link state routing - École polytechnique fédérale de ...ica · link state routing jean-yves le...
TRANSCRIPT
Link State Routing
Jean-Yves Le BoudecFall 2009
1
ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE
Contents
1. Link state flooding topology information
finding the shortest paths (Dijkstra)
2. Hierarchical routing with areas3. OSPF
database modelling
neighbor discovery - Hello protocol
database synchronization
link state updates
examples
2
1. Link State Routing
Principle of link state routingeach router keeps a topology database of whole network
link state updates flooded, or multicast to all network
routers compute their routing tables based on topologyoften uses Dijkstra’s shortest path algorithm
Used in OSPF (Open Shortest Path First), IS-IS (similar to OSPF)and PNNI (ATM routing protocol)
3
(a) Topology Database Synchronization
Neighbouring nodes synchronize before starting any relationshipHello protocol; keep alive
initial synchronization of database
description of all links (no information yet)
Once synchronized, a node accepts link state advertisementscontain a sequence number, stored with record in the database
only messages with new sequence number are accepted
accepted messages are flooded to all neighbours
sequence number prevents anomalies (loops or blackholes)
4
Example network
5
n1
A
B
n6
D E
n4
n3
C
n5n2
F
n7
Each router knows directly connected networks
Initial routing tables
6
net typen1 Ethern2 P-to-P
A
n1
A
B
n6
D E
n4
n3
C
n5n2
F
n7
net typen6 Ethern5 P-to-P
Dnet typen6 Ethern7 Ether
E
net typen1 Ethern7 Ether
F
net typen1 Ethern4 P-to-Pn5 P-to-P
C
net typen3 Ethern2 P-to-Pn4 P-to-P
B
After Flooding
7
rtr net costrtr net costA n1 10A n2 100B n3 10B n2 100B n4 100C n1 10C n4 100C n5 100D n6 10D n5 100E n6 10E n7 10F n1 10F n7 10
A n1 10A n2 100B n3 10B n2 100B n4 100C n1 10C n4 100C n5 100D n6 10D n5 100E n6 10E n7 10F n1 10F n7 10
The local metric information is flooded to all routersAfter convergence, all routers have the same information
n1
A
B
n6
D E
n4
n3
C
n5n2
F
n7
(b) Topology graph
8
Arrows routers-to-nets with a given metricexcept P-to-P, stub, and external networks
From nets to routers, metric = 0
A
B
C
D
F
E
100
10 10
10
10
100100
n1
100 n6
n7
n3
100
10
100
1010
10
0
0
0
0
0 0
externalnetwork
54
0
stub network
point to pointlink
broadcastnetwork
externalnetwork
(b) Path ComputationPerformed locally, based on topology databaseComputes one or several best paths to every destination from this nodeBest Path = shortest for OSPFOSPF uses Dijkstra’s shortest path
the best known algorithm for centralized operation
Paths are computed independently at every nodesynchronization of databases guarantees absence of persistent loops
every node computes a shortest path tree rooted at self
9
Simplified graph
10
Only arrows with metrics between routersEvery node executes the shortest path computation on the graph – same graph, but different sources
A
B
C
D
F
E
100
10
10
10
10
10
100 100
Dijkstra’s Shortest Path AlgorithmThe nodes are 0...N and the algorithm computes best paths from node 0c(i,j) is the cost of (i,j), pred(i) is the predecessor of node i on the tree M being builtm(j) is the distance from node 0 to node j.
as Bellman-Ford, works for any min-plus algebra11
m(0) = 0; M = {0};for k=1 to N {
find (i0, j0) that minimizes m(i) + c(i,j), with i in M, j not in M
m(j0) = m(i0) + c(i0, j0)pred(j0) = i0M = M ∪ {j0}
}
m(0) = 0; M = {0};for k=1 to N {
find (i0, j0) that minimizes m(i) + c(i,j), with i in M, j not in M
m(j0) = m(i0) + c(i0, j0)pred(j0) = i0M = M ∪ {j0}
}
Example: Dijkstra at A
12
A
B
C
D
F
E
100
10
10
10
10
10
100 100
init: M = { A }
step 1:i0=Aj0=Cm(C)=10
M = {A, C}
m(A)=0 m(C)=10
Example: Dijkstra at A
13
A
B
C
D
F
E
100
10
10
10
10
10
100 100
i0=Aj0=Fm(F)=10M = {A,C,F}
m(A)=0 m(C)=10 m(F)=10
Example: Dijkstra at A
14
A
B
C
D
F
E
100
10
10
10
10
10
100 100
i0=Fj0=Em(E)=20M = {A,C,F,E}
m(A)=0 m(C)=10 m(F)=10
m(F)=20
Example: Dijkstra at A
15
A
B
C
D
F
E
100
10
10
10
10
10
100 100
i0=Ej0=Dm(D)=40M = {A,C,F,E,D}
m(A)=0 m(C)=10 m(F)=10
m(E)=20m(D)=30
Example: Dijkstra at A
16
A
B
C
D
F
E
100
10
10
10
10
10
100 100
i0=Aj0=Bm(B)=100M = {A,C,F,E,D,B}
m(A)=0 m(C)=10 m(F)=10
m(E)=20m(D)=30m(B)=100
Routing table of A
17
net nextn1 directn2 directn3 Bn4 Cn5 Cn6 Fn7 F
A
n1
A
B
n6
D E
n4
n3
C
n5n2
F
n7
Test Your Understanding
Q1: Run Dijkstra at CQ2: What are the routing tables at C
solution
18
LS: SummaryAll nodes compute their own topology database
represents the whole network
strongly synchronized
All nodes compute their best path tree to all destinationsRouting tables are built from the tree
used for next hop routing only
LS versus DVLS avoids convergence problems of DV
supports flexible cost definitions; can be used for routing ATM connections
LS is much more complex
19
2. Divide large networks Why divide large networks?Cost of computing routing tables
update when topology changes
SPF algorithmn routers, k linkscomplexity O(n*k)
