Use of Link State Routing in Dynamic Hierarchical Link State DHLS Routing

1
Use of Link State Routing in Dynamic
Hierarchical Link State (DHLS) Routing

• George Lee

2
Outline

• Dynamic Colored-based Zone Routing (DCZR) ?
  Dynamic Hierarchical Link State Routing (DHLS)?
• Zone-based Hierarchical Link State (ZHLS)
• Future Work
• RFC 2328 - OSPF Version 2/The OSPF Working Group
• http//www.ietf.org/html.charters/ospf-charter.htm
  l
• Internet Draft - Optimized Link State Routing
  Protocol (OLSR) /The MANET Working Group
• http//www.ietf.org/html.charters/manet-charter.ht
  ml

3
DHLS (1)

• Dynamic Colored-based Zone Routing
  (DCZR) ?Dynamic Hierarchical Link State
  Routing(DHLS) ?
• Aiming at ZHLS (Zone-based Hierarchical Link
  State)

4
DHLS (2)

• GPS-based
• Zone-based
• Dynamic and Hierarchical zoning
• Merging
• Splitting

5
DHLS (3)

• DHLS includes the following procedures
• Dynamic and Hierarchical Zoning
• Joining a zone
• Updating the zone
• Merging zones to form a higher-level zone
• Splitting a zone into several lower-level zones
• Zone Coloring?
• Location Search?
• DHLS Routing

6
DHLS (4)

• Joining a zone
• When a node starts up
• It gets its position information from GPS.
• It maps its position to the zone map to obtain
  its Zone ID of the unit zone it stays in.
• It broadcasts a JOIN_REQUEST packet, along with
  its position, Node ID, and Zone ID.
• It starts a timer, tjr, and waits for a
  JOIN_REPLY sent from a certain cluster head,
  given the Zone ID of the current zoning level and
  the position of the cluster head..

7
DHLS (5)

• Updating the zone
• Each node periodically updates its position by
  GPS.
• When a node learns that it has moved out of the
  boundary of the zone of current level, it runs
  the joining-a-zone procedures as start-up.

8
DHLS (6)

• Merging
• Four neighboring level-i zones that form a square
  and the lower left zone is located exactly at the
  multiple of 2i horizontal and vertical unit zones
  from the origin may be merged to become a larger
  zone, namely level-(i1) zone.
• The four neighboring sub-zones that may be merged
  are called peers.
• The four peers are named counterclockwise as the
  0th, 1st, 2nd, and 3rd quadrant, individually.
• If all four level-l peers are colored black, they
  are merged to a become a level-(l1) zone.
• Splitting
• A level-i zone may be split into four level-(i-1)
  zones.

9
DHLS (7)

• An example of the hierarchy

10
DHLS (8)

• An example of the zone backbone (top-level zone)

11
DHLS (9)

• Why Link State routing?
• To minimize the time needed for location search
  and routing decision.
• Shortest path (hop or delay)
• Less routing info. (link state)
• To solve the hole problem
• Problems of Link State routing
• Flooding
• Hop increase
• Consistency/Synchronization
• Design Goals
• To reduce the amount of routing information to
  exchange.
• To minimize the hop increase

12
DHLS (10)

• Location Search overhead
• Assume NN unit zones, where N4p
• Number of levels 1log4N2 12log4N 2p1
• Number of zones ?

13
DHLS (11)

• Path Length
• Assume NN unit zones, where N4p
• Number of levels 1log4N2 12log4N 2p1
• Number of hops per path ?

14
ZHLS (1)

• Zone-based Hierarchical Link State
• A GPS-based Peer-to-Peer Hierarchical Link State
  Routing For Mobile Ad Hoc Networks
• Joa-Ng and Lu,VTC00
• A Peer-to-Peer Zone-Based Two-Level Link State
  Routing for Mobile Ad Hoc Networks
• Joa-Ng and Lu, JSAC, August 1999

15
ZHLS (2)

• Zone Map
• A node knows its physical location by GPS (Global
  Positioning System)
• It can determine its zone ID by mapping its
  physical location to a zone map, which has to be
  worked out at the design stage.
• The zone size depends on factors such as node
  mobility, network density, transmission power,
  and propagation characteristics.

