Routing in Switched Networks
Chapter 19 in Stallings 10th Edition
CS420/520 Axel Krings Page 1 Sequence 18 Chap 19
Routing in Packet Switching
Networks
• Key design issue for (packet) switched networks
• Select route across network between end nodes
• Characteristics required:
—Correctness
—Simplicity
—Robustness
—Stability
—Fairness
—Optimality
—Efficiency
CS420/520 Axel Krings Page 2 Sequence 18 Chap 19
1
Routing in Circuit Switched
Network
• Many connections will need paths through more
than one switch
• Need to find a route
—Efficiency
—Resilience
• Public telephone switches are a tree structure
—Static routing uses the same approach all the time
• Dynamic routing allows for changes in routing
depending on traffic
—Uses a peer structure for nodes
CS420/520 Axel Krings Page 3 Sequence 18 Chap 19
Elements of Routing Techniques for
Packet-Switching Networks
Table 19.1
Performance Criteria Network Information Source
Number of hops None
Cost Local
Delay Adjacent node
Throughput Nodes along route
All nodes
Decision Time
Packet (datagram) Network Information Update Timing
Session (virtual circuit) Continuous
Periodic
Decision Place Major load change
Each node (distributed) Topology change
Central node (centralized)
Originating node (source)
CS420/520 Axel Krings Page 4 Sequence 18 Chap 19
2
Performance Criteria
• Used for selection of route
—Minimum hop
—Least cost
—Delay
—Throughput
• Simplest is to choose “minimum hop”
• Can be generalized as “least cost” routing
• “least cost” is more flexible and is more
common than “minimum hop”
CS420/520 Axel Krings Page 5 Sequence 18 Chap 19
Example Packet Switched
Network
◆Example
–communicating nodes: node-1 to node-6
–what is of interest?
» Shortest path (1-3-6)
» least cost path (1-4-5-6)
CS420/520 Axel Krings Page 6 Sequence 18 Chap 19
3
Decision Time and Place
• Time
—Packet or virtual circuit basis
• Place
—Distributed
• Made by each node
—Centralized
• requires central node
—Source
• originating node
CS420/520 Axel Krings Page 7 Sequence 18 Chap 19
Network Information Source
and Update Timing
• Routing decisions usually based on knowledge of network
— (not always)
• Distributed routing
— Nodes use local knowledge
— May collect info from adjacent nodes
— May collect info from all nodes on a potential route
• Central routing
— Collect info from all nodes
• Update timing
— When is network info held by nodes updated?
• Fixed - never updated
• Adaptive - regular updates
• Continuous
• Periodic
• Major load change
• Topology change
CS420/520 Axel Krings Page 8 Sequence 18 Chap 19
4
Routing Strategies
• We will discuss several strategies:
—Fixed Routing
—Flooding Routing
—Random Routing
—Adaptive Routing
CS420/520 Axel Krings Page 9 Sequence 18 Chap 19
Routing Strategies - Fixed
Routing
• Use a single permanent route for each source to
destination pair of nodes
• Determined using a least cost algorithm
• Route is fixed
—Until a change in network topology
—Based on expected traffic or capacity
• Advantage is simplicity
• Disadvantage is lack of flexibility
—Does not react to network failure or congestion
CS420/520 Axel Krings Page 10 Sequence 18 Chap 19
5
Fixed Routing
Tables
CS420/520 Axel Krings Page 11 Sequence 18 Chap 19
Routing Strategies - Flooding
• Packet sent by node to every neighbor
• Eventually multiple copies arrive at destination
• No network information required
• Each packet is uniquely numbered so duplicates
can be discarded
• Need to limit incessant retransmission of
packets
—Nodes can remember identity of packets
retransmitted
—Can include a hop count in packets
CS420/520 Axel Krings Page 12 Sequence 18 Chap 19
6
Flooding 2 3
Example
6
3
3
1
4 5
Figure 19.3
(a) First hop
2 2 2
2 3
Assigned Hop
2
2
2
6
Count is 3
1 2
2
2
4 5
(b) Second hop
1
1
1 1 1
1
2 3 1
1
1
1
1
6
1
1
1
1
1
1
1 1 1 1 1
4 5
(c) Third hop
CS420/520 Axel Krings Page 13 Sequence 18 Chap 19
Figure 19.3 Flooding Example (hop count = 3)
Properties of Flooding
Can be used
All possible
Highly to send
routes are
robust emergency
tried messages
At least one
packet will have
taken minimum
hop route
Nodes directly or
indirectly
connected to
source are visited
High traffic
Security
Disadvantages: load
concerns
generated
CS420/520 Axel Krings Page 14 Sequence 18 Chap 19
7
Random Routing
• Node selects one outgoing path for retransmission of
incoming packet
• Selection can be random or round robin
— Can select outgoing path based on probability calculation, i.e.
