Nokia NRS II Composite (IS-IS) key terms

A router that runs IS-IS and participates in link-state routing. It originates link-state PDUs describing its own links and adjacencies, and computes shortest paths from the topology it learns.

The OSI address of an IS-IS-enabled system, composed of an authority and format identifier, an area address, a system identifier, and a selector byte identifying the service.

An NSAP whose selector byte is zero, used to identify the router itself rather than an upper-layer service. The NET supplies both the area address and the system ID used by IS-IS.

A fixed-length identifier, typically six bytes, that uniquely identifies an IS within its area. It is the last part of the NET before the selector byte and is the node identity used in SPF calculations.

The variable-length portion of a NET shared by all routers in the same IS-IS area. Routers with at least one matching area address can form Level 1 adjacencies.

An IS that maintains adjacencies and link-state data only with routers in its own area. It reaches other areas solely through an attached Level 1/Level 2 router using a default route.

An IS that maintains a backbone-level database and forms Level 2 adjacencies with other Level 2-capable routers, carrying inter-area prefix information across the routing domain.

A router that runs both levels simultaneously, holding separate databases for its area and for the backbone. It advertises its area's prefixes into Level 2 and sets the attached bit toward Level 1 neighbors.

A flag in the Level 1 link-state PDU of a Level 1/Level 2 router telling intra-area routers that a path out of the area exists. Level 1 routers use it to choose which default next hop to install.

The continuous set of Level 2 adjacencies across the routing domain. It must remain connected so that inter-area reachability information can flow between areas.

A failure condition in which an area's Level 1 topology splits into disconnected pieces while the backbone stays intact. Level 1/Level 2 routers on both sides can still relay traffic through Level 2, but intra-area optimal paths are lost.

A periodic link-local PDU used to discover neighbors, negotiate level and protocol parameters, and maintain adjacency. Its hold timer, when it expires without a refresh, tears the adjacency down.

A countdown carried in Hello PDUs that is refreshed on each received Hello. If it reaches zero, the neighbor is declared down and the adjacency is removed.

A relationship between two IS-IS routers on a common link, established after Hello parameters are exchanged and accepted. Its state progresses from initial to up, at which point link-state information can be exchanged.

An adjacency state in which a router has received a Hello from a neighbor that lists its own system ID as a neighbor. On broadcast links this must be reached before the DIS includes the router in the pseudonode.

A Hello field declaring whether the interface runs Level 1, Level 2, or both. Mismatched circuit types prevent adjacency formation, which is a common cause of silent neighbor failures.

On a broadcast link, the router elected to represent the segment as a pseudonode. It originates the pseudonode LSP, floods on behalf of the segment, and is chosen by highest priority then highest system ID.

A link-state PDU originated by the DIS describing a broadcast segment as a virtual node, with the attached routers listed as neighbors. It makes the multi-access link appear as a set of point-to-point links in the topology.

A multi-access medium such as Ethernet on which many routers share one segment. IS-IS models it with a DIS and pseudonode rather than forming a full mesh of adjacencies.

A relationship on a link with exactly two routers, such as a serial or point-to-point Ethernet interface. No DIS or pseudonode is used, and both routers describe the link directly in their own LSPs.

The unit of topology and reachability information a router originates about itself. Each LSP carries a sequence number, remaining lifetime, and TLVs, and is flooded to all routers at its level.

The fixed portion of a link-state PDU containing the LSP ID, sequence number, remaining lifetime, checksum, and flags such as the attached, overload, and partition-repair bits. These fields drive flooding and aging decisions.

An identifier made of the originating system ID, a pseudonode number, and a fragment number. It uniquely names a link-state PDU within a level's database.

A field in the LSP header that increases each time a router re-originates an LSP. A received LSP is accepted only if it is newer by sequence number, remaining lifetime, or checksum.

A decrementing field in the LSP header that ages a PDU from its maximum value toward zero. When it expires the LSP is purged and re-flooded with zero lifetime.

Re-flooding an LSP with a zero remaining lifetime to remove it from all databases. The originator keeps the sequence number so that stale copies cannot be reintroduced.

The reliable hop-by-hop distribution of link-state PDUs to all routers at a level. Each router acknowledges receipt and forwards the PDU on other interfaces until every database is synchronized.

The collection of all link-state PDUs received at a given level. Each router computes shortest paths from its own copy, so consistent databases across the level are essential for loop-free routing.

Splitting a router's link-state information across multiple LSPs when it exceeds the maximum PDU size. Fragments share the system ID and differ in the fragment number field of the LSP ID.

A type-length-value encoding used for all variable content in IS-IS PDUs, such as neighbor lists, IP prefixes, and metric values. Extensibility through new TLVs is why IS-IS supports IPv6 and traffic engineering.

The Dijkstra calculation that builds a shortest-path tree from the local link-state database, rooted at the computing router. It yields next hops and costs for all reachable nodes and prefixes.

A recalculation limited to the part of the shortest-path tree affected by a change, such as a prefix addition. It avoids a full SPF run and reduces CPU load in large networks.

The original IS-IS interface metric, a 6-bit value from 1 to 63 for each of the default, delay, expense, and error types. Its small range limits scalability and is why wide metrics were introduced.

A 24-bit interface metric carried in a dedicated TLV, allowing values up to roughly 16.7 million. It is required for traffic engineering and is the default metric style on Nokia SR OS.

The process of moving a network from narrow to wide metrics. Routers must be configured consistently, because mismatched metric styles break reachability even when adjacencies stay up.

A flag in the LSP header that tells other routers not to use the originating router for transit traffic while still allowing it to be reached directly. It is commonly set during startup or maintenance.

TLVs that advertise link bandwidth, administrative groups, and reservable resources so that constraint-based path computation can select MPLS traffic-engineered tunnels across the IS-IS domain.

An extension that lets a single IS-IS instance maintain separate topologies for IPv4, IPv6, or other address families. It allows different paths per family and avoids assuming congruent topologies.

A TLV that carries IPv6 prefix reachability with its metric and up/down bit. It is used in multi-topology deployments to advertise IPv6 prefixes independently of IPv4 topology.

An operating mode in which IPv6 prefixes are advertised in the same topology as IPv4, so both families share one shortest-path tree. It requires that all links support both families.

A TLV carrying a checksum or cryptographic digest that lets routers verify the origin and integrity of Hello and link-state PDUs. Mismatched keys or modes prevent adjacency formation.

The cost assigned to a link for SPF calculation, configured per interface and per level. Path cost is the sum of interface metrics along the shortest path.