IP Switching and Routing Essential Chapter 4 Link State Routing and OSPF PowerPoint PPT Presentation

presentation player overlay
1 / 59
About This Presentation
Transcript and Presenter's Notes

Title: IP Switching and Routing Essential Chapter 4 Link State Routing and OSPF


1
IP Switching and Routing EssentialChapter 4Link
State Routing and OSPF
  • Shuhei Tanigawa
  • 2005/7/07

2
Special Networks
  • Until now, the discussion of OSPF has treated all
    subnetworks as if they were simple point-to-point
    links.
  • Real network consist of variety of network
    technologies.

3
Special Networks
  • OSPF makes special allowance s for three special
    network types.
  • Broadcast networks
  • Nonbroadcast networks
  • Demand networks

4
Broadcast Networks
  • Broadcast networks provide an inherent broadcast
    or multicast capability.
  • They allow any system to communicate directly
    with any other system.
  • The most common type of broadcast network is a
    local area network such as Ethernet.

5
Broadcast Networks
  • Broadcast networks merit special treatment
    because of their any-to-any flexibility.
  • Consider, for example, the Token Ring LAN in
    figure 4.12
  • Each router can communicate with other four.
  • Five routers create a total of 20 entries in the
    link state database.
  • The number of entries grows as the square of the
    number of routers.

On large networks, this growth represents a
serious problem.
6
Broadcast Networks
  • OSPF elects a special router, known as designated
    router, from among those on networks.
  • This router treats all routers on the network as
    neighbors.
  • The other routers consider only the designate
    router as their neighbor.
  • As far as routing calculations are concerned,
    traffic between two such routers must pass
    through the designated router.

7
Broadcast Networks
  • Figure 4.14 shows the artificial topology OSPF
    create.
  • The designated router does not really exist.
  • One of the other routers takes on the role of
    designated router in addition to its other
    responsibilities.
  • Each true router on the network reports only the
    designated router as a neighbor.
  • These links to the designated router are known as
    network links.
  • The designated router reports each true routers
    as a neighbor.

8
Broadcast Networks
  • In order to force a correct route calculation,
    all neighbors advertised by the designated router
    have a distance of zero.

Which reflect the true distance
Defined to be zero
9
Broadcast Networks
  • The OSPF protocol takes advantage of broadcast
    networks when it floods LSA packets.
  • The designated router transmits LSAs to a special
    multicast address.
  • All OSPF routers listen to this address.
  • Regular routers simply send LSAs to the
    designated router.

10
Broadcast Networks
  • The designated router plays a key role in OSPFs
    operation on a broadcast network.
  • To reduce vulnerability, OSPF elects a backup
    designated router.
  • It keeps track of the same information as the
    designated router.
  • If the backup detects a failure of the designated
    router, it becomes active immediately.

11
Nonbroadcast Multi-Access Networks
  • The designers of OSPF originally developed the
    designated router concept for local area
    networks.
  • The way electing a special router can also work
    effectively on other networks.
  • Case study figure 4.15, 4.16-
  • Such a network is known as a nonbroadcast multi
    access (NBMA) network.
  • OSPF can handle the scaling problem with a
    designated router.

12
Nonbroadcast Multi-Access Networks
  • There are only two real differences between
    OSPFs treatment of broadcast networks and NBMA
    networks.
  • LSA flooding
  • Process for electing designated router

13
Point-to-Multipoint Networks
  • The designated router concept works only for
    networks in which every router communicate
    directly with every other router.
  • As the size of the network grows provisioning a
    separate virtual circuit between every pair of
    router can become both inefficient and expencive.

14
Demand Networks
  • Demand networks are networks whose expense is a
    direct function of usage.
  • Narrowband ISDN links
  • Such networks earn the name demand because they
    should be active only when actual application
    traffic demands their use.

15
Demand Networks
  • OSPF normally counts on links remaining active
    indefinitely.
  • Routers continually exchange hello packets.
  • LSAs are also periodically reflooded through a
    network.
  • Even without user traffic, these packets will
    consume bandwidth on a demand network.

16
Demand Networks
  • OSPF makes two changes to its normal behavior.
  • It eliminates the periodic hello packets.
  • It refrains from sending periodic LSA packets
    across demand networks.
  • Routers must remove age limit from LSA packets.

17
Multicast Routing
  • The OSPF protocol also provides experimental
    support for multicast routing.
  • Link state protocol require only slight
    enhancements to support multicast.

18
Multicast Routing
  • Consider the sample network of figure 4.18.
  • All the links have the same cost.
  • This example focuses on router B.
  • Consider how router B forwards a unicast packet
    from the personal computer to the server.
  • Router B must know where the server is located.
  • This is the information that Dijkstra computation
    provides.
  • Figure 4.19 shows the shortest path tree.

