Advanced Computer

Networks
ECE 5713

Advanced Internetworking
Global Internetworking, Multicast and MPLS

Dr. Amir Qayyum

 The Big Picture
Midterm
exam was
here

You are here

2
Advanced Internetworking

Outline
Global Internetworking
Multicast Routing
Multi-protocol Label Switching
Routing among Mobile Devices

3
 Where we are - What we understand
• Concepts of networking and network programming
– Elements of networks: nodes and links
– Building a packet abstraction on a link
• Issues related to bits/frames sent on a single link
– How to detect transmission errors in a frame after
encoding and framing it
– How to simulate a reliable channel (sliding window)
– How to arbitrate access to shared media in any network
• Design issues of direct link networks
– Functionality of network adaptors
4
Where we are – What we understand
• How switches may provide indirect connectivity
– Different ways to move through a network (forwarding)
– Bridge approach to extending LAN concept
– Example of a real virtual circuit network (ATM)
– How switches are built and contention within switches
• How different networks work together
– Common connecting protocol (IP) and its service model
– Routing protocols in simple networks
• Next: large networks working together

5
 Internetworking
• Reading: Peterson and Davie, Chapter 4
• Global internets (scale)
• Dealing with global scale
– Virtual geography and addresses
– Hierarchical routing
– Future internetworking: IPv6
• Multicast traffic
• MPLS
• Routing among mobile devices
6
How Should Hierarchy be Defined ?
• Hierarchy reduces information requirements for
forwarding

7
 Internetworking
• Main problems addressed by internetworking
– Heterogeneity
– Scale
• Bridges scale to level of medium organization
• IP addresses most end-to-end heterogeneity issues
• Today: issues of global scale
• Next couple of lectures: IPv6 , IP multicast, MPLS

8
 Global Internets
• Evolution of Internet structure
• Virtual geographies: networks, domains,
and translations
• Problems of scale
• IP address hierarchy evolution
• Notion of autonomous systems (AS’s)
• Common intradomain routing protocols
• Standard interdomain routing protocol

9
 Diameter of the Internet
• Connect all nodes of a network in a big ring
– Network diameter will be half the ring circumference
• Adding more links bring down network diameter
– Diameter can be quarter of the ring with one link …

How many
links should
be added to
bring the
diameter
below some
threshold ?

10
Internet Structure in Recent Past 1990
Stanford NSFNET backbone
ISU
BARRNET MidNet
regional Westnet
regional
… regional
Berkeley
PARC UNL KU
NCAR UNM

UA

• Tree structure, centered around one backbone

• Interconnection of thousands of “organizations”
– End users and service providers
• Scalability of routing protocols in global Internet
– Network entities are administratively independent

11
 Internet Structure Today
• Multiple backbone service providers
• Arbitrary graph structure
• Different QoS in different backbones (commercial!)
Large corporation
“Consumer” ISP
Peering
point
Backbone service provider Peering
point
“ Consumer” ISP

Large corporation “Consumer”ISP

Small Small
corporation corporation

12
 Virtual Geographies
• Internet domain names (understood by people)
• IP network numbers (understood by routers)
• Internet domain name hierarchy
– ASCII strings separated by periods; typical formats are:
• host.domain.domain_type
– xyz.carepvtltd.com
• host.subdomain.domain.domain_type
– abc.cs.berkeley.edu
• host.domain.domain_type.country
– www.case.edu.pk

13
 Virtual Geographies and IP Address
• Routers use IP network numbers
IP
address Class Network Host

• Original intention:
– One network number per physical network
• Closer to real geography (than domain names)
• Some jumps, e.g.,
– Divisions of one company
– Organizations with network research testbeds

14
 Virtual Geographies and DNS
• Translation from domain name to IP address
– Domain to IP: domain name service (DNS)
– IP to domain: relies on host
• Not secure
• Use reverse lookup - domain to IP - for security
• DNS is a hierarchical, (mostly) distributed
application (more detail later)

15
 Routing in
Large Scale Networks

16
 Route Propagation
• Know a smarter router
– Hosts know local router
– Local routers know site routers
– Site routers know core router
– Core routers know everything
• Autonomous System (AS)
– Corresponds to an administrative domain
– Examples: university, company, backbone network
– Assign each AS a 16-bit number
• Two-level route propagation hierarchy
– Interior gateway protocol (each AS selects its own)
– Exterior gateway protocol (Internet-wide standard)