16
ZHLS (3)

• Two levels of topology
• Node level topology physical link information
• If two nodes are within the communication range,
  a physical link exists.
• Zone level topology virtual link information
• If there exists at least one physical link
  connecting two zones, a virtual link exists.

17
ZHLS (4)

• Two types of hierarchical link state packets
  (LSP)
• Node LSP (Intra-cluster link state)
• Contains a list of connected neighbors
• Propagated locally within its zone
• Zone LSP (Inter-cluster link state)
• Contains a list of connected zones
• Propagated globally throughout the network

18
ZHLS (5)

• Two types of hierarchical link state packets
  (LSP)
• Node LSP (Intra-cluster link state)
• Contains a list of connected neighbors
• Propagated locally within its zone
• Zone LSP (Inter-cluster link state)
• Contains a list of connected zones
• Propagated globally throughout the network

19
ZHLS (6)

• Clustering
• The node starts the intra-zone clustering and
  then the inter-zone clustering procedures to
  build its routing tables.
• Two levels of clustering
• Intra-zone Clustering
• Inter-zone Clustering

20
ZHLS (7)

• Intra-zone Clustering
• Each node broadcasts asynchronously a link
  request.
• Nodes within its communication range reply with
  link response ltnode ID, zone IDgt.
• Each node generates its node LSP that contains
  (1) the node ID of its neighbors of the same
  zone and (2) the zone ID of its neighbors of the
  different zones.
• Each node propagates its node LSP locally
  throughout its zone
• Each node knows the node level topology of that
  zone. The Shortest Path algorithm is used to
  build its intra-zone routing table.
• A timer is set for every node LSP received. Any
  expired entry in the routing table will be
  deleted.

21
ZHLS (8)

• Inter-zone Clustering
• Gateway nodes
• Nodes that receive link response from the nodes
  of its neighboring zones.
• Since node LSP contains the zone IDs of the
  connected zones, each node will know which zones
  are connected to its zone.
• Each gateway node builds the zone LSP and
  broadcasts it throughout the network.
• Each node knows the zone level topology of that
  zone. The Shortest Path algorithm is used to
  build its inter-zone routing table.

22
ZHLS (9)

• Location Search
• The source node (say a) checks if the destination
  node (say b) is in its intra-zone routing table.
• If so, node a routes the data to the node b
  according to is intra-zone routing table.
• If not, node a sends a location request ltas node
  ID, as zone ID, bs node ID, zone Xgt to every
  other zone X.
• Every intermediate node routes the location
  request destined to zone X to zone X according to
  its inter-zone routing table.
• A gateway node will receive the location request
  and check its intra-zone routing table to see if
  node b exists in its zone.
• If so, the gateway node will reply a location
  response ltbs node ID, bs zone ID, as node ID,
  as zone IDgt to node a.

23
ZHLS (10)

• Routing
• First check the intra-zone routing table
• Then check the inter-zone routing table

24
ZHLS (11)

• Simulation
• Comparing ZHLS and flat LSR (Link State Routing)
• A Maisie simulator
• The total amount of communication overhead
  generated in ZHLS is much smaller than that
  generated in LSR.
• The average path increases by about 15 when ZHLS
  is used.

25
Future Work (1)

• In ZHLS
• It uses link state algorithm both in intra-zone
  and inter-zone routing.
• A Gateway node periodically floods zone-level
  link state to gateway nodes of other zones.
• The overhead of periodical flooding depends on
  the amount of zones.
• In DHLS
• It uses link state algorithm in all levels of
  intra-zone and inter-zone routing.
• A Zone Header periodically multicast zone-level
  link state to gateway nodes of other zones.
• The overhead of periodical multicasting depends
  on the amount of hierarchies.