• Pi probability of selecting link i
Ri
Pi =
• Ri data rate of link i ∑R j
• Sum is taken over all outgoing candidate links j
• No network info needed
• Route is typically not least cost nor minimum hop
€
CS420/520 Axel Krings Page 15 Sequence 18 Chap 19
Routing Strategies - Adaptive
Routing
• Used by almost all packet switching networks
• Routing decisions change as conditions on the
network change due to failure or congestion
• Requires information about network
Disadvantages: Decisions more complex
Tradeoff between quality of network information and
overhead
Reacting too quickly can cause oscillation
Reacting too slowly means information may be
irrelevant
CS420/520 Axel Krings Page 16 Sequence 18 Chap 19
8
Classification of Adaptive
Routing Strategies
• A convenient way to classify is on the basis of
information source
Local ••Route to outgoing link with shortest queue
••Can include bias for each destination
(isolated) ••Rarely used - does not make use of available
information
Adjacent ••Takes advantage of delay and outage information
nodes ••Distributed or centralized
All nodes ••Like adjacent
CS420/520 Axel Krings Page 17 Sequence 18 Chap 19
To 2
Node 4's Bias To 3
Table for To 1
Destination 6
Next Node Bias
1 9
2 6 To 5
3 3 Node 4
5 0
Figure 19.4 Example of Isolated Adaptive Routing
CS420/520 Axel Krings Page 18 Sequence 18 Chap 19
9
Isolated Adaptive Routing
Algorithm:
minimize Q + Bi
where
Q is queue length
Bi is bias for destination i
CS420/520 Axel Krings Page 19 Sequence 18 Chap 19
ARPANET Routing Strategies
1st Generation
• Distributed adaptive using estimated delay
—Queue length used as estimate of delay
• Version of Bellman-Ford algorithm
• Node exchanges delay vector with neighbors
• Update routing table based on incoming
information
• Doesn't consider line speed, just queue length
and responds slowly to congestion
CS420/520 Axel Krings Page 20 Sequence 18 Chap 19
10
Desti- Next Desti- Next
nation Delay node nation Delay node
1 0 — 3 7 5 1 0 —
2 2 2 0 4 2 2 2 2
3 5 3 3 0 2 3 3 4
4 1 4 2 2 0 4 1 4
5 6 3 3 1 1 5 2 4
6 8 3 5 3 3 6 4 4
I1,2 = 2
D1 S1 D2 D3 D4
I1,3 = 5
I1,4 = 1
(a) Node 1's Routing (b) Delay vectors sent to node 1 (c) Node 1's routing table
table before update from neighbor nodes after update and link
costs used in update
Figure 19.5 Original ARPANET Routing Algorithm
CS420/520 Axel Krings Page 21 Sequence 18 Chap 19
ARPANET Routing Strategies
2nd Generation
• Distributed adaptive using delay criterion
—Using timestamps of arrival, departure and ACK times
• Re-computes average delays every 10 seconds
• Any changes are flooded to all other nodes
• Re-computes routing using Dijkstra’s algorithm
• Good under light and medium loads
• Under heavy loads, little correlation between
reported delays and those experienced
CS420/520 Axel Krings Page 22 Sequence 18 Chap 19
11
ARPANET Routing Strategies (3)
• Third Generation (1987)
—Link cost calculations changed
—Measure average delay over last 10 seconds
—Normalize based on current value and
previous results
CS420/520 Axel Krings Page 23 Sequence 18 Chap 19
Least Cost Algorithms
• Basis for routing decisions
— Can minimize hop by setting each link cost to unity
— Can have link value inversely proportional to capacity
• Given network graph
— Nodes connected by bi-directional links
— Each link has a cost in each direction
• Define cost of path between two nodes as sum of costs
of links traversed
• For each pair of nodes, find a path with the least cost
• Link costs in different directions may be different
— E.g. length of packet queue
CS420/520 Axel Krings Page 24 Sequence 18 Chap 19
12
Dijkstra’s Algorithm Definitions
• Find shortest paths from given source to all other nodes,
by developing paths in order of increasing path length
• N =set of nodes in the network
• s = source node
• T = set of nodes so far incorporated by the algorithm
• w(i, j) = link cost from node i to node j
— w(i, i) = 0
— w(i, j) = ¥ if the two nodes are not directly connected
— w(i, j) ³ 0 if the two nodes are directly connected
• L(n) = cost of least-cost path from node s to node n
currently known
— At termination, L(n) is cost of least-cost path from s to n
CS420/520 Axel Krings Page 25 Sequence 18 Chap 19
Dijkstra’s Algorithm Method
• Step 1 [Initialization]
— T = {s} Set of nodes so far incorporated consists of only source node
— L(n) = w(s, n) for n ≠ s
— Initial path costs to neighboring nodes are simply link costs
• Step 2 [Get Next Node]
— Find neighboring node x not in T with least-cost path from s
— Incorporate node into T
• Step 3 [Update Least-Cost Paths]
— L(n) = min[L(n), L(x) + w(x, n)] for all n Ï T
— If latter term is minimum, path from s to n is path from s to x
concatenated with edge from x to n
• Algorithm terminates when all nodes have been added to T
CS420/520 Axel Krings Page 26 Sequence 18 Chap 19
13
Dijkstra’s Algorithm Notes
• At termination, value L(x) associated with each
node x is cost (length) of least-cost path from s
to x.