17
Notion of Autonomous Systems (AS)
• AS’s represent a third hierarchy
– Define routing domains
– Based on notion of autonomy of control
• Intradomain routing (within an AS)
– Performed using domain-specific algorithm
– Selected by domain administrators (autonomously)
– Allows heterogeneous interior gateway protocols
• Interdomain routing (between AS’s)
– Performed using standard global algorithm
– Homogeneous exterior gateway protocol

18
Intra-domain (Interior GW) Protocols
• Routing Information Protocol (RIP)
– From the early Internet, developed for XNS
– Part of Berkeley Software Distribution (BSD) Unix
– Distance-vector algorithm
– Based on hop count (infinity set to 16 hops)
• Open Shortest Path First (OSPF)
– Internet standard (RFC 2328), “open” means public
– Based on link-state algorithm
– Authenticates messages
– Load balances across links

19
Inter-domain (Exterior GW) Protocols
• Very complex and difficult
– Different metrics, security, large scale: 140K prefixes!
• Focuses on reachability rather than optimality
– Support for very flexible policies
• Exterior Gateway Protocol (EGP)
– Defined on Internet with tree structure
– Embodied (and enforced) tree structure
– Had to be replaced eventually
– Used distance-vector updates
– Replaced by Border Gateway Protocol (BGP)

20
 EGP Messages
• Neighbor acquisition
– One router requests that another be its peer
– Peers exchange reachability information
• Neighbor reachability
– One router periodically tests if the another is
still reachable
– Exchange HELLO/ACK messages
– Uses a k-out-of-n rule
• Routing updates
– Peers periodically exchange their routing
tables (distance-vector)
21
 BGP-4: Border Gateway Protocol
• Internet is an arbitrarily interconnected set of AS’s
• Two types of traffic
– Local: begins or ends within an AS
– Transit: moves through an AS
• Three types of AS’s
– Stub: one single connection to one other AS; carries
local traffic only
– Multihomed: connections to multiple other AS’s, but
refuses to carry transit traffic
– Transit: connections to multiple other AS’s and
designed to carry both transit and local traffic

22
 BGP-4: Borger Gateway Protocol
• Each AS has:
– One or more border gateways (routers) to handle
inter-AS traffic
– One or more BGP speakers that participate in routing
protocol: establish BGP sessions to exchange messages
• BGP speaker advertises:
– Names of networks within the AS
– Names of other reachable networks through the AS
(transit AS only)
– Full path information (intra-domain protocols use
heterogeneous metrics); path-vector routing
– Withdrawn routes/negative advertisements (cancel
previously advertised route)
23
 BGP Path-Vector Routing Example
• AS4 advertises 128.96 and 192.4.153 as local networks
• Speaker for AS2 advertises reachability to these networks
– Network 128.96 and 192.4.153 can be reached via AS4, and
network 192.4.32 and 192.4.3 via AS5
Customer P 128.96
(AS 4) 192.4.153
Regional provider A
(AS 2)
Customer Q 192.4.32
(AS 5) 192.4.3
Backbone network
(AS 1)
Customer R 192.12.69
(AS 6)
Regional provider B
(AS 3)
Customer S 192.4.54
(AS 7) 192.4.23

24
 BGP-4 - Details
• Full path in BGP messages to avoid loops
– Best route according to local policies is advertised
– No obligation to advertise route to known destinations
• 16 bit AS numbers are uniquely assigned
– Stub ASs do not need a unique AS number
• BGP-4 designed to support classless addresses
– Update messages contain prefix and its length (10.1/16)
• Update messages are reliably sent using TCP
– Occasional “keepalive” messages if nothing changes

25
Building Scalable Networks …
• Subdivided the routing problem into
manageable parts
– New level of hierarchy is introduced
• Complexity of interdomain routing:
– Order of number of ASs
• Complexity of interdomain routing:
– Ooder of networks in an AS

