




























































































Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
The basics of routing algorithms in telecommunications networks, focusing on distance vector and link-state methods. Key concepts, features, and differences between the two. Distance vector algorithms, such as rip and ospf, allow routers to determine the best path to a destination by exchanging distance vectors with neighbors. Link-state algorithms, like ospf and is-is, enable routers to have a complete view of the network topology and calculate the shortest path to each destination. The document also touches upon hierarchical routing, multicast routing, and routing with policy constraints.
Typology: Slides
1 / 101
This page cannot be seen from the preview
Don't miss anything!





























































































An Engineering Approach to Computer Networking An Engineering Approach to Computer Networking
Process of finding a path from a source to every destination inProcess of finding a path from a source to every destination inthe networkthe network
n n
Suppose you want to connect to Antarctica from your desktopSuppose you want to connect to Antarctica from your desktop
u u
what route should you take? what route should you take?
uu
does a shorter route exist? does a shorter route exist?
uu
what if a link along the route goes down? what if a link along the route goes down?
u u
what if you’re on a mobile wireless link? what if you’re on a mobile wireless link?
n n
Routing deals with these types of issuesRouting deals with these types of issues
How to make correct local decisions?How to make correct local decisions?
u u
each router must know each router must know
something something about global state
n n
Global stateGlobal state
uu
inherently large inherently large
uu
dynamic dynamic
u u
hard to collect hard to collect
n n
A routing protocol must intelligently summarize relevant A routing protocol must intelligently summarize relevantinformationinformation
Minimize routing table spaceMinimize routing table space
u u
fast to look up fast to look up
u u
less to exchange less to exchange
nn
Minimize number and frequency of control messagesMinimize number and frequency of control messages
nn
Robustness: avoidRobustness: avoid
u u
black holes black holes
u u
loops loops
uu
oscillations oscillations
nn
Use optimal pathUse optimal path
Routing in telephone networksRouting in telephone networks
n n
Distance-vector routingDistance-vector routing
n n
Link-state routingLink-state routing
nn
Choosing link costsChoosing link costs
nn
Hierarchical routingHierarchical routing
n n
Internet routing protocolsInternet routing protocols
n n
Routing within a broadcast LANRouting within a broadcast LAN
nn
Multicast routingMulticast routing
nn
Routing with policy constraintsRouting with policy constraints
n n
Routing for mobile hostsRouting for mobile hosts
3-level hierarchy, with a fully-connected core3-level hierarchy, with a fully-connected core
nn
AT&T: 135 core switches with nearly 5 million circuitsAT&T: 135 core switches with nearly 5 million circuits
n n
LECs may connect to multiple coresLECs may connect to multiple cores
Stable loadStable load
u u
can predict pairwise load throughout the day can predict pairwise load throughout the day
u u
can choose optimal routes in advance can choose optimal routes in advance
nn
Extremely reliable switchesExtremely reliable switches
uu
downtime is less than a few minutes per year downtime is less than a few minutes per year
u u
can assume that a chosen route is available can assume that a chosen route is available
u u
can’t do this in the Internet can’t do this in the Internet
nn
Single organization controls entire coreSingle organization controls entire core
uu
can collect global statistics and implement global changes can collect global statistics and implement global changes
n n
Very highly connected networkVery highly connected network
n n
Connections require resources (but all need the same)Connections require resources (but all need the same)
Simplicity of routing a historical necessitySimplicity of routing a historical necessity
n n
But requiresBut requires
u u
reliability in every component reliability in every component
uu
logically fully-connected core logically fully-connected core
nn
Can we build an alternative that has same features as theCan we build an alternative that has same features as thetelephone network, but is cheaper because it uses moretelephone network, but is cheaper because it uses moresophisticated routing?sophisticated routing?
u u
Yes: that is one of the motivations for ATM Yes: that is one of the motivations for ATM
u u
But 80% of the cost is in the local loop But 80% of the cost is in the local loop
FF
not affected by changes in core routingnot affected by changes in core routing
uu
Moreover, many of the software systems assume topology Moreover, many of the software systems assume topology
F F
too expensive to change themtoo expensive to change them
Burst of activity can cause network to enter metastable stateBurst of activity can cause network to enter metastable state
u u
high blocking probability even with a low load high blocking probability even with a low load
nn
Removed by trunk reservationRemoved by trunk reservation
uu
prevents spilled traffic from taking over direct path prevents spilled traffic from taking over direct path
DNHR measures traffic once a weekDNHR measures traffic once a week
n n
TSMR updates measurements once an hour or soTSMR updates measurements once an hour or so
u u
only if it changes “significantly” only if it changes “significantly”
nn
List of alternative paths is more up to dateList of alternative paths is more up to date
Routing in telephone networksRouting in telephone networks
n n
Distance-vector routingDistance-vector routing
n n
Link-state routingLink-state routing
nn
Choosing link costsChoosing link costs
nn
Hierarchical routingHierarchical routing
n n
Internet routing protocolsInternet routing protocols
n n
Routing within a broadcast LANRouting within a broadcast LAN
nn
Multicast routingMulticast routing
nn
Routing with policy constraintsRouting with policy constraints
n n
Routing for mobile hostsRouting for mobile hosts
EnvironmentEnvironment
u u
links and routers unreliable links and routers unreliable
u u
alternative paths scarce alternative paths scarce
uu
traffic patterns can change rapidly traffic patterns can change rapidly
nn
Two key algorithmsTwo key algorithms
u u
distance vector distance vector
u u
link-state link-state
nn
Both assume router knowsBoth assume router knows
uu
address of each neighbor address of each neighbor
u u
cost of reaching each neighbor cost of reaching each neighbor
n n
Both allow a router to determine global routing information byBoth allow a router to determine global routing information bytalking to its neighborstalking to its neighbors
‘
‘
2
‘
Each node knows its true cost to its neighborsEach node knows its true cost to its neighbors
n n
This information is spread to its neighbors the first time it sendsThis information is spread to its neighbors the first time it sendsout its distance vectorout its distance vector
n n
Each subsequent dissemination spreads the truth one hopEach subsequent dissemination spreads the truth one hop
nn
Eventually, it is incorporated into routing table everywhere in theEventually, it is incorporated into routing table everywhere in thenetworknetwork
nn
Proof: Bellman and Ford, 1957Proof: Bellman and Ford, 1957