-IS can easily grow to cover IPv6 or any other protocol because extending IS-IS consists of simply creating new type codes.

The NSAP address structure is shown in Figure 6-9.

Figure 6-9. Integrated IS-IS NSAP Address Structure

Cisco routers routing CLNS use addressing that conforms to the ISO 10589 standard. ISO NSAP addresses consist of the following:

The NSAP is called the NET when it has an NSEL of 0. Routers use the NET to identify themselves in the IS-IS PDUs.

For example, you might assign 49.0001.0000.0c12.3456.00, which represents the following:

The first part of an NSAP is the area address and is associated with the IS-IS routing process. Unlike OSPF, an IS-IS router can be a member of only one area, as shown in Figure 6-10.

Figure 6-10. IS-IS Routers Are Members of Only One Area

All routers in an area must use the same area address, which actually defines the area. The area address is used in L2 routing.

ESs recognize only ISs and other ESs on the same subnetwork that share the same area address.

The 6-byte NSAP system ID must be unique within an area. It is customary to use a MAC address from the router as the system ID, as shown in Figure 6-11, or, for Integrated IS-IS, to encode an IP address into the system ID.

Figure 6-11. System IDs Are Often the System MAC Address

All the system IDs in a domain must be of equal length. Cisco enforces this OSI directive by fixing the length of the system ID at 6 bytes in all cases.

Level 1 intra-area routing is based on system IDs; therefore, each ES and IS must have a unique system ID within the area.

All Level 2 ISs eventually recognize all other ISs in the Level 2 backbone; therefore, they must also have unique system IDs.

Thus, system IDs should remain unique across the domain. If the system IDs remain unique, there can never be a conflict at L1 or L2 if, for example, a device moves into a different area.

As discussed earlier, NSAP address have a one octet NSEL field that identifies a process on the device, corresponding roughly to a port number in IP. NET addresses are NSAP addresses with an NSEL value of 0. A NET address is used to uniquely identify an OSI host within an IS-IS routing domain. Because IS-IS originates from the OSI world, NET addresses are required even if the only routed protocol is IP.

The NET refers to the device itself; that is, it is the equivalent of that device’s Layer 3 OSI address.

Routers use the NET to identify themselves in the LSPs and, therefore, form the basis for the OSI routing calculation.

Three additional IS-IS terms related to NET addresses are introduced in Figure 6-12: subnetwork point of attachment (SNPA), circuit ID, and link.

Figure 6-12. Subnetwork Point of Attachment, Local Circuit ID, and Link

The SNPA is the point that provides subnetwork services. SNPA is the equivalent of the Layer 2 address corresponding to the NET or NSAP address. The SNPA is assigned by using one of the following:

A circuit is the IS-IS term for an interface. Because the NSAP and NET refer to the entire device, a circuit ID is used to distinguish a particular interface. A router assigns a circuit ID (1 octet) to each of its interfaces as follows:

A link is the path between two neighbor ISs and is defined as being up when communication is possible between the two neighbor SNPAs.

IS-IS routing flows naturally from the OSI address plan in which areas are identified and unique system IDs are given to each device.

The area address portion of the NSAP address can range from 1 to 13 bytes in length, as specified by the ISO standard. Therefore, an NSAP for an IS-IS network can be as little as 8 bytes in length; the NSAP is usually longer to permit some granularity in the allocation of areas. The area address is common to all devices in an area and is unique for each area. ISs and ESs are in the same area if they share the same area address.

Routing within an area involves collecting system IDs and adjacencies for all ISs and ESs in an area and using Dijkstra’s algorithm to compute best paths between devices. L1 routers are only aware of the local area topology. They pass the traffic destined outside the area to the closest L1/L2 router.

Routing between areas is based on area address. L2 (or L1/L2) routers in different areas exchange area address information and use Dijkstra’s algorithm to compute best paths between areas. They pass traffic destined to another area to the best L2 or L1/L2 router to reach that specific area.

When an ES is required to send a packet to another ES, the packet goes to one of the ISs on a network directly attached to the ES. The router then searches for the destination address and forwards the packet along the best route. If the destination ES is in the same area, the local IS recognizes the location by listening to ESH packets and forwards the packet appropriately; routing is by system ID within the area. If the destination address is an ES in another area, the L1 IS sends the packet to the nearest L1/L2 IS.