26
 Integrating Interdomain and
Intradomain Routing
• How routers in a domain get routing information ?
• In a stub AS with single border router
– Inject a default route in intradomain routing protocol
• In a domain with multiple border routers (any AS)
– Border routers inject specific routes learned from
outside, with some cost
• In backbone networks, too costly to inject too
many outside routes in intradomain protocol
– Use Interior BGP (IBGP) to redistribute outside routes

27
 Internetworking
• Basics of internetworking (heterogeneity)
– IP protocol, address resolution, control messages, …
• Routing
• Global internets (scale)
– Virtual geography and addresses
– Hierarchical routing
• Future internetworking: IPv6
• Multicast traffic
• MPLS

28
 IPv6 - History
• Next-generation IP, previously denoted IPng
• Intended to extend address space/routing limitations
– Requires header change: new protocol for all nodes!
– Try to include everything new with one change
• IETF (Internet Engg Task Force) solicited white papers
• Many suggestions, merged in time to several options
• IETF selected Simple Internet Protocol Plus (SIPP)
with some modifications, dubbed it version 6

29
 IPv6 - History
• Address space pressure reduced with NAT technique
– Few public addresses with Network Address Translation
– Driving force reduces and so the rapid transition!
– Will IPv6 succeed in completely replacing IPv4 ? ? ?

Local network
Internet
with many hosts
NAT Box

30
 IPv6 Wish List / Planned Support
• 128-bit addresses
• Multicast traffic
• Mobility
• Real-time traffic/quality of service guarantees
• Authentication and security
• Autoconfiguration: determining local IP address
• End-to-end fragmentation
• Protocol extensions (flexible protocol)
• Smooth transition spread over time is critical
– Islands of v6 and then v4; traffic is tunneled in Internet
31
 IPv6 Addresses
• Classless addressing/routing (similar to CIDR)
• Address notation
– String of eight 16-bit hex values separated by colons
e.g., 5CFA:0002:0000:0000:CF07:1234:5678:FFCD
– Only one set of contiguous zeroes can be elided, e.g.,
5CFA:0002::CF07:1234:5678:FFCD
• Address assignment
– Provider-based
– Geographic

010 region ID provider ID subscriber ID subnet host

32
 IPv6 Address - Prefix
• 0000 0000 reserved (includes transition addresses)
• 0000 001 ISO NSAP (Network Service Access
Point) allocation
• 0000 010 Novell IPX allocation
• 010 provider-based unicast
• 100 geographic unicast
• 1111 1110 10 link local addresses
• 1111 1110 11 site local addresses
• 1111 1111 multicast addresses
• Other unassigned

33
 IPv4 Packet Format Review
0 4 8 16 19 31
Version HLen TOS Length

Ident Flags Offset

TTL Protocol Checksum
SourceAddr

DestinationAddr

Options (variable) Pad

(variable)
Data

• 20-byte minimum
• Mandatory fields not always used (e.g.,
fragmentation)
• Options are unordered list of (name, value) pairs
34
 IPv6 Packet Format
0 4 8 16 24 31
version priority flow label
payload length next header hop limit
source address word 1
source address word 2
source address word 3
source address word 4
destination address word 1
destination address word 2
destination address word 3
destination address word 4
35
 IPv6 Packet Format
0 4 8 16 24 31
version priority flow label
payload length next header hop limit
source address (4 words)
destination address (4 words)
option (variable number, usually fixed length)
• 40-byte minimum
• Mandatory fields (probably) always used
• Strict order on options reduces processing time
(no need to parse irrelevant options)
36
 IPv6 Packet Format
• Priority and flow label
– Support service guarantees
– Allow “fair” bandwidth allocation
• Payload length (header not included, unlike IPv4)
• Next header
– Combines options and protocol
– Linked list of options
– Ends with higher-level protocol header (e.g., TCP)
• Hop limit is TTL field renamed to match usage

37
 IPv6 Extension Headers
• Extension headers (options) appear in order

1. Hop-by-hop options .. .. .. misc. info. for routers

2. Routing .. .. .. .. .. .. .. .. full/partial route to
follow
3. Fragmentation .. .. .. .. .. IP fragmentation info
4. Authentication .. .. .. .. .. sender identification
5. Encrypted security payload info. about contents
6. Destination options .. .. .. info. for destination
38
 IPv6 Extension Headers
• Hop-by-hop extension
– Length is in bytes beyond mandatory 8
0 8 16 24 31
next header length type
value

