Download Slide 1

Net-Centric Computing Division
Department of Computer Science
Bogor Agricultural University
KOM 312
KOMUNIKASI DATA
DAN JARINGAN KOMPUTER
Packet Switching Network
Sri Wahjuni
my_juni04(at)ipb.ac.id
AGENDA
 Network-layer
swj/11
Functions
 Packet Network Topology
 Network Service Model
 Routing Principles
 Shortest Path Routing
 QoS in Network Layer
2
Department of Computer Science -IPB
REVIEW
OSI Reference Model
 TCP/IP Architecture

swj/11
3
Department of Computer Science -IPB
TWO VIEWS OF NETWORKS
External view: what services provided to transport layer
by a network (i.e. network layer)
 Need connection setup or not
 What QoSs are provided
 Services should be independent from underlying
networks so that transport layer can run over any
networks as long as the networks provide the services
 Internal view:
 physical topology,
 datagram message transfer or virtual circuit
information transfer,
 addressing and routing, congestion control.
4

swj/11
Department of Computer Science -IPB
COMPARISONS OF TWO VIEWS BY EXAMPLES
Broadcast networks and packet-switched networks
 From external view, both networks provides transfer
of information between users, not too much different
 But from internal view, very different:
 A broadcast network (such as LANs) is small,
addressing is simple, frame transferred in one hop
so no routing is needed
 In a packet-switching network, addressing must
accommodate large-scale networks and routing is
necessary.

swj/11
5
Department of Computer Science -IPB
PACKET SWITCHING
swj/11
Transfer of information as payload in data packets
 Packets undergo random delays & possible loss
 Different applications impose differing requirements
on the transfer of information

Department of Computer Science -IPB
6
NETWORK SERVICE
swj/11



Network layer can offer a variety of services to transport layer
Connection-oriented service or connectionless service
Best-effort or delay/loss guarantees
Department of Computer Science -IPB
7
NETWORK-LAYER FUNCTIONS
8
Department of Computer Science -IPB
swj/11
Essential
 Routing: mechanisms for determining the set of best
paths for routing packets requires the collaboration of
network elements
 Forwarding: transfer of packets from NE inputs to
outputs
 Priority & Scheduling: determining order of packet
transmission in each NE
Optional: congestion control, segmentation &
reassembly, security
AGENDA
 Network-layer
swj/11
Functions
 Packet Network Topology
 Network Service Model
 Routing Principles
 Shortest Path Routing
 QoS in Network Layer
9
Department of Computer Science -IPB
ACCESS MULTIPLEXER
swj/11
10
Department of Computer Science -IPB
LOCAL AREA NETWORK
swj/11
LAN 1
Bridge
LAN 2
11
Department of Computer Science -IPB
INTRADOMAIN AND INTERDOMAIN
swj/11
12
Department of Computer Science -IPB
KEY ROLE OF ROUTING
swj/11
How to get packet from here to there?
 Decentralized nature of Internet makes routing a
major challenge
Interior gateway protocols (IGPs) are used to
determine routes within a domain
 Exterior gateway protocols (EGPs) are used to
determine routes across domains
 Routes must be consistent & produce stable flows


Scalability required to accommodate growth

Hierarchical structure of IP addresses essential to
keeping size of routing tables manageable
13
Department of Computer Science -IPB
AGENDA
 Network-layer
swj/11
Functions
 Packet Network Topology
 Network Service Model
 Routing Principles
 Shortest Path Routing
 QoS in Network Layer
14
Department of Computer Science -IPB
MESSAGE SWITCHING
swj/11
15
Department of Computer Science -IPB
MESSAGE SWITCHING DELAY
Source
t
t

Switch 2
swj/11
Switch 1
T
t
t
Destination
Delay
Minimum Delay = 3 + 3T,
: propagation delay (umumnya panjang hop/kecepatan cahaya)
T: message transmission time
16
Department of Computer Science -IPB
DATAGRAM PACKET SWITCHING