The L1/L2 IS routes by area address to other L1/L2 or L2 ISs. Forwarding through L1/L2 or L2 ISs, by area address, continues until the packet reaches an L1/L2 or L2 IS in the destination area. Within the destination area, ISs forward the packet along the best path, routing by system ID, until the destination ES is reached.

Because each router makes its own best-path decisions at every hop along the way, there is a significant chance that paths will not be reciprocal. That is, return traffic can take a different path than the outgoing traffic. For this reason, it is important to know the traffic patterns within your network and tune IS-IS for optimal path selection if necessary.

Using Figure 6-13, the following list analyzes traffic from Router 7 (R7) to Router 9 (R9):

Figure 6-13. Example of OSI Addressing

Figure 6-14 illustrates another example, this time with asymmetric routing. Area 1 contains the following two routers:

Figure 6-14. OSI Area Routing

Area 2 has many routers:

Area 3 contains one router that borders areas 2 and 4, yet it has no intra-area neighbors and is currently performing L2 functions only. If you add another router to area 3, you should change it to be an L1/L2 IS.

Recall, as shown in Figure 6-14, the border between the areas in an IS-IS network is on the link between L2 routers. (This is in contrast to OSPF, where the border exists inside the ABR itself.)

In this example, symmetric routing does not occur because L2 details are hidden from L1 routers, which only recognize a default route to the nearest L1/L2 router. For example, traffic from Router X to Router Y flows from Router X to its closest L1/L2 router. The L1/L2 router then forwards the traffic along the shortest path to the destination area (area 2), where it is routed along the shortest intra-area path to Router Y.

Router Y routes return packets to Router X via its nearest L1/L2 router. The L1/L2 router recognizes the best route to area 1 via area 4 based on the lowest-cost L2 path. Because L1 and L2 computations are separate, the path taken from Router Y back to Router X is not necessarily the least-cost path from Router Y to Router X.

Asymmetric routing is not detrimental to the network. However, troubleshooting can be difficult, and this type of routing is sometimes a symptom of suboptimal design. Like EIGRP and OSPF, a good IS-IS design is generally hierarchical and symmetrical.

Route leaking is a feature available since Cisco IOS Software Release 12.0 that helps avoid asymmetric routing and reduce suboptimal routing by providing a mechanism for leaking, or redistributing, L2 routes into L1 routers in a controlled manner. By having more detail about interarea routes, an L1 router is able to make a better choice with regard to which L1/L2 router to forward the packet.

Route leaking is defined in RFC 2966, Domain-wide Prefix Distribution with Two-Level IS-IS, for use with the narrow metric TLV types 128 and 130. The IETF has also defined route leaking for use with the wide metric (using TLV type 135).

To implement route leaking, an up/down bit in the TLV is used to indicate whether or not the route identified in the TLV has been leaked. If the up/down bit is set to 0 the route was originated within that L1 area. If the up/down bit is set to 1 the route has been redistributed into the area from L2. The up/down bit is used to prevent routing loops: An L1/L2 router does not re-advertise into L2 any L1 routes that have the up/down bit set.

Route leaking should be planned and deployed carefully to avoid the situation where any topology change in one area results in having to recompute many routes in all other areas.

The OSI stack defines a unit of data as a PDU. OSI recognizes a frame as a data-link PDU and a packet (or datagram, in the IP environment) as a network PDU.

Figure 6-15 shows examples of three types of PDUs (all with IEEE 802.2 Logical Link Control [LLC] encapsulation). IS-IS and ES-IS PDUs are encapsulated directly in a data-link PDU (frame); there is no CLNP header and no IP header. (In other words, IS-IS and ES-IS do not put routing information in IP or CLNP packets; rather, they put routing information directly in a data-link layer frame.) True CLNP (data) packets contain a full CLNP header between the data-link header and any higher-layer CLNS information.

Figure 6-15. OSI PDUs

The IS-IS and ES-IS PDUs contain variable-length fields, depending on the function of the PDU. Each field contains a type code, a length, and the appropriate values; this information is known as the TLVs.