size of DB, update messages grows with the network size
Use hierarchical routing to limit the scope of updates and computational overhead
divide the network into several areas
independent route computing in each area
inject aggregated information on routes into other areas
We explain hierarchical routing the OSPF wayIS-IS does things a bit differently
20
Hierarchical Routing
A large OSPF domain can be configured into areasone backbone area (area 0)
non backbone areas (areas numbered other than 0)
All inter-area traffic goes through area 0strict hierarchy
Inside one area: link state routing as seen earlierone topology database per area
21area 0
B1X3
X1
X4A1area 2area 1
X1
X3X4 B2A2
Principles Routing method used in the higher level:
distance vectorno problem with loops - one backbone area
Mapping of higher level nodes to lower level nodesarea border routers (inter-area routers) belong to both areas
Inter-level routing informationsummary link state advertisements (LSA) from other areas are injected into the local topology databases
22
Example
Assume networks n1 and n2 become visible at time 0. Show the topology databases at all routers
solution
23
area 0
B1X4
X1
X3A1area 2area 1
X2
X6X5 B2A2
n1
n2
10
10
10
66
666
6
10
1010
HintsAll routers in area 2 propagate the existence of n1 and n2, directly attached to B1 (resp. B2). Draw the topology database in area 2.
Area border routers X4 and X6 belong to area 2, thus they can compute their distances to n1 and n2
Area border routers X4 and X6 inject their distances to n1 and n2 into the area 0 topology database (item 3 of the principle). The corresponding summary link state record is propagated to all routers of area 0. Draw now the topology database in area 0.
All routers in area 0 can now compute their distance to n1 and n2, using their distances to X4 and X6, and using the principle of distance vector (item 1 of the principle). Do the computation for X3 and X5.
Area border routers X3 and X5 inject their distances to n1 and n2 into the area 1 topology database (item 3 of the principle). Draw now the topology database in area 1.
24
CommentsDistance vector computation causes none of the RIP problems
strict hierarchy: no loop between areas
External and summary LSA for all reachable networks are present in all topology databases of all areas
most LSAs are externalcan be avoided in configuring some areas as terminal: use defaultentry to the backbone
Area partitions require specific supportpartition of non-backbone area is handled by having the area 0 topology database keep a map of all area connected components
partition of backbone cannot be repaired; it must be avoided; can be handled by backup virtual area 0 links through non backbone area
25
*Example of issue : partitioned backbone
No connectivity between areas via backbone There is a route through Area 2Virtual link
X4 and X6 configure a virtual link through Area 2
virtual link entered into the database, metric = sum of links
26
area 0
B1X4
X1
X3A1area 2area 1
X2
X6X5 B2A2
n1
n2
10
10
10
6∞
66∞
6
10
1010
3. The OSPF ProtocolOSPF (Open Shortest Path First)
IETF standard for internal routing
used in large networks (ISPs)
Link State protocol + Hierarchical
27
*OSPF Components
On top of IP (protocol type = 89)Multicast
224.0.0.5 - all routers of a link
224.0.0.6 - all designated and backup routers
Sub-protocolsHello to identify neighbors, elect a designated and a backup router
Database description to diffuse the topology between adjacent routers
Link State to request, update, and ack the information on a link (LSA - Link State Advertisement)
28
*TOS and metric
TOSmapping of 4 IP TOS bits to a decimal integer
0 - normal service
2 - minimize monetary cost
4 - maximize reliability
8 - maximize throughput
16 - minimize delay
Metrictime to send 100 Mb over the interface
C = 108/bandwidth
1 if greater than 100 Mb/s
can be configured by administrator
29
*OSPF - summary
OSPF vs. RIPmuch more complex, but presents many advantages
no count to infinityno limit on the number of hops (OSPF topologies limited by Network and Router LSA size (max 64KB) to O(5000) links)less signaling traffic (LS Update every 30 min)advanced metriclarge networks - hierarchical routing
most of the traffic when change in topologybut periodic Hello messagesin RIP: periodic routing information traffic
drawbackdifficult to configure
30
Solutions
31
Test Your Understanding
Q1: Run Dijkstra at CA: (final step)Q2: What are the routing tables at C
32
A
B
C
D
F
E
100
10
10
10
10
10
100 100
m(F)=10m(C)=0m(A)=10
m(F)=20m(D)=30m(B)=100
Test Your Understanding
Q2: What are the routing tables at CA:
33
net nextn1 directn2 An3 Bn4 directn5 directn6 Fn7 F
C
n1
A
B
n6
D E
n4
n3
C
n5n2
F
n7
back
Solution
34
area 0
B1X4
X1
X3A1area 2area 1
X2
X6X5 B2A2
n1
n2
10
10
10
66
666
6
n1
n2
area 2 topology database
area 0 topology database
n1, d=10n2, d=16
n1, d=16n2, d=10
n1, d=28n2, d=22
n1, d=22n2, d=16
10
1010
area 1 topology database
back