• Jumbogram option (packet longer than 65,535 bytes)

– Payload length in main header set to 0
0 8 16 24 31
next header 0 194 0
payload length in bytes
39
 IPv6 Extension Headers
• Routing extension
0 8 16 24 31
next header 0 #addresses next address
strict/loose routing bitmap
1 – 24 addresses
– Up to 24 “anycast” addresses target AS’s or providers
– Next address tracks current target
– Strict routing requires direct link; loose routing allows
intermediate nodes
40
 IPv6 Extension Headers
• Fragmentation extension
0 8 16 29 31
next header reserved offset resv M
identification
• Similar to IPv4 fragmentation
– 13-bit offset,
– Last-fragment mark (M)
• Larger fragment identification field

41
 IPv6 Extension Headers
• Security without bothering the routers
• Authentication extension
– Designed to be very flexible
– Includes Security Parameters Index (SPI) and
Authentication data at the end
• Encryption extension
– Called Encapsulation Security Payload (ESP)
– Includes an SPI
– All headers and data after ESP are encrypted

42
 IPv6 Design Controversies
Address length
• 8-byte
– Less header overhead; Might run out in a few decades
• 16-byte
– More overhead; Good for foreseeable future
• 20-byte
– Even more overhead; Compatible with OSI
• Variable-length
– Difficult for router design

43
 IPv6 Design Controversies
Hop limit
• 65,535
– 32-hop paths are common now
– In a decade, much longer paths possible
• 255
– Limits lost packet lifetime (65,535 is far too long)
– Good network design makes long paths unlikely
• Source to backbone
• Across backbone
• Backbone to destination

44
 IPv6 Design Controversies
Maximum packet size
• > 64kB: supercomputer/high bandwidth
applications
– Too much overhead to fragment data
– Want much larger packets
• 64 kB: longer packets incompatible with
low-bandwidth lines
– Consider 1MB packet across 1.5 Mbps line
– Ties up line for >5 seconds
– Inconveniences interactive users
45
 IPv6 Design Controversies
Keep IP checksum ?
• Yes
– Removing checksum from IP analogous to removing
brakes from a car
• Lighter, so you go faster
• Unprepared for the unexpected
• No
– Typically duplicated in data link and transport layers
– Major expense in IPv4 routers
– In case of IPv6, it is mandatory in UDP and TCP

46
 IPv6 Design Controversies
Mobile hosts
• Direct or indirect ?
– Reconnect directly using canonical address
– Use home and foreign agents to forward traffic
• Mobility introduces asymmetry
– Base station signal is strong, heard by mobile units
– Mobile unit signal is weak and susceptible to
interference, not heard by base station
• No clearly superior design proposal
47
 IPv6 Design Controversies
Security
• Where ?
– Network layer: a standard service
– Application layer
• No viable standard
• Applications susceptible to errors in
network implementations
• Too clunky to turn off
• How ?
– Political export/import issues
– Cryptographic strength issues

48
Multicast

49
 Internetworking
• Basics of internetworking (heterogeneity)
– IP protocol, address resolution, control messages …
• Routing
• Global internets (scale)
– Virtual geography and addresses
– Hierarchical routing
• Future internetworking: IPv6
• Multicast traffic
• MPLS

50
Internet Multicast Outline
• Motivation and challenges
• Support strategy
• IP multicast service model
• Multicast in the Internet
• Routing
– Review of ELAN techniques
– Multicast routing
• Limitations

51
 Multicast
• Unicast: one destination; Broadcast: all
destinations; Multicast: subset of destinations
• When is multicast useful ?
– Send data to multiple receivers at once
• Videoconferencing, video-on-demand,
telecollaboration
• Software update to group of customers
– Limited broadcast/self-defined multicast
• Send question to unknown receiver
• Resource discovery; Distributed database

52
 Multicast
• Why not just use broadcast/unicast ?
– Broadcast not supported outside of LAN
– Unicast sends multiple copies across common links
• Multicast support
– Often supported by hardware in LAN’s (as broadcast,
if not multicast)
– But difficult to extend in scalable manner
• Multicast challenges
– Efficient distribution on an internetwork
– Specification of recipient group (abstraction must
support self-definition)
53
 Multicast Support Strategy
• IPv4 used as basis for experimental solutions
– Use class D addresses (1110 <28 bits>)
– Demonstrated with MBone
– Uses tunneling
• Multicast integrated into IPv6
• Group management protocol (IGMP)
• Several routing/forwarding schemes:
– Distance-vector
– Link-state
– Protocol-independent