IS-IS defines the following four types of PDUs:

The Hello PDUs were described earlier in section “The ES-IS Protocol”; IIHs use is further explored later in this chapter in the “Implementing IS-IS in Different Network Types” section. LSPs are described in the next section. PSNPs and CSNPs use is described in the later “Link-State Database Synchronization” section.

In IS-IS, router characteristics are defined by an LSP. A router’s LSP contains an LSP header followed by TLV fields, as shown in Figure 6-16.

Figure 6-16. Link-State Packets Represent Routers

An LSP header describes the following:

Cisco IOS Software automatically uses broadcast mode for LAN links and multipoint WAN links. It uses point-to-point mode for point-to-point links, such as point-to-point subinterfaces and dialer interfaces.

IS-IS has no specific support for NBMA networks. When implemented in broadcast mode, Cisco IOS Software assumes that the NBMA environment features a full mesh of PVCs. When creating static maps to map the remote IP address to the local DLCI on a Frame Relay interface, you should use the broadcast keyword; this is because broadcast mode uses multicast updates, which will not be sent without this keyword set.

When you use multipoint WAN links such as multipoint Frame Relay interfaces, you must also allow CLNS broadcasts and multicasts. This can be done by using the frame-relay map clns dlci-number broadcast command (in addition to creating the IP maps).

In IS-IS, broadcast networks are LAN interfaces or multipoint WAN interfaces.

Separate IS-IS adjacencies are established for L1 and L2 processes. If two neighboring routers in the same area run both L1 and L2, they establish two adjacencies, one for each level. The router stores the L1 and L2 adjacencies in separate L1 and L2 adjacency tables. On LANs, routers establish the two adjacencies with specific Layer 1 and Layer 2 IIH PDUs.

IIH PDUs announce the area address; separate IIH packets announce the L1 and L2 neighbors. Adjacencies form based on the area address communicated in the incoming IIH and the type of router (L1 or L2). L1 routers accept L1 IIH PDUs from their own area and establish adjacencies with other routers in their own area. L2 routers (or the L2 process within any L1/L2 router) accept only L2 IIH PDUs and establish only Level 2 adjacencies.

Pseudonode and DIS

In IS-IS, a broadcast link itself is modeled as a pseudo-node that connects all attached routers to a star-shaped topology. The pseudo-node is represented by the DIS. Dijkstra’s algorithm requires this virtual router (pseudo-node) to build a directed graph for broadcast media.

Criteria for DIS selection are, first, highest priority and second, highest SNPA (recall that on LANs the SNPA is the MAC address). Cisco router interfaces have a default L1 and L2 priority of 64. You can configure the priority from 0 to 127 using the isis priority number-value[level-1 | level-2] interface configuration command. The L1 DIS and L2 DIS on a LAN may or may not be the same router, because an interface can have different L1 and L2 priorities.

In IS-IS, all routers on a LAN establish adjacencies with all other routers and with the DIS. Therefore, if the DIS fails, another router takes over immediately with little or no impact on the network’s topology. There is no backup DIS. Contrast this behavior with OSPF, where the DR and BDR are selected and the other routers on the LAN establish full adjacencies only with the DR and BDR. In case of DR failure, the BDR is promoted to DR, and a new BDR is elected.

A selected router is not guaranteed to remain the DIS. Any adjacent IS with a higher priority automatically takes over the DIS role; this is called preemptive behavior. Because the IS-IS LSDB is synchronized frequently on a LAN, giving priority to another IS over the DIS is not a significant issue.

Rather than having each router connected to the LAN advertise an adjacency with every other router on the LAN, each router (including the DIS) just advertises a single adjacency to the pseudo-node. Otherwise, each IS on a broadcast network with n connected ISs would require (n)(n – 1)/2 adjacency advertisements. Generating LSPs for each adjacency creates considerable overhead in terms of LSDB synchronization.

As shown in Figure 6-17, the DIS generates the pseudo-node LSPs. A pseudo-node LSP details only the adjacent ISs (for example, the ISs connected to that LAN). The pseudo-node LSP is used to build the map of the network and to calculate the SPF tree. The pseudo-node LSP is the equivalent of a network LSA in OSPF.