Department of Computer Science -IPB
swj/11

Messages broken into
smaller units (packets)
Source & destination
addresses in packet header
Connectionless, packets
routed independently
Packet may arrive out of
order
Pipelining of packets across
network can reduce delay,
increase throughput
Lower delay that message
switching, suitable for
interactive traffic
17
DATAGRAM PACKET SWITCHING DELAY
Assume three packets corresponding to one message
traverse same path
T
swj/11
τ
Minimum
Delay = 3τ + 5(T/3) (single path assumed)
Additional queuing delays possible at each link
Packet pipelining enables message to arrive sooner
18
Department of Computer Science -IPB
DELAY FOR K-PACKET MESSAGE OVER L HOPS
Source
t
1
2
3
1
2
3
t
1
2
3 hops
3
t
L hops
3 + 2(T/3) first bit received
L + (L-1)P first bit received
3 + 3(T/3) first bit released
L + LP first bit released
3 + 5 (T/3) last bit released
Department of Computer Science -IPB
swj/11
t
L + LP + (k-1)P last bit released
where T = k P ; P disebut delay transmisi
19
DATAGRAM VS MESSAGE SWITCHING
Delay in message switching : L + L(kP)
 Delay in packet switching (k packets): L + LP +
(k-1)P
 Additional delay in message switching : P(k-1)(L1)

swj/11
20
Department of Computer Science -IPB
ROUTING TABLES IN CONNECTIONLESS
PACKET SWITCHING




Department of Computer Science -IPB
swj/11

Route determined by table
lookup
Routing decision involves
finding next hop in route
to given destination
Routing table has an entry
for each destination
specifying output port that
leads to next hop
Size of table becomes
impractical for very large
number of destinations
21
EXAMPLE: INTERNET ROUTING

Internet protocol uses datagram packet switching
across networks

Hosts have two-part IP address:


Networks are treated as data links
swj/11

Network address + Host address
Routers do table lookup on network address

This reduces size of routing table
22
Department of Computer Science -IPB
VIRTUAL-CIRCUIT PACKET SWITCHING






swj/11

Network-layer connection-oriented
service
Call set-up phase sets ups pointers
in fixed path along network
All packets for a connection follow
the same path
Abbreviated header identifies
connection on each link
Packets queue for transmission
Variable bit rates possible,
negotiated during call set-up
Delays variable, cannot be less
than circuit switching
23
Department of Computer Science -IPB
CONNECTION SETUP
Department of Computer Science -IPB
swj/11
Signaling messages propagate as route is selected
 Signaling messages identify connection and setup
tables in switches
 Typically a connection is identified by a local tag,
Virtual Circuit Identifier (VCI)
 Each switch only needs to know how to relate an
incoming tag in one input to an outgoing tag in the
corresponding output
 Once tables are setup, packets can flow along path

24
CONNECTION SETUP DELAY
swj/11
Connection setup delay is incurred before any packet
can be transferred
 Delay is acceptable for sustained transfer of large
number of packets
 This delay may be unacceptably high if only a few
packets are being transferred
25

Department of Computer Science -IPB
VC IMPLEMENTATION
A VC consists of:
1.
3.


swj/11
2.
Path from source to destination
VC numbers, one number for each link along path
Entries in forwarding tables in routers along path
Packet belonging to VC carries a VC number.
VC number must be changed on each link.

New VC number comes from forwarding table
26
Department of Computer Science -IPB
VC FORWARDING TABLE
swj/11
􀁺 Each input port of packet
switch has a forwarding
table
􀁺 Lookup entry for VCI of
incoming packet
􀁺 Determine output port (next
hop) and insert VCI for
next link
􀁺 Very high speeds are
possible
􀁺 Table can also include
priority or other information
about how packet should
be treated
27
Department of Computer Science -IPB
AGENDA
 Network-layer




swj/11
Functions
 Packet Network Topology
 Network Service Model
 Routing Principles
Routing algorithm
Routing tables
Hierarchical routing
Flooding and Deflection Routing
 Shortest
Path Routing
 QoS in Network Layer
Department of Computer Science -IPB
28
ROUTING ALGORITHM REQUIREMENTS

Responsiveness to changes
Topology or bandwidth changes, congestion
 Rapid convergence of routers to consistent set of routes
 Freedom from persistent loops

Optimality


Resource utilization, path length
Robustness


swj/11

Continues working under high load, congestion, faults,
equipment failures, incorrect implementations
Simplicity

Efficient software implementation, reasonable processing
29
load
Department of Computer Science -IPB
ROUTING ALGORITHM CLASSIFICATION
Centralized vs Distributed Routing
 Centralized Routing
All routes determined by a central node
 All state information sent to central node
 Problems adapting to frequent topology changes
 Does not scale

swj/11

Distributed Routing
Routes determined by routers using distributed algorithm
 State information exchanged by routers
 Adapts to topology and other changes
 Better scalability

30
Department of Computer Science -IPB
ROUTING ALGORITHM CLASSIFICATION -2
Set up manually, do not change; requires administration
 Works when traffic predictable & network is simple
 Used to override some routes set by dynamic algorithm
 Used to provide default router


swj/11
Static vs Dynamic Routing
 Static Routing
Dynamic Routing
Adapt to changes in network conditions
 Automated
 Calculates routes based on received updated network
state information