54
 IP Multicast Service Model
• Each group uses a single address
– Class D addresses (1110 <28 bits>)
– Some well-known, some dynamically assigned
• Group membership
– Members located anywhere in the Internet
– Number of receivers is arbitrary
– Members can join/leave dynamically
– Hosts can belong to more than one group

55
 IP Multicast Service Model
• Senders simply use group address as destination
– Sender need not be in group
– LAN loopback needed for sender in group
• Multicast scope
– LAN (local scope)
– Administrative scope (e.g., campus); may overlap; can
assign group addresses dynamically
– TTL scope (no more than N hops)
• Scope is exposed to protocols and applications
(by exposing IP TTL)
56
 IP Multicast Service Model
• Multicast reception requires membership in
group
– Internet Group Management Protocol (IGMP),
RFC 1112
– New operations to join and leave group
– LAN routers track local membership
– Forwarding depends on routing scheme
– Last hop typically uses LAN broadcast
• Packet reception same as IP unicast
57
Internet Multicast Backbone - MBone
• Existing infrastructure for multicast in the Internet
• Multicast route propagation using DVMRP
• Problem: most IP routers do not support multicast
• Solution: tunneling by multicast-capable routers
– Encapsulate multicast traffic in IP packets
– Send to other multicast-capable routers
– Recipients unpack and forward original multicast packet
• Passes through multicast-incapable areas of
Internet
58
 ELAN Multicast Techniques
• Direct support (Ethernet)
– Application subscribes to group
– IP layer notifies Ethernet card to listen to
packets with group address
• Support through broadcast (LANE)
• Flooding in ELANs
– Each packet sent on all but incoming link
– Switches must remember each packet!
• Spanning tree: every host gets one copy

59
 ELAN Multicast Techniques
• Spanning tree selection
– Elect a leader; spanning tree is shortest path to leader
(Perlman)
– Distribute topology everywhere, compute in parallel
(link-state)
• Problems with spanning trees
– Bandwidth wasted for groups with few receivers;
solution: prune LAN’s with no receivers from tree
– For very large ELAN’s, no single tree is efficient;
solution: define tree per group or tree per source
• The same solutions are used in the Internet!
60
 Spanning Tree Tradeoffs
• Tree per group or tree per source ?

• Per group advantage

– One routing entry per group
• Per source advantages
– More efficient distribution
– Spread load better across links
– Leverage unicast routing tables

61
Multicast Routing in the Internet

• Multicast Open Shortest Path First (MOSPF)

• Distance-Vector Multicast Routing Protocol
(DVMRP, used in MBONE)
• Protocol-independent multicast (PIM)
– Deals with scalability issues of above protocols
– Dense Mode (PIM-DM)
– Sparse Mode (PIM-SM)

62
 Multicast Routing in the Internet
• How do senders find receivers?
– Receivers inform all senders of interest (MOSPF)
– Send to all receivers; uninterested receivers prune
(DVMRP, PIM-DM)
– Agree on set of rendezvous points (PIM-SM)
• Types of distribution trees
– Separate tree from each sender (DVMRP, MOSPF,
PIM-DM, PIM-SM)
– Tree rooted at rendezvous point (PIM-SM)

63
 Link State Multicast (MOSPF)
• Each host on a LAN
– Periodically announces its group memberships,
via Internet Group Management Protocol (IGMP)
• Extend LSP to include set of groups with
members on a given LAN
• MOSPF routing extends OSPF
– Uses Dijkstra’s algorithm
– Computes shortest-path spanning tree for source-
group pairs
– Forward packet on local portion of tree

