Two more OSPF questions

Answered Question
Oct 26th, 2009

Quiz:

1. What is a neighbor on a local LAN segment in OSPF?

a. A neighbor is a router in the same area

b. A neighbor is a router in the same classful network

c. A neighbor is a router on a multiaccess link, with an adjacency with a DR

d. A neighbor is another router with the same network address

8. In learning a new route, what will an internal OSPF router do if a received LSA is not found in the topological database?

a. The LSA is flooded immediately out of all the OSPF interfaces, except the interface from which the LSA was received.

b. The LSA is dropped and a message is sent to the transmitting router.

c. The LSA is placed in the topological database and an acknowledgment is sent to the transmitting router

d. The sequence numbers are checked, and if the LSA is valid, it is entered into the topology database.

These are the only two questions I got wrong on the chapter quiz, except I still kinda stand by my answers. I'd like to see how others weigh in on this, to see if I'm really misunderstanding something or if the book is wrong (which these BSCI prep books so often are...)

I have this problem too.
0 votes
Correct Answer by marikakis about 7 years 1 month ago

The books do need editing. However, your last post refers to a different OSPF procedure than the 2nd question posted initially. An "unrecognized/unknown LSA" is different from an LSA not already installed in the link-state database. An "unrecognized/unknown LSA" is an LSA with an LS-Type not supported by the OSPF implementation of a particular router (e.g. not one of the LS types 1-5). (Unless I misunderstood something, and you are referring to a question in the book you did not post in this thread.) Anyway, the phrase "unknown LSA" can be misleading.

For question 1, the answers a, b are obviously wrong, so we are left with c,d. The defence I can think of for the book answer (that is answer d) is the following: c is wrong because the router that achieved adjacency with a DR, is not necessarily our neighbor, unless our interface is up/up and we follow proper Hello procedures, etc. Questions involving 'neighbors' are often a mess because no single router is a neighbor on its own. You need 2 routers to define what a neighbor is. Answer d is better in this respect. To be honest, when I first read answer d I thought it was referring to a router having the exact IP address as we do, so I thought of it as being very wrong! Only after the process of elimination did I understand that it refers to the "same subnet" criterion for neighbors. Of course, other criteria have to be met as well as others have said.

The second question needs more work. b is obviously wrong. d is also wrong because sequence numbers are irrelevant in this case (LSA is not in the database, we are not trying to see if it is more recent than some other in the database), and an LS sequence number check is not a validity check to solely determine installation of the LSA. So, we are left with a and c.

Answer c is not completely wrong, but the phrase "transmitting router" suggests a "direct acknowledgement" procedure, which is used in special cases (sending ack immediately to the neighbor's IP address when a duplicate LSA is received) and not always, while also acks can be sent as multicasts. See RFC 2328 (Section 13.5. Sending Link State Acknowledgment packets).

RFC 2328 (Section 13. The Flooding Procedure) says: "(5) Otherwise, find the instance of this LSA that is currently contained in the router's link state database. If there is no database copy, or the received LSA is more recent than the database copy...the following steps must be performed:

(a) If there is already a database copy ...

(b) Otherwise immediately flood the new LSA out some subset of the router's interfaces (see Section 13.3). In some cases (...) the LSA will be flooded back out the receiving interface. This occurrence should be noted for later use by the acknowledgment process (Section 13.5)...

(e) Possibly acknowledge the receipt of the LSA by sending a Link State Acknowledgment packet back out the receiving interface."

Note point (e) above, especially the words "possibly" and "back out receiving interface", which does not necessarily mean "to the transmitting router". I am not saying answer a is always true (see point b above), but it is more generally true, which possibly makes it a better answer than the others.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
mahmoodmkl Tue, 10/27/2009 - 00:13

Hi

I would say as below.

for the 1 the answer would be c.

for the 2 the answer would be c.

Thanks

Mahmood

CriscoSystems Tue, 10/27/2009 - 10:33

I also put "c" for question 1 as well. The @#*%!ing Cisco book's answer key says it's "d." (The question DID specify "local LAN segment, which tells us there must be a DR.)

For question 8, I put "a" and "c." The book says only "a" is correct. Why not "c" though? I thought LS updates were explicitly acknowledged with LSAck packets in OSPF - or is that _only_ in the "LOADING" state?

royalblues Tue, 10/27/2009 - 12:28

Well it is possible that all the routers in the LAN segment are having a ospf priority of 0 in which they would still become neighbors (you can see them using the sh ip osp neighbor command) and are not having an adjacency with DR.

