The Network Layer

The Network Layer
Chapter 5
Computer Networks, Fifth Edition by Andrew Tanenbaum and David Wetherall, © Pearson Education-Prentice Hall, 2011
Network Layer Design Issues
• Store-and-forward packet switching

• Services provided to transport layer
• Implementation of connectionless service
• Implementation of connection-oriented service
• Comparison of virtual-circuit and datagram
networks
Store-and-Forward Packet Switching
ISP’s equipment
The environment of the network layer protocols.
Services Provided to the Transport Layer
1. Services independent of router technology.

2. Transport layer shielded from number, type,
topology of routers.
3. Network addresses available to transport
layer use uniform numbering plan
– even across LANs and WANs
Implementation of Connectionless Service
ISP’s equipment
A’s table (initially) A’s table (later) C’s Table E’s Table
Routing within a datagram network

Implementation of
Connection-Oriented Service
ISP’s equipment
A’s table C’s Table E’s Table
Routing within a virtual-circuit network

Comparison of Virtual-Circuit
and Datagram Networks
Comparison of datagram and virtual-circuit networks

Routing Algorithms (1)
• Optimality principle
• Shortest path algorithm
• Flooding
• Distance vector routing
• Link state routing
• Routing in ad hoc networks
Routing Algorithms (2)
• Broadcast routing
• Multicast routing
• Anycast routing
• Routing for mobile hosts
• Routing in ad hoc networks
Fairness vs. Efficiency
Network with a conflict between fairness and efficiency.
The Optimality Principle
(a) A network. (b) A sink tree for router B.
Shortest Path Algorithm (1)
The first five steps used in computing the shortest path from A
to D. The arrows indicate the working node
Distance Vector Routing
a) Completely decentralized
b) No node has complete information about the costs of all
network links
c) Gradual calculation of path by exchanging information with
neighbors
– Remember -- network as a graph ?
An Example
B A B C D E F G
C
A
D A 0 1 1 ∞ 1 1 ∞
E
B 1 0 1 ∞ ∞ ∞ ∞
F G
C 1 1 0 1 ∞ ∞ ∞
• Internal Information at each D ∞ ∞ 1 0 ∞ ∞ 1
node ----->
E 1 ∞ ∞ ∞ 0 ∞ ∞
F 1 ∞ ∞ ∞ ∞ 0 1
G ∞ ∞ ∞ 1 ∞ 1 0
Routing Tables
Cost Next
B
Hop
A
D
B 1 B
E
C 1 C
F G
a) With this
information, routing
D ∞ -
table at A is --> E 1 E
F 1 F
G ∞ -
Evolution of the table.
a) Each node sends a message to

neighbors with a list of
Cost Next
distances. Hop
b) F --> A with G is at a distance 1
c) C --> A with D at distance 1. B 1 B
C 1 C
B D 2 C
E 1 E
C
A
D
E F 1 F
F G G 2 F
Final Distance Matrix
B A B C D E F G
C
A
D A 0 1 1 2 1 1 2
E
B 1 0 1 2 2 2 3
F G
C 1 1 0 1 2 2 2
D 2 2 1 0 3 2 1
E 1 2 2 3 0 2 3
F 1 2 2 2 2 0 1
G 2 3 2 1 3 1 0
Routing updates
a) When are routing updates sent ?
1. Periodic updates
– Even if nothing has changed, send periodically. Main reason is to
let other nodes know that the sender is alive.
– Refresh information that might be needed if some of the routes
were to become unavailable.
2. Triggered updates
– When a node receives an update from one of its neighbors which
may lead to a change in its routing tables (could be due to
change in link cost).
o Note: typically order of periodicity is seconds to
several minutes.
Link/Node Failures
a) Nodes that first notice send new lists of distances to

neighbors.
b) How do they detect failures ?
– Route updates don’t arrive
– Probing with test packets.
Example Revisited
a) Let link from F to G fail.