Figure 6-17. LSP Representing Routers: LAN Representation

This section describes the LSPs and IIHs used by L1 and L2 routers.

Table 6-4 summarizes the differences between broadcast and point-to-point links.

An IS-IS update process is responsible for flooding the LSPs throughout the IS-IS domain. An LSP is typically flooded to all adjacent neighbors except the neighbor from which it was received. L1 LSPs are flooded within their local areas. L2 LSPs are flooded throughout the backbone.

Each IS originates its own LSPs (one for L1 and one for L2). These LSPs are identified by the originator’s system ID and an LSP fragment number starting at 0. If an LSP exceeds the maximum transmission unit (MTU), it is fragmented into several LSPs, numbered 1, 2, 3, and so on.

IS-IS maintains the L1 and L2 LSPs in separate LSDBs.

When an IS receives an LSP, it examines the checksum and discards any invalid LSPs, flooding them with an expired lifetime age. If the LSP is valid and newer than what is currently in the LSDB, it is retained, acknowledged, and given a lifetime of 1200 seconds. The age is decremented every second until it reaches 0, at which point the LSP is considered to have expired. If an LSP expires, it is kept for an additional 60 seconds before it is flooded as an expired LSP.

Sequence number PDUs (SNPs) are used to acknowledge the receipt of LSPs and to maintain LSDB synchronization. There are two types of SNPs: CSNP and PSNP. The use of SNPs differs between point-to-point and broadcast media.

Separate CSNPs and PSNPs are used for L1 and L2 adjacencies.

Adjacent IS-IS routers exchange CSNPs to compare their LSDB. In broadcast networks, only the DIS transmits CSNPs. All adjacent neighbors compare the LSP summaries received in the CSNP with the contents of their local LSDBs to determine whether their LSDBs are synchronized (in other words, whether they have the same copies of LSPs as other routers for the appropriate levels and area of routing). CSNPs are periodically multicast (every 10 seconds) by the DIS on a LAN to ensure LSDB accuracy.

If there are too many LSPs to include in one CSNP, they are sent in ranges. The CSNP header indicates the starting and ending LSP ID in the range. If all LSPs fit the CSNP, the range is set to default values.

Adjacent IS-IS routers use PSNPs to acknowledge the receipt of LSPs and to request transmission of missing or newer LSPs.

On point-to-point networks, CSNPs are sent when the link comes up to synchronize the LSDBs.

LSDB Synchronization on a LAN Example

On a LAN the DIS periodically (every 10 seconds) sends CSNPs that list the LSPs it holds in its LSDB. This update is a multicast to all L1 or L2 IS-IS routers on the LAN.

Figure 6-18 illustrates an example. In this network, R1 compares the list of LSPs from the DIS with its topology table and realizes it is missing one LSP. Therefore, it sends a PSNP to the DIS (R2) to request the missing LSP.

Figure 6-18. LSDB Synchronization on a LAN

The DIS reissues only that missing LSP (LSP 77), and R1 acknowledges it with a PSNP.

LSDB Synchronization on Point-to-Point Links Example

Unlike on broadcast links, such as LAN links, CSNPs are not periodically sent on point-to-point links. A CSNP is sent only once, when the point-to-point link first comes up. After that, LSPs are sent to describe topology changes, and they are acknowledged with a PSNP.

Figure 6-19 shows an example of what happens on a point-to-point link when a link failure is detected.

Figure 6-19. LSDB Synchronization: Point to Point

The steps shown in this example are as follows:

1. A link fails.

2. R2 notices this failure and issues a new LSP noting the change.

3. R1 receives the LSP, stores it in its topology table, and sends a PSNP back to R2 to acknowledge receipt of the LSP.

IIH PDUs announce the area address. On LANs, separate IIH packets announce the L1 and L2 neighbors.

For example, when a LAN has routers from two areas attached, as shown in Figure 6-20, the following process takes place:

Figure 6-20. L1 and L2 Adjacencies on a LAN

On point-to-point WAN links, the IIH PDUs are a common format to both levels. The level type and the area address are announced in the hellos, as follows:

Two L1 routers that are physically connected but are not in the same area can exchange IIHs, but they do not establish an adjacency, because the area addresses do not match.