64
 Link State Multicast (MOSPF)
• Tree computation
– Can’t precompute for all source-group pairs
– Compute on demand when first packet from a
source S to a group G arrives
– Cache trees for active source-group pairs
– Recompute when link-state changes
• Scalability limitations
– Reasonable intra-AS scalability
– But meaningless for inter-AS
– Source-group pairs scale with sources (needs
to be hierarchical)
65
Distance Vector Multicast (DVMRP)
• Idea
– Graph of directed next-hop edges to a destination S
form a tree
– Use reverse edges to broadcast from S
• Implementation (reverse path broadcast, or RPB)
– Forward multicast packet on all links
– Iff packet came from next hop for packet source
• Avoid repetition on LAN’s
– Assign parent router for each LAN
– Has shortest path to source, ties broken by ID
– Track parenthood via vector exchanges

66
 RPB and RPM
M M G

RPM from S to G
Member of RPB from S
M multicast
group G S Unicast route to S
Pruned
67
RPB to RPM (reverse path multicast)
• Identify leaf networks
– Only one router on network
– Thus no distance packets received on interface
• Prune leaf networks
– Without hosts in a group
– Hosts must self-identify using IGMP
• Forward pruning information
– Extend distance vector with group information
– Forward packets only to interested parties
– Only when multicast source active
68
 Distance Vector Multicast
RPM Implementation
• Assume that everyone is interested
• Respond to unwanted packets with
prune requests
• Prune requests
– Canceled by graft request
– Time out periodically
• Need ARQ for prune or graft ?

69
Distance Vector Multicast - Scalability
• Packets are periodically broadcast (thus
guaranteed to reach all interested members)
• High overhead for sparse groups; consider
– Multicast group of 10 members
– Scattered around the world
– Packets periodically reach all routers in Internet
• High overhead for routers
– All off-tree routers maintain pruning state
– And periodically retransmit

70
Protocol Independent Multicast (PIM)
• Approach
– Define rendezvous points (RP) for each group
– Need multiple RP’s to handle failures
• Two versions
– Dense mode
• Explicit prune messages
• Shared tree
– Sparse mode
• Explicit join messages
• Shared or source-specific tree

71
Protocol Independent Multicast (PIM)
• Rendezvous points (RP) for each multicast
group

Specific
multicast
RP tree
RP
RP

72
Protocol Independent Multicast

• Joins
– Receiver: send packet to one RP
– Source: send to all RP’s
• Tree selection
– Rooted at rendezvous points
– Shared for infrequent traffic
– Source-specific if merited by traffic level

73
 Limitations on Multicast
• Scalability (addressed to some extent by PIM)
– Explosive growth of the Internet population
– Explosive growth of multicast, multimedia
applications
• Control of network resources
– Applications have different performance needs
– Different resource commitments by clients and/or
organizations
– Different ASs provide different QoS …

74
 Internetworking
• Basics of internetworking (heterogeneity)
– IP protocol, address resolution, control messages …
• Routing
• Global internets (scale)
– Virtual geography and addresses
– Hierarchical routing
• Future internetworking: IPv6
• Multicast traffic
• MPLS

75
 Multiprotocol Label Switching
• Combines properties of virtual circuits with
flexibility and robustness of datagrams
– Relies on IP addresses and IP routing protocols
– Forwards packets using short, fixed length labels with
local scope
• Marriage of two seemingly opposed technologies

76
 MPLS - Capabilities
• Enable IP capabilities on devices that do not have
the capability to forward IP datagrams
• Forward IP packets along explicit routes: routes
precalculated separate to IP routing protocol
• Support certain types of virtual private networks
(VPN) services
• Where is performance improvement ?
– Depends on factors other than header processing

77
 Destination Based Forwarding
• MPLS – attaching labels with IP datagrams
• Router allocates a label for each prefix in its
routing table

78
 Destination Based Forwarding
• Advertisement of label and prefixes to neighbors
via Label Distribution Protocol (LDP)
– Attach corresponding label to all packets sent to that
router for that prefix

79
 Destination Based Forwarding
• Other routers store these labels as remote labels
in their routing table along with prefixes

80
 Destination Based Forwarding
• Other routers store these labels as remote labels
in their routing table along with prefixes

81
 Destination Based Forwarding
Example
• R1 acts as Label Edge Router (LER)
– applies labels to arriving IP packets after complete IP lookup
• Packet destined to 10.1.1.5 arrives at R1
– R1 matches the prefix 10.1.1, attaches label ’15’ to packet, send
to R2

82
 Destination Based Forwarding
Example
• R2 checks the label of incoming packet, consults its
table for outgoing interface (1), updates the label
value to ’24’ and forwards the packet to R3
• R2 doesn’t consult IP address for forwarding packet
to R3 !!!