Department of Computer Science -IPB
31
ROUTING TABLES IN DATAGRAM
swj/11
32
Department of Computer Science -IPB
ROUTING TABLE IN VIRTUAL-CIRCUIT
swj/11
• Route determined during connection setup
• Tables in switches implement forwarding that realizes
33
selected route
Department of Computer Science -IPB
ROUTING TABLES IN VC - CONT
swj/11
34
Department of Computer Science -IPB
NON-HIERARCHICAL ADDRESSES AND
ROUTING
swj/11
No relationship between addresses & routing
proximity
 Routing tables require 16 entries each

Department of Computer Science -IPB
35
HIERARCHICAL ADDRESSES AND
ROUTING
swj/11
Prefix indicates network where host is attached
 Routing tables require 4 entries each

Department of Computer Science -IPB
36
FLOODING

Send a packet to all nodes in a network
No routing tables available
 Need to broadcast packet to all nodes (e.g. to
propagate link state information)

swj/11

Approach
Send packet on all ports except one where it arrived
 Exponential growth in packet transmissions

37
Department of Computer Science -IPB
swj/11
Flooding is initiated from Node 1: Hop 1 transmissions
Department of Computer Science -IPB
38
swj/11
39
Flooding is initiated from Node 1: Hop 2 transmissions
Department of Computer Science -IPB
swj/11

Flooding is initiated from Node 1: Hop 3 transmissions
Department of Computer Science -IPB
40
LIMITED FLOODING
Time-to-Live field in each packet limits number
of hops to certain diameter
 Each switch adds its ID before flooding; discards
repeats
 Source puts sequence number in each packet;
switches records source address and sequence
number and discards repeats

swj/11
41
Department of Computer Science -IPB
DEFLECTION ROUTING
Network nodes forward packets to preferred port
 If preferred port busy, deflect packet to another
port
 Works well with regular topologies






swj/11

Manhattan street network
Rectangular array of nodes
Nodes designated (i,j)
Rows alternate as one-way streets
Columns alternate as one-way avenues
Bufferless operation is possible
Proposed for optical packet networks
 All-optical buffering currently not viable
Department of Computer Science -IPB

42
EXAMPLE: NODE (0,2)→(1,0)
swj/11
Tunnel from last column
to first column or vice
versa
43
Department of Computer Science -IPB
EXAMPLE: BUSY NODE
swj/11
44
Department of Computer Science -IPB
AGENDA
 Network-layer
swj/11
Functions
 Packet Network Topology
 Network Service Model
 Routing Principles
 Shortest Path Routing
 QoS in Network Layer
45
Department of Computer Science -IPB
SHORTEST PATHS & ROUTING
Many possible paths connect any given source and
to any given destination
 Routing involves the selection of the path to be used
to accomplish a given transfer
 Typically it is possible to attach a cost or distance to
a link connecting two nodes
 Routing can then be posed as a shortest path
problem

swj/11
46
Department of Computer Science -IPB
ROUTING METRICS





swj/11
Means for measuring desirability of a path
 Path Length = sum of costs or distances
 Possible metrics
Hop count: rough measure of resources used
Delay: sum of delays along path; complex & dynamic
Bandwidth: “available capacity” in a path
Load: Link & router utilization along path (congestion)
Cost: $$$
47
Department of Computer Science -IPB
SHORTEST PATH APPROACHES
Department of Computer Science -IPB
swj/11
Distance Vector Protocols
 Neighbors exchange list of distances to destinations
 Best next-hop determined for each destination
 Ford-Fulkerson (distributed) shortest path
algorithm
Link State Protocols
 Link state information flooded to all routers
 Routers have complete topology information
 Shortest path (& hence next hop) calculated
 Dijkstra (centralized) shortest path algorithm
48
DISTANCE VECTOR ALGORITHM
Iterative, asynchronous:
swj/11
each local iteration caused by:
 local link cost change
 DV update message from
neighbor
Each node:
wait for (change in local link
cost of msg from neighbor)
Distributed:

each node notifies neighbors
only when its DV changes

neighbors then notify their
neighbors if necessary
recompute estimates
if DV to any dest has
changed, notify neighbors
49
Department of Computer Science -IPB
BELLMAN-FORD ALGORITHM