b) F sets new distance to G to ∞;
sends update to A.
c) A was initially routing to G via F. So
it now sets link cost to G to ∞.
d) Next update from C; A learns that C
has 2 hop path to G.
e) A now can reach G in 3 hops via C.
B
f) A sends an update to F. Thus, F C
now, can reach G via A in 4 hops. A
D
F G
Count to Infinity
a) A discovers that link to E is lost. • The process

b) If before A’s message (saying that link continues and
cost to E is ∞) is received, if B or C
thus, the system
advertise that they can reach E in two
hops, then A can be confused. does not stabilize.
c) Another possibility, B gets A’s update • This is the count
followed by C’s update which says to infinity problem.
that E is reachable in 2 hops.
d) So B tells A this, and A thinks it can
now reach E via B in 3 hops. B
e) This information reaches C who now C
thinks that it can reach E in 4 hops via A
D
A.
E
F G
Split Horizon
a) One solution would be to approximate ∞ to say 16
hops.
b) With Split Horizon, when a node sends a routing table
update to its neighbors, it “does not” send those routes
it learned from “a particular” neighbor, back to that
neighbor.
– For example, B had E, 2, A. When it sends a route update to
A, it does not include this.
c) With split horizon with poison reverse, this update is
reported but the link weight is set to ∞.
– For example B sends (E, ∞) to A.
Does this work ?
a) Works only if there is a loop with only 2 nodes.

b) Typically, in static networks where link failures/node
failures are rare, this may be enough.
c) Speed of convergence is why, link state routing may
be preferable -- it takes a while before routes
converge.
Link State Routing
a) Also called shortest path first (SPF) forwarding

– Named after Dijkstra’s algorithm (1959) which it uses to compute
routes
b) All routers have tables which contain a representation of the
entire network topology
– In the form of lists of routers and information about each router’s
neighbours and the connection between the two
Link State Routing
a) Each router creates a link state packet (LSP) which

contains names (e.g. network addresses) and cost
to each of its neighbours
– The LSP is transmitted to all other routers, who each
update their own records
– When a routers receives LSPs from all routers, it can use
(collectively) that information to make topology-level
decisions
Link State Packets
a) LSPs are generated and distributed when:

– A time period passes
– New neighbours connect to the router
– The link cost of a neighbour has changed
– A link to a neighbour has failed (link failure)
– A neighbour has failed (node failure)
Link State Packets
a) LSP are essentially a list of tuples, containing:

– The name of a neighbour to a router
• Which may be a router or a network
– The cost of the link to that neighbour
Link State Packets
a) Distribution of LSPs can be difficult

– Routers themselves are the means for delivering messages
– How do routers deliver their own messages, particularly when routers
are in an inconsistent state
• e.g. During link failure, before each router has been notified
of the problem
Dijkstra’s LSR Algorithm
a) Consider the following network:
6 2
A B C 5
2 1 2
G
2 4
D E F 1
Link state database:
A B C D E F G
B 6 A 6 B 6 A 2 B 1 C 2 C 5
D 2 C 2 F 2 E 2 D 2 E 4 F 1
E 1 G 5 F 4 G 1
a) Now, if we want to generate a PATH for
C:
– First, we add (C,0,0) to PATH
C (0)
a) Examine C’s LSP
– Add F, G, and B to TENT
C (0)
(2) (5) (2)

F G B
a) Place F in PATH (shown as solid line)
– Add G and E to TENT (adding costs)
C (0)
(2) (5) (2)

F G B
(3) (6)
G
E
a) G exists in TENT twice, keep only the
best
– The new G is a better path than the old (3 <
5)
C (0)
(2) (5) (2)

F G B
(3) (6)
G
E
a) Put B into path (shown as solid line)
– Add A and E to TENT
C (0)
(2) (2)
F B
(3) (6) (3) (8)

G A
E E
a) E exists in TENT twice, keep only the
best
– The new E is better than the old (3 < 6)
C (0)
(2) (2)
F B
(3) (6) (3) (8)

G A
E E
a) Place E in PATH (shown as solid line)
– Add D to TENT
C (0)
(2) (2)
F B
(3) (3) (8)

G A
E
(5)
D
a) Place G in PATH (shown as solid line)
– All G’s LSP elements already exist in TENT
C (0)
(2) (2)
F B
(3) (3) (8)