Figure 6-21 shows the different permutations for WAN adjacencies.

Figure 6-21. WAN Adjacencies

IS-IS uses an OSI forwarding database (routing table) to select the best path to a destination.

When the LSDBs are synchronized, routers use the LSDBs to calculate the SPF tree to OSI destinations (NETs). The total of the link metrics along each path determines the shortest path to any given destination. L1 and L2 routes have separate LSDBs; therefore, routers may run the SPF algorithm twice, once for each level, and create separate SPF trees for each level. Routers insert the best paths into the CLNS routing table (also called the OSI forwarding database).

Routers calculate ES reachability with a partial route calculation (PRC), based on the L1 and L2 SPF trees.

The PRC generates best-path choices for IP routes and offers the routes to the IP routing table, where they are accepted based on normal IP routing table rules. For example, if more than one routing protocol is running, the router compares the administrative distances of routes to the same destination. When the IP IS-IS routes are entered into the routing table, they are shown as via L1 or L2, as appropriate.

The separation of IP reachability from the core IS-IS network architecture provides Integrated IS-IS better scalability than OSPF, as follows:

Recall that all interarea traffic in IS-IS must traverse the L2 backbone area. Thus, CLNS addresses must be planned to execute a two-level hierarchy. You must decide which routers will be backbone (L2) routers, which routers will be L1/L2, and which will be internal area (L1) routers. If some routers must do both L1 and L2 routing, the specific interfaces that will participate in each type of routing should be identified.

Remember that a router’s CLNS address is called the NET, and it consists of three main parts:

It is not enough to plan the IS-IS area addressing. You must also plan IP addressing to allow for summarization of addresses so that the network is scalable. Route summarization is the key that enables all the benefits of hierarchical addressing design. Route summarization minimizes routing update traffic and resource utilization.

Be particularly careful when you configure the IP addressing on the router, because it is more difficult to troubleshoot IP address misconfigurations with IS-IS. The IS-IS neighbor relationships are established over OSI CLNS, not over IP. Because of this approach, two ends of a CLNS adjacency can have IP addresses on different subnets, with no impact on the operation of IS-IS.

The router isis [area-tag] global configuration command enables Integrated IS-IS on the router. You can use the optional tag to identify multiple IS-IS processes. (Just as multiple OSPF processes can be present on the same router, multiple IS-IS processes are possible.) The process name is significant only to the local router. If it is omitted, the Cisco IOS Software assumes a tag of 0. If more than one IS-IS process is used, the network plan should indicate which interfaces will participate in which IS-IS process.

IP routing is enabled by default; CLNS routing is disabled by default. To enable CLNS routing, use the clns routing global configuration command. In addition, you must enable CLNS routing at each interface.

After the IS-IS process is enabled, the router’s NET must be assigned, with the net network-entity-title router configuration command.

Even when you use IS-IS for IP routing only (with CLNS routing not enabled), you must still configure a NET. The NET is a combination of area number and a unique system identification number for each particular router, plus an NSEL of 00 at the end. The area number must be between 1 and 13 bytes in length. The system ID has a fixed length of 6 bytes in Cisco routers and must be unique throughout each area (L1) and throughout the backbone (L2).

The final step is to select which interfaces participate in IS-IS routing. Interfaces that use IS-IS to route IP (and thus must establish IS-IS adjacencies) must be configured using the ip router isis [area-tag] interface configuration command. Enable Integrated IS-IS on the appropriate interfaces—do not forget interfaces to stub IP networks, such as loopback interfaces (even though there are no CLNS neighbors on those interfaces).

If there is more than one IS-IS process, the IS-IS process to which the interface belongs must be specified using the appropriate process name in the optional area-tag field. If no area-tag is listed, the Cisco IOS Software assumes a tag value of 0. If only one IS-IS process is active on the router, no area-tag value is needed.

Use the clns router isis [area-tag] interface configuration command to enable the IS-IS routing process on an interface to support CLNS routing.