Now consider parallel computations for all destinations d
Initialization




Send Step


Send new distance vector to immediate neighbors across local link
Receive Step


swj/11

Each node has 1 row for each destination d
Distance of node d to itself is zero: Dd(d)=0
Distance of other node j to d is infinite: Dj(d)= ∝ , for j ≠ d
Next node nj = -1 since not yet defined
For each destination d, find the next hop that gives the minimum
distance to d,
 Minj { Cij+ Dj(d) }
 Replace old (nj, Di(d)) by new (nj*, Dj*(d)) if new next node or
distance found
Go to send step
Department of Computer Science -IPB
50
BELLMAN-FORD ALGORITHM(3)
swj/11
Table entry
@ node 1
for dest Node 6
Table entry
@ node 3
for dest Node 6
51
Department of Computer Science -IPB
BELLMAN-FORD ALGORITHM(4)
swj/11
52
Department of Computer Science -IPB
BELLMAN-FORD ALGORITHM(5)
swj/11
53
Department of Computer Science -IPB
BELLMAN-FORD ALGORITHM(6)
swj/11
54
Department of Computer Science -IPB
SHORTEST PATH TREE TO NODE 6
swj/11
55
Department of Computer Science -IPB
COMPARISON OF LS AND DV ALGORITHMS
Message complexity

LS:

DV:

exchange between neighbors only

convergence time varies
Speed of Convergence

LS:

O(n2) algorithm requires O(nE)
msgs
may have oscillations
 DV: convergence time varies
 may be routing loops
 count-to-infinity problem
Department of Computer Science -IPB

router malfunctions?
LS:


node can advertise
incorrect link cost
each node computes only
its own table
swj/11

with n nodes, E links, O(nE) msgs
sent
Robustness: what happens if
DV:


DV node can advertise
incorrect path cost
each node’s table used by
others

error propagate thru
network
59
EXAMPLE: INTRADOMAIN ROUTING
Also known as Interior Gateway Protocols (IGP)
 Most common Intra-AS routing protocols:

RIP: Routing Information Protocol

OSPF: Open Shortest Path First

IGRP: Interior Gateway Routing Protocol (Cisco
proprietary)
swj/11

60
Department of Computer Science -IPB
RIP ( ROUTING INFORMATION PROTOCOL)
Distance vector algorithm
 Distance metric: # of hops (max = 15 hops)
 Distance vectors: exchanged among neighbors every
30 sec via Response Message (also called
advertisement)
 Each advertisement: list of up to 25 destination nets
within AS

swj/11
61
Department of Computer Science -IPB
RIP: EXAMPLE
z
w
y
D
B
swj/11
A
x
C
Destination Network
Next Router
Num. of hops to dest.
w
y
z
x
A
B
B
--
2
2
7
1
….
….
....
Routing table in D
Department of Computer Science -IPB
62
RIP: EXAMPLE
Advertisement
from A to D
Next hops
- - C 4
… ...
w
z
x
A
swj/11
Dest
w
x
z
….
y
D
B
C
Destination Network
Next Router
Num. of hops to dest.
w
y
z
x
A
B
BA
--
2
2
75
1
….
….
....
Routing
table in D
Department of Computer Science
-IPB
63
RIP: LINK FAILURE AND RECOVERY
If no advertisement heard after 180 sec --> neighbor/link
declared dead




routes via neighbor invalidated
new advertisements sent to neighbors
neighbors in turn send out new advertisements (if tables
changed)
link failure info quickly propagates to entire net
poison reverse used to prevent ping-pong loops (infinite
distance = 16 hops)
swj/11

64
Department of Computer Science -IPB
OSPF (OPEN SHORTEST PATH FIRST)
“open”: publicly available
 Uses Link State algorithm



LS packet dissemination
Topology map at each node
Route computation using Dijkstra’s algorithm
swj/11

OSPF advertisement carries one entry per neighbor
router
 Advertisements disseminated to entire AS (via
flooding)

65
Department of Computer Science -IPB
INTERDOMAIN ROUTING:

BGP (Border Gateway Protocol): the de facto
standard
Path vector protocol instead of distance vector
swj/11

BGP
66
Department of Computer Science -IPB
AGENDA
 Network-layer
swj/11
Functions
 Packet Network Topology
 Network Service Model
 Routing Principles
 Shortest Path Routing
 QoS in Network Layer
67
Department of Computer Science -IPB
QOS PARAMETER

End-to-end delay :