83
 MPLS – Achievements
• Fixed-length label lookup instead of variable
length IP prefix lookup
– Simpler to implement exact match instead of longest
match algorithm
• Only forwarding algorithm is changed
– Any standard routing algorithm may still be used
– Packets will follow the path selected by IP routing
• Devices that don’t know how to forward IP
packets can be used as Label Switching Routers
– ATM switches without changing forwarding hardware

84
 How to Attach Labels to Packets ?
• Depends upon the type of link carrying the
packets
• When IP packets are carried as complete frames
– Label is inserted as a “shim” between layer 2 header
and layer 3 header
– Like on most link types: Ethernet, Token Ring, PPP
• When the switch function as an MPLS LSR
– Label need to be in a place where switch can use it
– In ATM, labels are inserted in ATM cell header
(combination of VPI and VCI).
85
Methods of Attaching Labels

86
 An MPLS Network
• A mixture of conventional IP routers, label edge
routers (LER) and ATM switches (as LSRs)
– All are using the same routing protocols
• Reduction in number of adjacencies that each
router must maintain
– Greatly reduce the amount of work of each router
• Edge routers have a full view of complete
topology of the network
– Edge routers pick a new path in case of node failure

87
 MPLS – Example of MPLS Network
• Conventional Network
– Overlay of virtual circuits
– Each router connect to other by a virtual circuit

88
 MPLS – Example of MPLS Network
• MPLS Network
– Routers peer directly with LSRs
– No virtual circuits interconnecting routers
– Each router has only one adjacency

89
 MPLS – Explicit Routing

• MPLS provides a convenient way to add

capabilities similar to source routing to IP
networks – Explicit routing
• Not really source of the packet that picks the route
– different from source routing
• One of the routers inside the service provider’s
network does this
• Application
– Traffic Engineering

90
 Example – Explicit Routing
• “Fish” network
• Route for traffic from R1 to R7 is R1-R3-R6-R7
• Route for traffic from R2 to R7 is R2-R3-R4-R5-R7
• Good use of capacity available along two distinct paths
• Different from normal IP routing

91
 MPLS – Explicit Routing
• With MPLS enabled routers, very easy to achieve
desired routing
• How can we make sure to choose different paths?
– If labels of R1 and R2 are different, R3 can send the
packets along different paths
• How to agree on labels in a network?
– Normal label switch forwarding doesn’t work
– RSVP (Resource Reservation Protocol)

92
 MPLS – Explicit Routing

• Fast reroute – a capability to make networks more

resilient in the face of failure
• Possible to precalculate a path (backup path)
between router ‘A’ to router ‘B’ that explicitly
avoids a certain link ‘L’
• Can significantly reduce the time taken to reroute
packets around a point of failure
• CSPF (Constrained Shortest Path First) – most
common algorithm to calculate explicit routes

93
 MPLS VPN - An ATM Circuit

• Pseudowire Emulation
• Routers interconnected by a tunnel
• Tunnel header

94
Forwarding ATM Cells in MPLS VPN

95
Example of Layer 3 VPN

96
Routing for Mobile Hosts

97
 Routing for Mobile Hosts
• Mobile hosts, fixed infrastructure
MN
– e.g. cellular networks

MN

• Mobile hosts, dynamic infrastructure

– Ad hoc networks
MN MN

MN MN MN

98
 Routing for Mobile Hosts
• How can mobility be supported in view of the fact
that a portion of an IP address is a network address?
– Host-specific information is out of the routers to make
routing scalable
– If host changes the network, it would become
unreachable
• One solution is to acquire new address via DHCP
• If host changes the network frequently, keeping a
single address would be beneficial
– Avoids termination of connection while in move

99
 Mobile IP
• Mobile-IP is a solution that works without any
changes to non-mobile hosts (typical of IETF)
– Supports both mobility and geographic aggregation
• A router known as home agent (HA) is required at
the home network of the mobile host
• Mobile host (MH) uses permanent home address,
and remains connected as it roams
• A router known as foreign agent (FA) is required
at the network to which the mobile host attaches