Example 6-1 shows a simple Integrated IS-IS configuration for IP routing only; CLNS routing is not enabled. This configuration specifies only one IS-IS process, so the optional area-tag is not used. The net command configures the router to be in area 49.0001 and assigns a system ID of 0000.0000.0002. IS-IS is enabled on the Fast Ethernet 0/0 and serial 0/0/1 interfaces. This router acts as an L1/L2 router by default.

interface FastEthernet0/0
 ip address 10.1.1.2 255.255.255.0
 ip router isis
!
interface Serial 0/0/1
 ip address 10.2.2.2 255.255.255.0
 ip router isis
!

<output omitted>

router isis
net 49.0001.0000.0000.0002.00

The default configuration of IS-IS leaves the router with an IS type of L1/L2. Although this configuration has the advantage of allowing all routers to learn of each other and pass routes without too much administrative oversight, it is not the most efficient way to build an IS-IS network. Routers with the default configuration send out both L1 and L2 hellos and maintain both L1 and L2 LSDBs.

Each router should be configured to support the minimum level of routing required, which does the following:

If a router is to operate only as an internal area router or a backbone router, specify this configuration by entering the is-type {level-1 | level-1-2 | level-2-only} router configuration command. To specify that the router act only as an internal area (L1) router, use is-type level-1. To specify that the router act only as a backbone (L2) router, use is-type level-2-only. If the level type has been changed from the default, you can return to the default with the is-type level-1-2 command.

Although a router can be an L1/L2 router, it might not be required to establish both types of adjacencies over all interfaces.

If a particular interface only has L1 routers connected to it, there is no need for the router to send L2 hellos out that interface. Similarly, if an interface only has L2 routers connected to it, there is no need for the router to send L1 hellos out that interface. Extraneous hellos would waste bandwidth and router resources trying to establish adjacencies that do not exist.

To make IS-IS more efficient in these types of situations, configure the interface to send only the needed type of hellos by using the isis circuit-type {level-1 | level-1-2 | level-2-only} interface configuration command. The default is level-1-2; the router will attempt to establish both types of adjacencies over the interface.

Unlike most other IP protocols, IS-IS on a Cisco router does not take into account line speed or bandwidth when it sets its link metrics; all interfaces are assigned a metric value of 10 by default. In a network with links of varying types and speeds, this assignment can result in suboptimal routing.

To change the metric value, use the isis metric metric [delay-metric [expense-metric [error-metric]]] {level-1 | level-2} interface configuration command. The metric can have different values for L1 and L2 over the same interface. The metric value is from 1 to 63.

If the metric value for all interfaces needs to be changed from the default value of 10, then the change needs to be performed one by one on all IS-IS interfaces, which can be time-consuming and error-prone, especially for routers with many IS-IS interfaces. Alternately, the metric default-value {level-1 | level-2} router configuration command can be used to change the metric value for all IS-IS interfaces. If the keyword level-1 or level-2 is not entered, the metric will be applied to both L1 and L2 IS-IS interfaces. This command is available only in Cisco IOS Software Release 12.3(4)T and later; it only supports the cost metric.

Figure 6-22 shows two different areas. The configurations of the routers are provided in Example 6-2.

Figure 6-22. Tuning IS-IS Example

R1(config)#router isis
R1(config-router)#net 49.0001.0000.0000.0001.00
R1(config-router)#is-type level-1

R1(config)#interface FastEthernet0/0
R1(config-if)#ip router isis
R1(config-if)#isis circuit-type level-1
R2(config)#router isis
R2(config-router)#net 49.0001.0000.0000.0002.00

R2(config)#interface FastEthernet0/0
R2(config-if)#ip router isis
R2(config-if)#isis circuit-type level-1

R2(config)#interface serial 0/0/1
R2(config-if)#ip router isis
R2(config-if)#isis circuit-type level-2-only
R2(config-if)#isis metric 35 level-2
R3(config)#router isis
R3(config-router)#net 49.0002.0000.0000.0003.00
R3(config-router)#is-type level-2-only

R3(config)#interface Serial0/0/1
R3(config-if)#ip router isis
R3(config-if)#isis circuit-type level-2-only
R3(config-if)#isis metric 35 level-2

In this example, area 49.0002 contains only one router (R3) and needs to do only L2 routing. Therefore, it is appropriate to change R3’s IS type to level-2-only.

