出自：
http://www.gxlaoyou.com/home.php?mod=space&uid=46&do=blog&id=10
·OSPF屬於IGP，是Link-State協議，基於IP 協議號 89。
·採用SPF算法（Dijkstra算法）計算最佳路徑。（1.最短 2.無環）
·快速響應網絡變化。
·以較低頻率（每隔30分鐘）發送定期更新，被稱爲鏈路狀態刷新。
·網絡變化時是觸發更新。
·支持等價的負載均衡，我們所學的協議中，只有EIGRP支持不等價負載。(默認4，最大16)
一個路由器產生的LSA會把自己所有接口的信息都包含其中，對端鄰居會把LSA放入LSDB中，繼續泛洪此LSA，直到網絡中的所有路由器都收到此LSA。
OSPF既有周期更新(30分)，又有觸發更新。
鏈路狀態路由協議（也可以說OSPF）工作原理：
1-首先每臺路由器通過使用Hello報文與它的鄰居之間建立鄰接關係 。
2-然 後每臺路由器向每個鄰居發送鏈路狀態通告(LSA),有時叫鏈路狀態報文(LSP).今天你先把理解成某條路由(當然它不僅包含路由信息還包含拓撲信 息)，這個LSA我們後面會專門拿1天課程來講！ 每個鄰居在收到(LSA)LSP之後要依次向它的鄰居轉發這些(LSA)LSP(一般我們也叫做Flooding泛洪LSA/LSP)。
3-接下來每臺路由器要在數據庫中保存一份它所收到的LSA/LSP的備份(LSDB)，所有區域內路由器的數據庫(LSDB)應該相同。這張表相對EIGRP的TOP表更詳細！那麼(LSDB)存放的就是一條一條LSA！
4-最後，依照拓撲數據庫(LSDB)-每臺路由器使用Dijkstra算法（SPF算法）計算出到每個網絡的最短無環路徑樹(shortest path first)，並將Cost(其實就是Metric)值最小放到到路由選擇表中！
OSPF的簡化原理：發Hello報文(問候網友是否在？)——建立鄰居關係(組團)——形成鏈路狀態數據庫(使用LSA發送出去)[相當打怪撿到寶物放入寶物庫]——SPF算法(通過寶貝瑪雅之石合成好裝備)——形成路由表(裝備打造成功)！
SPF算法：
1、在一個區域內的所有路由器有同樣的LSDB
2、每一個路由器在計算時都將自已做爲樹根
3、具有去往目標的最低cost值的路由是最好的路徑
4、最好的路由被放入轉發表
Link-State Advertisement-LSA的選擇操作：
LSU是什麼呢？
它就是一個信封(容器)-存放真正的一條或者多條LSA！
當一臺路由器收到Link-State update(LSU)，它會取出裏面的LSA條目，放到鏈路狀態數據庫LS-datebase進行比較:
(1) 如果我沒有！我就會添加到鏈路狀態數據庫LS-datebase！然後反饋一個LSack給LSA的發送源-表明我收到了！（如果對方沒有收到 LSack，對方就會重發一次LSA過來），這樣自己又要向鄰居洪泛(flood)LSA,保證你的鄰居也能收到LSA，洪泛(flood)LSA完成之 後，運行SPF算法選出最新路由表!
(2)如果我已經有一條(SEQ=100)，那麼就比較序列號，就要分三種情況：
①如果你的序列號(seq=100)跟我一樣，那我就直接忽略丟棄掉這條LSA!
②如果我的序列號跟你不一樣,你的序列號(seq=101)比我高,那就添加LSA到LS-datebase返回到(1)操作
③如果我的序列號跟你不一樣,但你序列號(seq=99)比我低，那我就要把LSA從LS-datebase取出來封裝到LSU，發送給信息源，告訴它-要更新的是你而不是我！
Question 1:
Which of the following statements below best describe the process identifier that is used to run OSPF on a router? (Choose two)
A – It is an optional parameter required only if multiple OSPF processes are running on the router
B – It is locally significant
C – It is needed to identify a unique instance of an OSPF database
D – All routers in the same OSPF area must have the same process ID if they are to exchange routing information
Answer: B C
Question 2:
Why R1 can’t establish an OSPF neighbor relationship with R3 according to the following graphic? (Choose two)
A – Configure EIGRP on these routers with a lower administrative distance
B – All routers should be configured for backbone Area 1
C – R1 and R3 have been configured in different areas
D – The hello and dead interval timers are not configured the same values on R1 and R3
Answer: C D
Explanation:
A is not correct because configure EIGRP on these routers (with a lower administrative distance) will force these routers to run EIGRP, not OSPF.
B is not correct because the backbone area of OSPF is always Area 0.