26
Future Work (2)

• Simulation of DHZR
• Borland C Builder ? Network Simulator 2 (ns-2)
• Parameters
• R 250 m
• Dense vs. Sparse
• Mobility
• 3km/h (Pedestrians)
• 30km/h (Vehicles in the cities)
• 100km/h (Vehicles on highway)

27
OSPF Version 2 (1)

• Open Shortest Path First (OSPF)
• A link-state routing protocol
• A SPF-based protocol
• A distributed-database protocol
• Interior Gateway Protocol (IGP), Intra-AS Routing
• IETF OSPF Working Group
• RFC 2328 98 OSPF Version 2

28
OSPF Version 2 (2)

• Abstract of OSPF
• Each OSPF router maintains an identical database
  describing the AS's topology.
• From this database, a routing table is calculated
  by constructing a shortest-path tree, (thus,
  loop-free) with itself as root.
• OSPF recalculates routes quickly in the face of
  topological changes, utilizing a minimum of
  routing protocol traffic.
• Traffic is distributed equally among equal-cost
  routes.

29
OSPF Version 2 (3)

• An area (i.e. IP subnetting) routing capability
  is provided, enabling an additional level of
  routing protection and a reduction in routing
  protocol traffic.
• All OSPF routing protocol exchanges are
  authenticated.
• Externally derived routing data (e.g., routes
  learned from an Exterior Gateway Protocol such as
  BGP) is advertised throughout the AS, but is
  kept separate from the OSPF protocol's link state
  data.

30
OSPF Version 2 (4)

• Classification of Routers
• Internal routers
• belonging to the same area
• Area border routers
• attaching to multiple areas
• Backbone routers
• having an interface to the backbone area
• do not have to be area border routers
• AS boundary routers
• A router that exchanges routing information with
  routers belonging to other ASs.
• The paths to each AS boundary router are known by
  every router in the AS.
• AS boundary routers may be internal or area
  border routers, and may or may not participate in
  the backbone.

31
OSPF Version 2 (5)

• Routing protocol packets
• Type Packet name Protocol
  function ____________________________________
  _________________
• 1 Hello
  Discover/maintain neighbors
• 2 Database Description
  Summarize database contents
• 3 Link State Request
  Database download
• 4 Link State Update
  Database update
• 5 Link State ACK
  Flooding acknowledgment
• The Hello protocol
• Hello ? ? Hello
• The Database Exchange Process
• Database Description ?Link State Request ? ?
  Link State UpdateLink State ACK ?

32
OSPF Version 2 (6)

• The Hello Protocol
• Establishing and maintaining neighbor
  relationships.
• Hello packets are sent periodically out all
  router interfaces.
• The Hello Protocol elects a Designated Router and
  a Backup Designated Router for the network.
• Receipt of an Hello Packet may also cause an
  Hello Packet to be sent back to the neighbor in
  response.

33
OSPF Version 2 (7)

• The Designated Router performs two main
  functions
• DR originates a network-LSA (Link State
  Advertisement) on behalf of the network
• Since DR is adjacent to all other routers on the
  network, thus it plays a central part in the
  synchronization process.
• Backup Designated Router

34
OLSR (1)

• Optimized Link State Routing Protocol
• For mobile ad hoc networks
• Proactive routing
• Hop-by-hop routing
• Well-suited for large and dense networks
• The MANET Working Group
• MANET (Mobile Ad Hoc Network)
• Internet DraftOptimized Link State Routing
  Protocol (OSLR)
• http//www.ietf.org/internet-drafts/draft-ietf-man
  et-olsr-06.txt

35
OLSR (2)

• Each node selects a set of its one-hop neighbor
  nodes, with bidirectional link, as "multipoint
  relays" (MPRs).
• The MPRs are used to form the route