G A
E
(5)
D
a) Place D in PATH (shown as solid line)
– Add path to A since it is better than old A
C (0)
(2) (2)
F B
(3) (3) (8)

G A
E
(5)
D
(7)
A
a) Place A in PATH (shown as solid line)
– All A’s LSP elements already exist in PATH
C (0)
(2) (2)
F B
(3) (3)
G
E
(5)
D
(7)
A
a) We are done since all routes from TENT
were placed into PATH
C (0)
(2) (2)
F B
(3) (3)
G
E
(5)
D
(7)
A
a) We can now create a forwarding database:
Forwarding Database
C (0) Destination Port
(2) (2) C C
F B
F F
(3) (3)
G
E
G F
B B
(5) E B
D
D B
(7)
A A B
LSR Topology Changes
a) LSR forwarding tables must be recalculated
whenever a topology change occurs
– For example, a new router and/or link is added to the
network
• This new link may provide a more efficient route to one or more other
nodes
– For example, a given link’s cost is reduced
• This new link may now provide the lowest total cost route to a destination
that was previously forwarded in another direction
– For example, a given link’s cost is increased
• This new link may no longer provide the lowest total cost route to a given
destination, and another route should now be chosen
LSR Topology Changes
a) In a nutshell, LSR routers should invalidate (indicate that it

needs to be regenerated) its PATH data structure, and thus its
forwarding table
– The entire PATH generation algorithm (e.g. Dijkstra’s algorithm) should
be reapplied
Topology Change Example
a) Let’s consider our previously generated PATH
structure for the router C
C (0)
(2) (2)
F B
(3) (3)
G
E
(5)
D
(7)
A
a) Say we receive an LSP from router B, indicating the
link cost from B to E is now 6
C (0)
(2) (2)
F B
(3) (3)
G
E
(5)
D
(7)
A
a) The total route costs are different in PATH:
C (0)
(2) (2)
F B
(3) (8)
G
E
(10)
D
(12)
A
a) Consider for now, only the cost to A
C (0)
(2) (2)
F B
(3) (8)
G
E
(10)
D
(12)
A
a) Recall that another path to A existed
b) Now, that path is more efficient
C (0)
(2) (2)
F B
(3) (8) (8)

G A
E
(10)
D
(12)
A
a) The PATH data structure is complete, the
forwarding table can now be regenerated
C (0)
(2) (2)
F B
(3) (8) (8)

G A
E
(10)
D
a) In a router, which will be running as a

computer program, finding if a new path
exists essentially requires complete re-
execution of Dijkstra’s algorithm
b) For example, there could have been many
routes to A, each of which would have to be
compared to find the most efficient route
Hierarchical Routing
Hierarchical routing.
Broadcast Routing
Reverse path forwarding. (a) A network. (b) A sink tree.

(c) The tree built by reverse path forwarding.
Multicast Routing (1)
(a) A network. (b) A spanning tree for the leftmost router. (c) A
multicast tree for group 1. (d) A multicast tree for group 2.
Multicast Routing (2)
(a) Core-based tree for group 1.

(b) Sending to group 1.
Anycast Routing
(a) Anycast routes to group 1.

(b) Topology seen by the routing protocol.
Routing for Mobile Hosts
Packet routing for mobile hosts
Routing in Ad Hoc Networks
The shaded nodes are

(a) Range of A’s broadcast. new recipients. The
(b) After B and D receive it. dashed lines show
(c) After C, F, and G receive it. possible reverse routes.
(d) After E, H, and I receive it. The solid lines show the
discovered route.
Congestion Control Algorithms (1)
• Approaches to congestion control

• Traffic-aware routing
• Admission control
• Traffic throttling
• Load shedding
Congestion Control Algorithms (2)
When too much traffic is offered, congestion sets in and

performance degrades sharply.
Approaches to Congestion Control
Timescales of approaches to congestion control
Traffic-Aware Routing
A network in which the East and West parts

are connected by two links.
Traffic Throttling (1)
(a) A congested network. (b) The portion of the network that is

not congested. A virtual circuit from A to B is also shown.
Traffic Throttling (2)
Explicit congestion notification
Load Shedding (1)
A choke packet that affects only the source..