C and D are correct because these entries must match on neighboring routers:
- Hello and dead intervals
– Area ID (Area 0 in this case)
– Authentication password
– Stub area flag
Question 3:
Which items are correct about the routing protocol OSPF? (Choose three)
A – Support VLSM
B – Increase routing overhead on the network
C – Confine network instability to one area of the network
D – Allow extensive control of routing updates
Answer: A C D
Explanation:
Routing overhead is the amount of information needed to describe the changes in a dynamic network topology. All routers in an OSPF area have identical copies of the topology database and the topology database of one area is hidden from the rest of the areas to reduce routing overhead because fewer routing updates are sent and smaller routing trees are computed and maintained (allow extensive control of routing updates and confine network instability to one area of the network).
Question 4:
Which three features are of OSPF routing protocol? (Choose three)
A – Converge quickly
B – OSPF is a classful routing protocol
C – Identify the best route by use of cost
D – Before exchanging routing information, OSPF routers find out neighbors
Answer: A C D
Question 5:
OSPF routing uses the concept of areas. What are the characteristics of OSPF areas? (Chose three)
A – Each OSPF area requires a loopback interface to be configured
B – Areas may be assigned any number from 0 to 65535
C – Area 0 is called the backbone area
D – Hierarchical OSPF networks do not require multiple areas
E – Multiple OSPF areas must connect to area 0
F – Single area OSPF networks must be configured in area 1
Answer: B C E
Explanation:
I used to think the answers should be C D E and here is my explanation:
OSPF can use an active interface for its router ID, so a loopback interface is not a must -> A is incorrect.
OSPF Area is a 32-bit number so we can use up to 232 – 1 = 4294967296 – 1 (since Area 0 is the first area). Remember that only process ID is a 16-bit number and ranges from 1 to 65535 -> B is incorrect.
F is incorrect too because single area OSPF netwoks must be configured in Area 0, which is called the backbone area.
For answer D, it is a bit hard to guess what they want to say about “hierarchical” but we should understand “Hierarchical OSPF networks” as “OSPF networks”. D is correct bercause we can only have one area (area 0 – the backbone area) for our networks.
But TT commented on 01-11-2010:
Especially to note on choice B, D, and E:
Choice B: we all know that The areas can be any number from 0 to 4.2 billion and 1 to 65,535 for the Process ID. As choice B specifies ‘area’ (be aware, it’s not saying ‘process id), there is no reason to say that we cannot assign numbers from 0 to 65535 for area # (it is using ‘may be’, not ‘have to be’ or ‘ought to be’). Hence, we do not worry about assigning ’0′.
Choice E: as Area 0 is the backbone, we all understand that any areas in a OSPF network have to be connected to it. And actually this is implicitly saying that multiple areas form a hierarchical OSPF network, as Area 0 being a root and others being its leaves.
Choice D: when it specifies ‘Hierarchical’, at least 2 areas should be required to form such topology (of course that includes Area 0)
Although Choice B is not an absolutely accurate statement since it not only can be assigned up to 65535, it is still a correct answer. And again, it specifies ‘area’, not ‘process id’, so ’0′ can be included. Finally, it would be meaningless to call OSPF a hierarchical network if no more than one area is present.
—————————————————————————————————-
I reviewed the question and think it is a more suitable solution with choice B than choice D, surely it is a tricky question!
Question 6:
Part of the OSPF network is shown below:
Configuration exhibit:
R1 routing commands:
ip route 0.0.0.0 0.0.0.0 serial0/0
router ospf 1
network 172.16.100.0 0.0.0.3 area 0
network 172.16.100.64 0.0.0.63 area 0
network 172.16.100.128 0.0.0.31 area 0
default-information originate
You work as a network technician, study the exhibits carefully. Assume that all router interfaces are operational and correctly configured. In addition, assume that OSPF has been correctly configured on router R2. How will the default route configured on R1 affect the operation of R2?
A – Any packet destined for a network that is not directly connected to router R2 will be dropped immediately
B – Any packet destined for a network that is not directly connected to router R1 will be dropped
C – Any packet destined for a network that is not directly connected to router R2 will be dropped immediately because of the lack of a gateway on R1
D – The network directly connected to a router R2 will not be able to communicate with the 172.16.100.0, 172.16.100.28 and 172.16.100.64 subnetworks.
E – Any packet destined for a network that is not referenced in the routing table of router R2 will be directed to R1. R1 will then send that packet back to R2 and a routing loop will occur
Answer: E
Explanation:
First, notice that the more-specific routes will always be favored over less-specific routes regardless of the administrative distance set for a protocol. In this case, because we use OSPF for three networks (172.16.100.0 0.0.0.3, 172.16.100.64 0.0.0.63, 172.16.100.128 0.0.0.31) so the packets destined for these networks will not be affected by the default route.
The default route configured on R1 “ip route 0.0.0.0 0.0.0.0 serial0/0″ will send any packet whose destination network is not referenced in the routing table of router R2 to R1, it doesn’t drop anything so answers A, B and C are not correct. D is not correct too because these routes are declared in R1 and the question says that “OSPF has been correctly configured on router R2″, so network directly connected to router R2 can communicate with those three subnetworks.
As said above, the default route configured on R1 will send any packet destined for a network that is not referenced in its routing table to R2; R2 in turn sends it to R1 because it is the only way and a routing loop will occur.