skip to content

In OSPFv2, what do LSA types 1, 2, 3, 4, 5 and 7 describe, who originates each, and how far does each flood?

level: middleimportance: must knowfreq 45%

answer

  1. topology versus reachability
  2. the DR speaks for a segment
  3. ABRs rewrite at every border
  4. one type floods unchanged AS-wide
  5. a not-so-stubby exception

basics

~20 s

Types 1 (router) and 2 (network, from the DR) map an area's topology and stay inside it; ABRs originate 3 (network summary) and 4 (ASBR summary) per area; ASBRs' type 5 floods AS-wide except stub areas and NSSAs, whose own externals use type 7.

solid answer

~40 s

A **type 1 router-LSA** comes from every router and lists its links in one area; a **type 2 network-LSA** comes from the designated router of each multi-access segment and lists the routers on it. Both are flooded through one area only, and together they are the area's topology. ABRs originate **type 3 summary-LSAs** (a network or range outside the area) and **type 4 summary-LSAs** (the route to an ASBR) into each area; they are area-scoped too, so the next ABR originates its own copy with its own cost. An ASBR originates **type 5 AS-external-LSAs**, the only kind that floods unchanged through the whole AS, except stub areas and NSSAs. Inside an NSSA (RFC 3101) an ASBR originates **type 7** instead, which never leaves that area; an NSSA ABR translates it into a type 5.

go deeper

for a junior

Know that routers describe their links in one LSA type, ABRs summarise between areas, and external routes have their own LSA type.

for a middle

Give all six types with originator and scope, and explain why type 4 exists when type 5 already names its ASBR.

for a senior

Explain that summaries are regenerated at each ABR, which types each area type admits, and how to localise a missing route by LSA type.

for a principal

Weigh how much of the database each area type removes against the path information that loss costs routers inside it.

## Why OSPF has several LSA types A **link-state advertisement (LSA)** is one unit of the OSPF database. RFC 2328 splits the database into kinds because areas need two different sorts of information: - **Topology** inside an area: which router connects to which segment, at what cost. Every router in the area needs all of it to run SPF. - **Reachability** from outside the area: which destinations exist beyond the border and at what cost, with no detail of the path. Each LSA is identified by its **LS type**, its **Link State ID** and its **Advertising Router**, and each type has a **flooding scope** that decides how far it travels. ## The six types interviewers ask about | Type | Name | Originated by | Describes | Flooding scope | |---|---|---|---|---| | 1 | Router-LSA | Every router, once per attached area | Its links in that area and their costs; B, E and V bits | One area | | 2 | Network-LSA | The designated router of a broadcast or NBMA segment with two or more routers | The routers attached to that segment | One area | | 3 | Summary-LSA (IP network) | ABRs | A network or area address range outside the area, with a cost | One area; each ABR re-originates | | 4 | Summary-LSA (ASBR) | ABRs | The cost to reach an ASBR located outside the area | One area; each ABR re-originates | | 5 | AS-external-LSA | ASBRs | A destination outside the AS, with external metric type 1 or 2 and a forwarding address | The whole AS, except stub areas and NSSAs | | 7 | Type-7 AS-external-LSA (RFC 3101) | ASBRs inside an NSSA, and NSSA ABRs for a default | An external destination, same layout as type 5 | One NSSA only | Some identifiers worth knowing: - A router-LSA's Link State ID is the router's **Router ID**; its links are typed point-to-point, transit network, stub network or virtual link. - A network-LSA's Link State ID is the designated router's **interface address**, and it has no metrics: the cost from a segment to its routers is zero. - A type 3 LSA's ID is the **network address**; a type 4 LSA's ID is the **ASBR's Router ID**. ## Which LSA crosses which border The phrase "crosses a border" hides the real mechanism: 1. **Types 1 and 2 never cross.** They are the area's topology, and that topology is exactly what areas hide. 2. **Types 3 and 4 do not cross either: they are regenerated.** An ABR runs SPF in each attached area, then originates fresh summary-LSAs into every other attached area, with **itself** as advertising router and **its own** cost. The next ABR down the line repeats the process. A summary seen in area 2 names area 2's ABR, not the router that first had the route. 3. **Type 5 is the one that floods unchanged.** Its advertising router stays the original ASBR everywhere. That is why type 4 exists: a router in another area needs a route to that ASBR, which its own area's topology does not contain. 4. **Type 7 stops at the NSSA border.** An NSSA ABR elected or configured as translator originates a new type 5 for each type 7 whose **P-bit** is set and whose forwarding address is non-zero, and floods that type 5 into the other areas that accept externals. A type 7 address range on the translator can fold several type 7s into one type 5. Two rules decide what ABRs put in summaries: - AS-external routes are **never** advertised in summary-LSAs. - Only **intra-area** routes are summarised into the backbone; both intra-area and inter-area routes are summarised into other areas. ## How area types prune the list | Area type | Types present inside it | |---|---| | Normal area or backbone | 1, 2, 3, 4, 5 | | Stub area | 1, 2, 3 (including a default), no 4, no 5 | | NSSA | 1, 2, 3, 7, no 4, no 5 | An implementation feature often called "totally stubby" (a term in no RFC) goes further and lets the ABR send only the default type 3 into a stub area. ## Reading a database by type When a route looks wrong, the type tells you where to look: - A missing **intra-area** route points at types 1 and 2 inside the area. - A missing **inter-area** route points at the type 3 from the ABR and at the ABR's own view of the other area. - A missing **external** route needs both the type 5 and a path to its ASBR: a router or type 4 inside the area, or a forwarding address it can reach.

  • Why does an OSPF type 3 LSA's advertising router change at each ABR, while a type 5's never does?
    Summary-LSAs are area-scoped: each ABR computes routes in one area and originates new summaries into another, signed with its own Router ID and cost. AS-external-LSAs are AS-scoped and flooded as originated, so the ASBR stays the advertising router everywhere; routers in other areas then need a type 4 to find that ASBR.
  • Which LSA types does an internal router in an OSPF stub area hold?
    Types 1 and 2 for its own area, plus type 3 summaries for inter-area destinations and a type 3 default that its ABRs originate. It holds no type 5, because externals are not flooded into stub areas, and no type 4, because without type 5 there is no ASBR to reach. With the ABR suppressing summaries, only the default type 3 remains.

saying these in an interview costs you the question

  • Type 3 summary-LSAs are flooded across the whole AS unchanged.
  • Type 4 LSAs carry the external prefixes themselves.
  • Every router on a broadcast segment originates its own network-LSA.
  • Type 7 LSAs are flooded into area 0 so the backbone learns NSSA routes.
  • Type 5 LSAs reach every area, stub areas included.
  • An ABR summarises type 5 external routes into type 3 LSAs.