19
Multicast Routing
  • Note the two important facts about unicast
    routing.
  • The root of shortest path tree is router B
    itself.
  • It does not matter where the packet being routed
    originated.
  • Figure 4.20 shows that the PC sends a single
    multicast packet to all workstations on the
    network.
  • Router B must construct a different tree like in
    figure 4.21

20
Multicast Routing
  • There is a major difference between the multicast
    tree and the unicast tree.
  • Two trees have different roots.
  • With the multicast tree, the root is the source
    of the packets.
  • There maybe many destinations on the multicast
    tree.

21
Multicast Routing
  • Consider what happens when the minicomputer sends
    a multicast packet.
  • Figure 4.22 shows what router B should do with
    such packet.
  • It should forward a packet to router C.
  • This next hop differs from the last case.

22
Multicast Routing
  • A correct shortest path tree leads to the right
    forwarding decision.
  • Figure 4.23 shows the tree for multicast packets
    from the minicomputer.
  • It is clearly different from figure 4.21.
  • The new tree correctly points to router C as the
    next hop.

23
Multicast Routing
  • Multicast routing can present a significant
    problem for OSPF routers.
  • Those routers must calculate a different shortest
    path tree for each source system.
  • Dijkstras calculation can be very
    computationally intensive, particularly with
    large networks.
  • ODPF strongly recommends that routers calculate
    multicast trees only when a multicast packet
    arrives for forwarding.
  • They should then cache the results of those
    calculations.

24
OSPF Message Format
  • OSPF protocol packets are themselves carried as
    payload of IP datagrams.
  • A specific next header value of 89 identifies the
    payload of OSPF.
  • All packets begin with a common OSPF header.
  • The header includes eight fields.

25
OSPF Message Format
vers
hlen
diffserv
ECN
payload length
fragment identifier
0
D F
M F
fragment offset
IP header
hop limit
next hdr89
header checksum
source address
destination address
version 2
type
message length
router ID
area ID
OSPF header
checksum
authentication type
authentication data
rest of OSPF message
26
OSPF Message Format
  • Version
  • The current version number is 2.
  • Type
  • The OSPF protocol uses five different types of
    packets.
  • Message length
  • Router ID
  • One of the routers IP addresses
  • area ID
  • Checksum
  • Authentication type
  • Authentication data

27
Authenticating OSPF Messages
  • Value Authentication type
  • 0 Null authentication
  • 1 Password authentication
  • 2 Cryptographic Authentication
  • Routers configured to use one of these
    authentication schemes on each of interfaces.

28
Authenticating OSPF Messages
  • The null authentication scheme is the simplest of
    all.
  • Its really no authentication at all.
  • The password authentication scheme is only
    slightly less simple.
  • It doesnt really protect against malicious
    parties attacking on OSPF network.
  • It does provide some protection against
    accidental misconfigurations.

29
Authenticating OSPF Messages
  • Cryptographic authentication offers the strongest
    possible authentication.
  • It relies on a special mathematical function
    known as a cryptographic digest.
  • The OSPF specification details support for
    message digest 5 (MD5)
  • MD5 algorithm performs a set of convoluted
    calculation on its input and derives 128-bit
    digest.

30
Authenticating OSPF Messages
version 2
type
message length
router ID
area ID
checksum
authentication type2
0
auth.len16
key ID
cryptographic sequence number
OSPF data
message digest (16 bytes)
31
Meeting Neighbors
version 2
type
message length
router ID
area ID
checksum
authentication type
authentication data
network mask
hello interval
priority
options
router dead interval
designated router
backup designated router
neighbor 1
neighbor 2
other neighbors
neighbor n
32
Advertising Link State
version 2
type4
message length
router ID
area ID
checksum
authentication type
authentication data
number of advertisement
LSA header
LSA data
other LSA
LSA header
LSA data
33
Advertising Link State
  • Link state header

LS age
LS type
options
link state ID
advertising router
link state sequence number
link state checksum
length
34
Advertising Link State
  • Value LSA Type
  • 1 Router link
  • 2 Network link
  • 3 Summary link to network
  • 4 Summary link to AS boundary router
  • 5 External link
  • 6 Group membership advertisement
  • 7 NSSA link
  • 9 Opaque link confined local network
  • 10 Opaque link confined to an area
  • 11 Opaque link for an entire AS

35
Advertising Link State
  • Link state header

LS age
LS type
options
link state ID
advertising router
link state sequence number
link state checksum
length
36
Router Links
  • The simplest link type is a router link.
  • It represents a normal link between two routers.

37
Router Links
LS age
LS type1
options
link state ID
advertising router
link state sequence number
link state checksum
length
Links header
Router type
0
Number of links
Link ID
Link data
Default metric
Link type
TOS count
Link 1
TOS value
0
TOS metric
Other TOS metric
Link ID
Link data
Default metric
Link type
TOS count
Link 2
TOS value
0
TOS metric
Other TOS metric
other links
38
Network Links
  • The second type of LSA is the network links.
  • Network that have a designated router use the
    network LSA.