• In addition, T defines least-cost path from s to
each other node
• One iteration of steps 2 and 3 adds one new
node to T
—Defines least cost path from s to that node
CS420/520 Axel Krings Page 27 Sequence 18 Chap 19
Example of Dijkstra’s Algorithm
CS420/520 Axel Krings Page 28 Sequence 18 Chap 19
14
Results of Example
Dijkstra’s Algorithm
CS420/520 Axel Krings Page 29 Sequence 18 Chap 19
Bellman-Ford Algorithm
Definitions
• Essential idea
— first, find shortest paths from given node subject to the constraint that
the paths contain at most 1 link
— next, find the shortest paths with a constraint of paths of at most 2
links
— and so on
• Definitions
— s = source node
— w(i, j) = link cost from node i to node j
• w(i, i) = 0
• w(i, j) = ¥ if the two nodes are not directly connected
• w(i, j) ³ 0 if the two nodes are directly connected
— h = maximum number of links in path at current stage of the algorithm
• i.e. h = max length of a path currently considered
— Lh(n) = cost of least-cost path from s to n under constraint of no more
than h links
CS420/520 Axel Krings Page 30 Sequence 18 Chap 19
15
Bellman-Ford Algorithm Method
• Step 1 [Initialization]
— L0(n) = ¥, for all n ¹ s
— Lh(s) = 0, for all h
• Step 2 [Update]
— For each successive h ³ 0
• For each n ≠ s, compute
Lh+1 (n) = min[Lh ( j) + w( j, n)]
j
— Connect n with predecessor node j that achieves minimum
— Eliminate any connection of n with different predecessor node
formed during an earlier iteration
— Path from s to n terminates with link from j to n
CS420/520 Axel Krings Page 31 Sequence 18 Chap 19
Example of Bellman-Ford
Algorithm
CS420/520 Axel Krings Page 32 Sequence 18 Chap 19
16
Results of Bellman-Ford
Example
CS420/520 Axel Krings Page 33 Sequence 18 Chap 19
Comparison
• Bellman-Ford • Dijkstra
— Calculation for node n needs — Each node needs complete
link cost to neighboring nodes topology
plus total cost to each — Must know link costs of all
neighbor from s links in network
— Each node can maintain set of — Must exchange information
costs and paths for every with all other nodes
other node
— Can exchange information with
direct neighbors
— Can update costs and paths
based on information from
neighbors and knowledge of
link costs
CS420/520 Axel Krings Page 34 Sequence 18 Chap 19
17
Evaluation
Dependent on
••Processing Implementation Both converge
time of specific under static
algorithms topology and
••Amount of costs
information
required from
other nodes If link costs
change,
algorithms
attempt to
If link costs depend catch up
on traffic, which Both
depends on routes converge
chosen, may have to same
feedback instability solution
CS420/520 Axel Krings Page 35 Sequence 18 Chap 19
Summary
• Routing in packet- • Internet routing
switching networks protocols
—Autonomous systems
—Characteristics
—Approaches to routing
—Routing strategies
—Border gateway
• Examples: Routing in protocol
ARPANET —OSPF protocol
—First generation:
Distance Vector Routing • Least-cost algorithms
—Dijkstra’s algorithm
—Second generation:
Link-State Routing —Bellman-Ford
—Third generation algorithm
—Comparison
CS420/520 Axel Krings Page 36 Sequence 18 Chap 19
18