100
 Mobile IP
• Home & foreign agents send advertisement
messages
• Mobile host attaches to a nearby foreign agent
– Mobile host communicates home agent address
– Foreign agent contacts home agent to communicate
care-of-address
Sending host
Home agent Foreign agent
(10.0.0.3) (12.0.0.6)
(10.0.0.9)
(network 10) Internetwork Mobile host
Home network IP tunnel

101
 Mobile IP - Features
• Transparent routing of packets to a mobile host
• No modification of existing routers or non-
mobility supporting hosts
• Mobile Host (MH) is assigned a unique home
address within its home network
• Other hosts communicating with the MH always
use MH’s home address
• Foreign agent (FA) maintains a mapping of the
MH’s home address to its care-of-address

102
 Mobile IP – Home Agent (HA)
• An agent on the MH’s home network
• Maintains registry of MH’s current location
• MH’s location is care-of-address
• Mobility binding is the connection between the
MH’s home address and care-of-address
• Each time the MH establishes a new care-of-
address, it must register with its HA

103
 Mobile IP - Challenges

• How does the MH get a local IP address ?

• How does the HA intercept a packet that
is destined for the MH ?
• How does the HA then deliver the packet
to the FA ?
• How does the FA deliver the packet to the
MH ?

104
 Mobile IP
Assigning Care-of-address

• MH discovers foreign agent (FA) using an

agent discovery protocol
• MH registers with FA and FA’s address
becomes MH’s care-of-address
• MH obtains a temporary IP address from FA
or via DHCP-like procedures

105
 Mobile IP
• Home agent is required to intercept all packets for
the mobile host
– it sends a gratuitous ARP to use proxy-ARP when
mobile host registers with foreign agent
• Home agent uses tunneling (IP-in-IP) to forward
packets of mobile host to foreign agent
• Foreign agent strips the extra IP header and
forwards the packet to the mobile host
• Mobile host can behave as foreign agent, if absent

106
 Mobile IP – Routing Problems
• Indirect (triangle) routing
– Places unnecessary burden on the internet
– Significantly increases latency
– Route from sending node to mobile node can be
significantly sub-optimal
• Both nodes on the same network, far from
home agent
• Handoff between FA
– Causes connection to be rerouted through HA

107
 Route Optimization in Mobile IP
• Making sending node aware of the care-of-address
• Home agent sends a binding update to the source
• Source uses new address and records it in its
binding cache
• Foreign agent sends binding warning if the source
uses out of date binding cache (if mobile host is
not FA)
• Security issue if someone announces itself as FA

108
 Route Optimization in Mobile IP
• Correspondent node caches MH location and care-
of-address

Sender FA MH

HA

109
 Handoff Optimization in Mobile IP
• MH informs old FA of new FA. Old FA forwards
MH’s packets through new FA

Sender FA1 MH

Home
Agent FA2 MH

110
 Routing in Ad Hoc Networks
• Collection of mobile nodes
• Form a temporary network
• In a distributed manner
• No server or access points
• Management protocols
– Periodic polling (proactive)
– On-demand routing (reactive)

111
 Proactive Routing Protocols
• Driven by timer based mechanisms
• Protocols send periodic routing advertisements
• Link status detection is beacon-based (hello’s)
• Concerns
– Updates waste bandwidth and power (especially if
nothing changes)
– Topology changes may be too dynamic/rapid to be
captured by periodic updates
– Shortest path may not be best path (signal strength,
energy consumption)

112
 Reactive Routing Protocols
• Driven by data packets requiring delivery
• Discover route only when needed
– Soft-state routing table contains forward and reverse
route
– Uni-directional links may foil routing
• Detect link status when forwarding
• Concerns
– Latency to set up route
– Overhead for route discovery/maintenance
113
 Ad Hoc Routing Protocols
• Proactive protocols
– OLSR (Optimized Link State Routing)
– TBRPF (Topology Broadcast based on Reverse Path
Forwarding)
• Reactive protocols
– DSR (Dynamic Source Routing)
– AODV (Ad hoc On-demand Distance Vector) Routing
• IETF’s MANET working group deals with them
– www.ietf.org/html.charters/manet-charter.html

114