This makes "d" the best answer, htough it would be hard to understand why would someone configure something like this in a real world.

HTH

Narayan

Peter Paluch Tue, 10/27/2009 - 12:31

Hi Seth,

To be completely honest, I am confused by the available answers to your quiz questions.

Regarding the question about who is a neighbor - by definition, a neighbor is a router whose selected OSPF parameters align to ours, in particular:

- The network mask must be identical (except on point-to-point links)

- The Hello and Dead intervals must be identical

- The area number must be identical

- The area type must be identical

- If authentication is used, it must be successful

Of the available answers, I cannot really choose which one tackles the point.

Regarding the receipt of a previously unknown LSU, you are completely correct - LSUs are sent reliably so there must be some sort of acknowledgement. That acknowledgement must either be done using the LSAck packet, or it can be done implicitely by the receiving router sending the LSU back. This is done on multiaccess segments where a non-DR router send a LSU to the DR/BDR. The DR will not acknowledge its arrival explicitely with LSAck, rather, it will flood it back to all routers on a segment which serves as an implicit acknowledgement.

Best regards,

Peter

CriscoSystems Tue, 10/27/2009 - 19:06

Official 642-901 Exam Certification Guide, Ch 9 (Fundamentals of IS-IS), page 259. Discussing the differences between IS-IS and OSPF: "Unrecognized LSPs are ignored and flooded in IS-IS; OSPF ignores and drops unrecognized LSAs."

And of course the answer-key to the OSPF quiz in the exact same book says an OSPF router floods an unknown LSA out all interfaces.

WHY.

DOES.

CISCO.

NOT.

EDIT.

THEIR.

FREAKIN'.

TEXTBOOKS.

Correct Answer
marikakis Tue, 10/27/2009 - 21:05

The books do need editing. However, your last post refers to a different OSPF procedure than the 2nd question posted initially. An "unrecognized/unknown LSA" is different from an LSA not already installed in the link-state database. An "unrecognized/unknown LSA" is an LSA with an LS-Type not supported by the OSPF implementation of a particular router (e.g. not one of the LS types 1-5). (Unless I misunderstood something, and you are referring to a question in the book you did not post in this thread.) Anyway, the phrase "unknown LSA" can be misleading.

For question 1, the answers a, b are obviously wrong, so we are left with c,d. The defence I can think of for the book answer (that is answer d) is the following: c is wrong because the router that achieved adjacency with a DR, is not necessarily our neighbor, unless our interface is up/up and we follow proper Hello procedures, etc. Questions involving 'neighbors' are often a mess because no single router is a neighbor on its own. You need 2 routers to define what a neighbor is. Answer d is better in this respect. To be honest, when I first read answer d I thought it was referring to a router having the exact IP address as we do, so I thought of it as being very wrong! Only after the process of elimination did I understand that it refers to the "same subnet" criterion for neighbors. Of course, other criteria have to be met as well as others have said.

The second question needs more work. b is obviously wrong. d is also wrong because sequence numbers are irrelevant in this case (LSA is not in the database, we are not trying to see if it is more recent than some other in the database), and an LS sequence number check is not a validity check to solely determine installation of the LSA. So, we are left with a and c.

Answer c is not completely wrong, but the phrase "transmitting router" suggests a "direct acknowledgement" procedure, which is used in special cases (sending ack immediately to the neighbor's IP address when a duplicate LSA is received) and not always, while also acks can be sent as multicasts. See RFC 2328 (Section 13.5. Sending Link State Acknowledgment packets).

RFC 2328 (Section 13. The Flooding Procedure) says: "(5) Otherwise, find the instance of this LSA that is currently contained in the router's link state database. If there is no database copy, or the received LSA is more recent than the database copy...the following steps must be performed:

(a) If there is already a database copy ...

(b) Otherwise immediately flood the new LSA out some subset of the router's interfaces (see Section 13.3). In some cases (...) the LSA will be flooded back out the receiving interface. This occurrence should be noted for later use by the acknowledgment process (Section 13.5)...

(e) Possibly acknowledge the receipt of the LSA by sending a Link State Acknowledgment packet back out the receiving interface."

Note point (e) above, especially the words "possibly" and "back out receiving interface", which does not necessarily mean "to the transmitting router". I am not saying answer a is always true (see point b above), but it is more generally true, which possibly makes it a better answer than the others.

Actions

This Discussion