Jitter


Variability in the packet delays
swj/11

sum of individual delays experienced at each system
Packet loss

Sum of packet discarded by he queue system
68
Department of Computer Science -IPB
QOS REQUIREMENTS
swj/11
69
Department of Computer Science -IPB
QUEUEING SYSTEM
swj/11
Queue scheduling :
- FIFO
- Priority (discard priority, HOL priority)
70
Department of Computer Science -IPB
(A) FIFO AND (B) FIFO W/ DISCARD PRIORITY
swj/11
71
Department of Computer Science -IPB
HEAD OF LINE (HOL) PRIORITY
swj/11
 High priority queue serviced until empty
 High priority queue has lower waiting time
Department of Computer Science -IPB
72
FAIR QUEUEING
Each flow has its own logical queue: prevents hogging;
allows differential loss probabilities
 C bits/sec allocated equally among non-empty queues

transmission rate = C / n(t), where n(t)=# non-empty queues
swj/11

Idealized system assumes fluid flow from queues
 Implementation requires approximation: simulate fluid
system; sort packets according to completion time in
ideal system

73
Department of Computer Science -IPB
RANDOM EARLY DETECTION (RED)
swj/11
Background :
 Packets produced by TCP will reduce input rate in
response to network congestion
 Early drop: discard packets before buffers are full
 Random drop causes some sources to reduce rate
before others, causing gradual reduction in aggregate
input rate
74
Department of Computer Science -IPB
RED ALGORITHM
swj/11
Algorithm:
 Maintain running average of queue length
 If Qavg < minthreshold, do nothing
 If Qavg > maxthreshold, drop packet
 If in between, drop packet according to probability
 Flows that send more packets are more likely to have
packets dropped
75
Department of Computer Science -IPB
QOS PROVISIONING

Open-loop control : does not rely on feedback
information to react to congestion

swj/11
Admission Control : common for virtual-circuit
 Policing (Leaky Bucket algorithm)
 Traffic Shaping

Closed-loop control : maximize link utilization
while preventing buffer overflows due to
congestion

Congestion control (in TCP)
76
Department of Computer Science -IPB
ADMISSION CONTROL
Flows negotiate contract
with network
 Specify requirements:

swj/11
Peak, Avg., Min bit rate
 Maximum burst size
 Delay, Loss requirement

o Network computes resources needed
o“Effective” bandwidth
o If flow accepted, network allocates resources to
ensure QoS delivered as long as source conforms to
contract
Department of Computer Science -IPB
77
POLICING
Network monitors traffic flows continuously to
ensure they meet their traffic contract
 When a packet violates the contract, network can
discard or tag the packet giving it lower priority
 If congestion occurs, tagged packets are discarded
first
 Leaky Bucket Algorithm is the most commonly used
policing mechanism

swj/11
Bucket has specified leak rate for average contracted
rate
 Bucket has specified depth to accommodate variations in
arrival rate
 Arriving packet is conforming if it does not result in
overflow
Department of Computer Science -IPB

78
LEAKY BUCKET PRINCIPLES
swj/11
Let X = bucket content at last conforming packet
arrival
 Let ta – last conforming packet arrival time =
depletion in bucket

Department of Computer Science -IPB
79
LEAKY BUCKET ALGORITHM
swj/11
80
Department of Computer Science -IPB
LEAKY BUCKET EXAMPLE
swj/11
81
Department of Computer Science -IPB
TRAFFIC SHAPING
swj/11
Networks police the incoming traffic flow (ingress)
 Traffic shaping is used to ensure that a packet
stream conforms to specific parameters (engress)
 Networks can shape their traffic prior to passing it
to another network

Department of Computer Science -IPB
82
LEAKY BUCKET TRAFFIC SHAPER
swj/11
Buffer incoming packets
 Play out periodically to conform to parameters
 Surges in arrivals are buffered & smoothed out
 Possible packet loss due to buffer overflow
 Too restrictive, since conforming traffic does not need
to be completely smooth

83
Department of Computer Science -IPB
TOKEN BUCKET TRAFFIC SHAPER
swj/11
Token rate regulates transfer of packets
 If sufficient tokens available, packets enter network
without delay
 K determines how much burstiness allowed into the
network

Department of Computer Science -IPB
84
TOKEN BUCKET SHAPING EFFECT
swj/11
The token bucket constrains the traffic from a
source to be limited to b + r t bits in an interval of
length t
Department of Computer Science -IPB
85
REFERENCES
Garcia A.L., Widjaja A. 2004. Networks
Communication : Fundamental Concepts and Key
Architectures 2nd ed. – Chapter 7. McGraw-Hill
Companies, Inc.
 Kurose J.F., Ross K.W. 2003. Computer
Networking : A Top-Down Approach Featuring Internet
2nd ed. – Chapter 4. Pearson Education.

swj/11
86
Department of Computer Science -IPB