Load Shedding (2)
A choke packet that affects each hop it passes through.

Quality of Service
• Application requirements
• Traffic shaping
• Packet scheduling
• Admission control
• Integrated services
• Differentiated services
Application Requirements (1)
How stringent the quality-of-service requirements are.
Categories of QoS and Examples
1. Constant bit rate

• Telephony
2. Real-time variable bit rate
• Compressed videoconferencing
3. Non-real-time variable bit rate
• Watching a movie on demand
4. Available bit rate
• File transfer
Traffic Shaping (1)
(a) Shaping packets. (b) A leaky bucket. (c) A token bucket
Traffic Shaping (2)
(a) Traffic from a host. Output shaped by a token bucket of rate

200 Mbps and capacity (b) 9600 KB, (c) 0 KB.
Traffic Shaping (3)
Token bucket level for shaping with rate 200 Mbps and capacity
(d) 16000 KB, (e) 9600 KB, and (f) 0KB..
Packet Scheduling (1)
Kinds of resources can potentially be

reserved for different flows:
1. Bandwidth.
2. Buffer space.
3. CPU cycles.
Round-robin Fair Queuing
(a) Weighted Fair Queueing.

(b) Finishing times for the packets.
Admission Control (1)
An example flow specification
Admission Control (2)
Bandwidth and delay guarantees with token buckets and WFQ.
Integrated Services (1)
(a) A network. (b) The multicast spanning tree for host 1.

(c) The multicast spanning tree for host 2.
Integrated Services (2)
(a) Host 3 requests a channel to host 1. (b) Host 3 then

requests a second channel, to host 2.
(c) Host 5 requests a channel to host 1.
Differentiated Services (1)
Expedited packets experience a traffic-free network
Differentiated Services (2)
A possible implementation of assured forwarding
Internetworking
• How networks differ

• How networks can be connected
• Tunneling
• Internetwork routing
• Packet fragmentation
How Networks Differ
Some of the many ways networks can differ
How Networks Can Be Connected
(a) A packet crossing different networks.

(b) Network and link layer protocol processing.
Tunneling (1)
Tunneling a packet from Paris to London.
Tunneling (2)
Tunneling a car from France to England
Packet Fragmentation (1)
Packet size issues:
1. Hardware
2. Operating system
3. Protocols
4. Compliance with (inter)national standard.
5. Reduce error-induced retransmissions
6. Prevent packet occupying channel too long.
(a) Transparent fragmentation.

(b) Nontransparent fragmentation
Fragmentation when the elementary data size is 1 byte.

(a) Original packet, containing 10 data bytes.
Fragmentation when the elementary data size is 1 byte

(b) Fragments after passing through a network
with maximum packet size of 8 payload bytes plus header.
Fragmentation when the elementary data size is 1 byte

(c) Fragments after passing through a size 5 gateway.
Path MTU Discovery
The Network Layer Principles (1)
1. Make sure it works

2. Keep it simple
3. Make clear choices
4. Exploit modularity
5. Expect heterogeneity
...
The Network Layer Principles (2)
...
6. Avoid static options and parameters
7. Look for good design (not perfect)
8. Strict sending, tolerant receiving
9. Think about scalability
10. Consider performance and cost
The Network Layer in the Internet (1)
• The IP Version 4 Protocol

• IP Addresses
• IP Version 6
• Internet Control Protocols
• Label Switching and MPLS
• OSPF—An Interior Gateway Routing Protocol
• BGP—The Exterior Gateway Routing Protocol
• Internet Multicasting
• Mobile IP
The Network Layer in the Internet (2)
The Internet is an interconnected collection of many networks.
The IP Version 4 Protocol (1)
The IPv4 (Internet Protocol) header.
The IP Version 4 Protocol (2)
Some of the IP options.
IP Addresses (6)
IP address formats
IP Addresses (7)
Special IP addresses
IP Addresses (1)
An IP prefix.
IP Addresses (2)
Splitting an IP prefix into separate networks with subnetting.
IP Addresses (3)
A set of IP address assignments
IP Addresses (4)
Aggregation of IP prefixes
IP Addresses (5)
Longest matching prefix routing at the New York router.
IP Addresses (8)
Placement and operation of a NAT box.
IP Version 6 Goals
• Support billions of hosts
• Reduce routing table size
• Simplify protocol
• Better security
• Attention to type of service
• Aid multicasting
• Roaming host without changing address
• Allow future protocol evolution
• Permit coexistence of old, new protocols. . .
IP Version 6 (1)
The IPv6 fixed header (required).
IP Version 6 (2)
IPv6 extension headers
IP Version 6 (3)
The hop-by-hop extension header for