Router link (from true router to designated
router)
Network link (from designated router to true
router)
39
Network Links
LS age
LS type2
options
link state ID
advertising router
link state sequence number
link state checksum
length
Network mask
Attached router 1
Attached router 2
Other attached routers
other LSAs
40
Summary Links
  • The next two type of LSAs are summary LSAs.
  • Area border routers distribute these within their
    areas to advertise destinations outside of the
    area.
  • The different LSA types indicate what those
    destinations represents.
  • Type 3 LSAs identify other networks within the
    AS.
  • Type 4 LSAs identify AS boundary routers.

41
Summary Links
LS age
LS type3
options
link state ID
advertising router
link state sequence number
link state checksum
length
Network mask
0
metric
TOS value
TOS metric
Other TOS metrics
other LSAs
42
Summary Links
LS age
LS type4
options
link state ID
advertising router
link state sequence number
link state checksum
length
Network mask0
0
metric
TOS value
TOS metric
Other TOS metrics
other LSAs
The main purpose of a type 4 LSAs is simply to
announce the presence of an AS boundary router.
43
External Links
  • Information of type 4 LSAs alone does not tell
    other ASes what destinations are available beyond
    the AS.
  • This is the job of external(type 5)
    advertisement.

44
External Links
LS age
LS type5
options
link state ID
advertising router
link state sequence number
length
Link state checksum
E
0
metric
Forwarding address
External route tag
Other TOS metrics, forwarding addresses, and tags
other LSAs
45
Group Membership Advertisements
  • Group addresses require their own advertisement
    type, and designated routers originate it.
  • For each group that has any members, designated
    routers build a type 6 LSAs.

46
Group Membership Advertisements
LS age
LS type6
options
link state ID destination group
advertising router
link state sequence number
length
Link state checksum
Vertex type
Vertex ID
Other vertex types and IDs
other LSAs
47
NSSA Advertisements
LS age
LS type7
options
link state ID
advertising router
link state sequence number
length
Link state checksum
E
0
metric
Forwarding address
External route tag
Other TOS metrics, forwarding addresses, and tags
other LSAs
48
Opaque Advertisements
  • The final three types of LSAs are opaque
    LSAs.(type 9,10, and 11)
  • The only difference between three is the three is
    the extent to which OSPF floods them.
  • Local network (type 9)
  • A single area (type10)
  • An entire AS (type 11)
  • Opaque LSAs provide a way to give OSPF new
    capability in the future.

49
Reliable Flooding
  • The different types of LSAs define the complete
    topology of network.
  • In order to distribute that information to all of
    the networks router, OSPF floods link state
    update packet throughout the network.
  • The OSPF protocol takes the flooding procedure
    one step further.
  • It requires routers to explicitly acknowledge
    when they receive an advertisement.

50
Reliable Flooding
  • The link state acknowledgment packet contains a
    list of link state headers.
  • Because the header is sufficient to identify an
    advertisement.
  • A single acknowledgment packet can acknowledge
    many link state updates.

Link state update
Link state acknowledgment
51
Updating Neighbor
  • This chapter has presented OSPF as if networks
    operated in a completely orderly manner.
  • Real networks never function this neatly.
  • In particular, routers are usually introduced to
    networks that are already functioning.
  • Router must rapidly catch up and learn the
    networks topology.

52
Updating Neighbor
  • To catch up with the rest of the network, a newly
    introduced router relies on its neighbor.
  • As soon as two routers greet each other, they
    exchange information about their link state
    database.
  • They do so with database description packet.

53
Updating Neighbor
version 2
type2
message length
router ID
area ID
authentication type
checksum
options
0
I
S
M
Database description sequence number
LSA 1 header
LSA 2 header
other LSA headers
LSA n header
54
Updating Neighbor
  • Once a router receives a complete set of database
    description packets from its neighbor, it
    examines its own link state database.
  • Most likely, the router will find that its
    neighbor has at least some information that
    lacks.
  • The router requests the updated information from
    its neighbor.
  • It does so with a link state request packet.

55
Updating neighbor
  • The link state request contains a list of LSAs
    that the sender wish to receive.
  • These LSAs are identified solely by their type,
    link state ID, and advertising router.
  • When the neighbor receives a request, it finds
    the advertisements in its link state database and
    forwards them in link state update packet.

56
Updating Neighbor
  • After exchanging three packets, two routers will
    have successfully synchronized their link state
    databases.
  • database description packets
  • link state requests
  • link state updates

57
Summary
  • Routers rely on routing protocols like OSPF to
    learn their map.
  • They see how to reach the networks destinations.
  • The OSPF is one of the family of link state
    routing protocol.
  • Link state protocols proceed in three steps.
  • The OSPF protocol organizes networks into
    hierarchies.
  • AS
  • Area

58
Summary
  • The OSPF protocol has flexibility to operate over
    a wide variety of links.
  • Point-to-point links
  • Broadcast networks
  • Nonbroadcast multi access networks
  • Point-to-multipoint networks
  • Demand networks
  • In many case, OSPF routers elects a designated
    router to reduce traffic demands on those
    networks.

59
Summary
  • The OSPF protocol also has experimental support
    for multicast routing.
  • As the use of multicast increases, network
    engineers should gain greater understanding of
    OSPFs limitaitons.
Write a Comment
User Comments (0)
About PowerShow.com