Area 49.0001 has two routers. R1 is strictly an internal area router; it does not connect to routers in any other area. It is appropriate to configure this router as IS type level-1.

R2 connects to the internal area router R1 and also to R3, in a different area. Therefore, R2 must do both L1 and L2 routing, so it is left at the default IS L1/L2 type setting. However, there is no need for R2 to send L2 hellos out the interface connected to R1, so it is appropriate to set the IS-IS circuit type of R2’s FastEthernet 0/0 to L1. Similarly, because R2’s Serial 0/0/1 interface connects only to an L2 router, the IS-IS circuit type should be set to level-2-only.

Remember that the default IS-IS metric for all interfaces is 10. In this topology, the serial link is slower than the Fast Ethernet link. Using the default metric does not give the routers a true picture of the value of each link, so the routers cannot make truly informed routing decisions. Therefore, the IS-IS metric on each serial interface should be changed to reflect the link preference; in Example 6-2, the metric is set to 35.

You can use the following show clns commands to verify the router configuration and to troubleshoot the CLNS portion of an Integrated IS-IS network:

You can use the following show commands to verify the router configuration and to troubleshoot the CLNS and IS-IS portions of the Integrated IS-IS network:

Figure 6-23 is a copy of Figure 6-22, repeated here for your convenience. The figure shows three routers in two areas; routers R1 and R2 belong to area 49.0001 and router R3 belongs to area 49.0002. R1 is an L1 router doing only L1 routing. R2 is an L1/L2 router doing both L1 and L2 routing. R3 is an L2 router doing only L2 routing. The configuration of the three routers is shown earlier, in Example 6-2.

Figure 6-23. Routing in a Two-Level Area Structure

Figure 6-23 forms the basis of the following show command output examples.

R1#show isis topology
IS-IS paths to level-1 routers
System Id            Metric     Next-Hop  Interface  SNPA
R1                   —
R2                   10         R2        Fa0/0      0016.4650.c470
R2#show isis topology
IS-IS paths to level-1 routers
System Id            Metric     Next-Hop  Interface  SNPA
R1                   10         R1        Fa0/0      0016.4610.fdb0
R2                   —
IS-IS paths to level-2 routers
System Id            Metric     Next-Hop  Interface  SNPA
R1                   **
R2                   —
R3                   35         R3        Se0/0/1    *HDLC*

R2#show clns protocol
IS-IS Router: <Null Tag>
  System Id: 0000.0000.0002.00  IS-Type: level-1-2
  Manual area address(es):
        49.0001
  Routing for area address(es):
        49.0001
  Interfaces supported by IS-IS:
        Loopback0 - IP
        Serial0/0/1 - IP
        FastEthernet0/0 - IP
  Redistribute:
    static (on by default)
  Distance for L2 CLNS routes: 110
  RRR level: none
  Generate narrow metrics: level-1-2
  Accept narrow metrics:   level-1-2
  Generate wide metrics:   none
Accept wide metrics:     none

R2#show clns neighbors
System Id      Interface   SNPA             State  Holdtime  Type Protocol
R3             Se0/0/1     *HDLC*           Up     28        L2   IS-IS
R1             Fa0/0       0016.4610.fdb0   Up     23        L1   IS-IS

R2#show clns interface s0/0/1
Serial0/0/1 is up, line protocol is up
  Checksums enabled, MTU 1500, Encapsulation HDLC
  ERPDUs enabled, min. interval 10 msec.
  CLNS fast switching enabled
  CLNS SSE switching disabled
  DEC compatibility mode OFF for this interface
  Next ESH/ISH in 45 seconds
  Routing Protocol: IS-IS
    Circuit Type: level-2
    Interface number 0x1, local circuit ID 0x100
    Neighbor System-ID: R3
    Level-2 Metric: 35, Priority: 64, Circuit ID: R2.00
    Level-2 IPv6 Metric: 10
    Number of active level-2 adjacencies: 1
    Next IS-IS Hello in 5 seconds
if state UP