large datagrams (jumbograms).
IP Version 6 (4)
The extension header for routing.
Internet Control Protocols (1)
The principal ICMP message types.
Internet Control Protocols (2)
Two switched Ethernet LANs joined by a router
Label Switching and MPLS (1)
Transmitting a TCP segment using IP, MPLS, and PPP.
Label Switching and MPLS (2)
Forwarding an IP packet through an MPLS network
OSPF—An Interior Gateway
Routing Protocol (1)
An autonomous system
A graph representation of the previous slide.
The relation between ASes, backbones, and areas in OSPF.
The five types of OSPF messages
BGP—The Exterior Gateway
Examples of routing constraints:
1. No commercial traffic for educat. network

2. Never put Iraq on route starting at Pentagon
3. Choose cheaper network
4. Choose better performing network
5. Don’t go from Apple to Google to Apple
Routing policies between four Autonomous Systems
Propagation of BGP route advertisements
Mobile IP
Goals
1. Mobile host use home IP address anywhere.
2. No software changes to fixed hosts
3. No changes to router software, tables
4. Packets for mobile hosts – restrict detours
5. No overhead for mobile host at home.
End
Chapter 5

The Network Layer

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

The Network Layer

Caricato da

Copyright:

Formati disponibili

The Network Layer

• Store-and-forward packet switching

The environment of the network layer protocols.

1. Services independent of router technology.

Routing within a datagram network

A’s table C’s Table E’s Table

Routing within a virtual-circuit network

Comparison of datagram and virtual-circuit networks

Network with a conflict between fairness and efficiency.

(a) A network. (b) A sink tree for router B.

a) Each node sends a message to

a) Nodes that first notice send new lists of distances to

a) Let link from F to G fail.

a) A discovers that link to E is lost. • The process

a) Works only if there is a loop with only 2 nodes.

a) Also called shortest path first (SPF) forwarding

a) Each router creates a link state packet (LSP) which

a) LSPs are generated and distributed when:

a) LSP are essentially a list of tuples, containing:

a) Distribution of LSPs can be difficult

Link state database:

(2) (5) (2)

(2) (5) (2)

(2) (5) (2)

(3) (6) (3) (8)

(3) (6) (3) (8)

(3) (3) (8)

(3) (3) (8)

(3) (3) (8)

a) In a nutshell, LSR routers should invalidate (indicate that it

(3) (8) (8)

(3) (8) (8)

a) In a router, which will be running as a

Reverse path forwarding. (a) A network. (b) A sink tree.

(a) Core-based tree for group 1.

(a) Anycast routes to group 1.

Packet routing for mobile hosts

The shaded nodes are

• Approaches to congestion control

When too much traffic is offered, congestion sets in and

Timescales of approaches to congestion control

A network in which the East and West parts

(a) A congested network. (b) The portion of the network that is

Explicit congestion notification

A choke packet that affects only the source..

A choke packet that affects each hop it passes through.

How stringent the quality-of-service requirements are.

1. Constant bit rate

(a) Shaping packets. (b) A leaky bucket. (c) A token bucket

(a) Traffic from a host. Output shaped by a token bucket of rate

Kinds of resources can potentially be

Round-robin Fair Queuing

(a) Weighted Fair Queueing.

An example flow specification

Bandwidth and delay guarantees with token buckets and WFQ.

(a) A network. (b) The multicast spanning tree for host 1.

(a) Host 3 requests a channel to host 1. (b) Host 3 then

Expedited packets experience a traffic-free network

A possible implementation of assured forwarding

• How networks differ

Some of the many ways networks can differ

(a) A packet crossing different networks.

Tunneling a packet from Paris to London.