For many years I’ve been struggling to memorize OSPF LSA types. Finally, I have decided to create a lab that shows all those LSA types that are supported by Cisco in OSPFv2 (IPv4-OSPF).

Here is my lab topology:

Type 1 – Router LSA

According to the documentation LSA type 1 is sent by any router within the area and contains the information about its directly connected links.

To see it I start packet capture in R1 Fa0/0 interface and reset the OSPF process at R2 to provoke LSA flood. Theoretically I should see two Type 1 LSA coming from R2: one for link 10.0.2.0/30 and another one for 10.0.1.0/24.

Here is what actually appears:

Two LSAs. One with unicast destination and one destined to “ALL OSPF ROUTERS” multicast address. Both contain the same information:

The information on both links is passed inside one Type 1 LSA. Links are presented with their IDs. There is a proper alogithm to derive Link ID depending on link type, in case of Transit that appears in this LSA network it’s DR IP address. The reason why R2 sends two LSA probably has to do with Cisco implementation of OSPF and the fact that R2 has broadcast multicaccess segment connected to it’s Fa0/1 interface.

This is how Type 1 LSA has been sent:

Type 2 – Network LSA

Type 2 LSA is generated by DR elected for multiaccess segment to inform its neighbors about this segment existence. An example of such a segment is the interconnecion between R2 R3 and R4. From R2 we can see that R4 got elected as DR (it’s router ID is 172.16.2.1):

R2#sh ip ospf neighbor Neighbor ID Pri State Dead Time Address Interface 172.16.1.1 1 FULL/BDR 00:00:37 10.0.1.1 FastEthernet0/1 172.16.2.1 1 FULL/DR 00:00:34 10.0.1.2 FastEthernet0/1

So putting capture on R2 fa0/0 and resetting its OSPF process we should get Type 2 LSA from R4 containing information on R2-R3-R4 segment. Here it is, packed together with some other LSAs:

Notice the subnet mask for the segment, DR ID (Advertising Router) and three attached router IDs.

So Type 2 LSA flow is the following:

LSA Type 2

This type of LSA then gets resent by all the routers within the area (in this example R2 will resend it to R1). Advertising router field within the LSA will retain its original value as R4 being the DR is the one responsible for the LSA.

Type 3 – Summary LSA

Type 3 LSAs are sent by ABR. When a link-state change occurs within one of the areas connected to it, it sends this type of LSA to all the other areas thus notifying them about that change. For example, when R2 Fa0/0 link goes down we can expect R3 to notify R6 about this change with Type 3 LSA.

Setting capture for Fa0/0 on R6 and shutting down Fa0/0 on R2 we get the following:

Notice the Metric field value. Setting it to 16777215 (FFFFFF hexadecimal) is a way to mark link as unreachable and signal to other OSPF routers to withdraw its data from their databases.

R3 generates Type 3 LSA because it gets notified of the change by Type 1 LSA coming from R2. Here is the flow diagram:

Please note that the name “summary LSA” has nothing to do with route summarization feature. Nothing gets summarized here, at least by default.

Type 4 – ASBR summary LSA

Type 4 LSAs are also sent by ABR. By using them ABR notifies it’s connected areas about the presence of ASBR in any of these areas. In my topology R3(ABR) would notify R6 which is in Area 1 about the presence of R1(ASBR) in Area 0 using LSA of this type. Later on R6 will receive some external route information from R1 and the Type 5 LSA announcing this route will contain R1 Router ID, saying that this external network is behind R1. Data obtained by R6 from Type 4 LSA will tell it that R1 is behind R3.

What makes router an ASBR and thus provokes LSA Type 4 generation? It’s route redistribution from any other routing process or protocol into OSPF occurring at this router. In my topology R1 is an ASBR as it redistributes routes from RIP into OSPF.

So, again, R3, being an ABR, should notify R6 about the presence of R1 using Type 4 LSA. But how does R3 get to know that R1 is an ASBR and not just any common intra-area router? The answer is: R1 notifies everyone within Area 0 of itself being an ASBR by setting a special (E) flag in Type 1 LSAs it sends. Here it is:

Wireshark already interprets E flag set as “AS boundary router”.

So when R3 receives this LSA it sends Type 4 LSA to notify R6:

Advertising router field contains R3 Router-ID, Link State ID contains R1 (ASBR) Router ID.

Here is what the process looks like:

Type 5 – External LSA

LSA Type 5 is used to advertise external routes in all the OSPF areas. External in this case means not calculated or generated by OSPF process, but redistributed from other routing protocols, redistributed static or connected routes. As an example route for 192.168.99.0/24 network should be advertised by R1 using this type of LSA.

Putting R7 Loopback0 interface into shutdown provokes LSA flooding out of R1 FastEthernet0/0 notifying it as unreachable. Here is the capture:

Type 7 – NSSA External LSA

This LSA Type contains external routes information as well as Type 5 LSA. The difference is that Type 7 LSA is propagated across Not-so-Stubby area and Type 5 LSA is not. On ABR that connects NSSA to Backbone area Type 7 LSA gets converted into Type 5 LSA.

In my topology R5 should notify R4 about changes with interface Loopback0 (172.16.102.0/24 subnet) using this type of LSA.

Here is a Type 7 LSA sent by R5 on it’s Loopback0 shutdown:

and here is LSA 5 generated by R4 for Area0, containing information on the same link: