Chapter 4 LinkState Routing and Hierarchical Routing

1
Chapter 4Link-State Routing and Hierarchical
Routing

• Professor Rick Han
• University of Colorado at Boulder
• rhan_at_cs.colorado.edu

2
Announcements

• Last weeks lectures are online
• Homework 4 is online, due March 6
• PA 1 should be returned to you, see TA
• PA 2 due Saturday by 1159 pm
• Midterm likely Thursday, March 13
• Next, link-state routing, hierarchical routing

3
Recap of Previous Lecture

• Dijkstra as alternative Shortest Path algorithm
  to distributed Bellman-Ford Equation
• Iteratively grow a shortest path spanning tree
  from the root outwards
• Link-state routing
• Dijkstra shortest path algorithm,
• Reliable flooding of LSPs to all nodes
• Link-state Metrics
• Queue size
• More advanced

4
Revised LS ARPANET Metric

• If a loaded link looks very bad then everyone
  will move off of it
• Want some to stay on to balance load and avoid
  oscillations
• It is still an OK path for some
• Use a hop-normalized metric that
• Has a limited range of values for a given link
• Doesnt jump too much between updates for a given
  link
• Has a limited range of values across different
  link types
• ? gradual change

5
Revised LS ARPANET Metric (2)

• Revised metric is a function of
• Smoothed bounded link utilization
• Link type
• Link utilization
• Measured link util. is sampled over 10sec period
• Link utilization .5current sample .5last
  average
• Normalized according to link type
• Satellite should look good when queuing on other
  links increases

6
Routing Metric vs. Link Utilization
225
9.6 satellite
140
New metric (routing units)
90
9.6 terrestrial
75
56 satellite
60
56 terrestrial
30
0
50
100
25
75
Utilization
7
Observations

• Utilization effects
• High load never increases cost more than 3cost
  of idle link
• Cost f(link utilization) only at moderate to
  high loads
• LSPs flood only if change in Cost exceeds
  threshold
• Link types
• Most expensive link is 7 least expensive link
• High-speed satellite link is more attractive than
  low-speed terrestrial link
• Allows routes to be gradually shed from link

8
Revised LS ARPANET Metric (3)

• Better utilization than earlier two metrics
• Less oscillation
• Allows routes to be gradually shed from link
• But is it sufficiently responsive to avoid stale
  updates?
• Cant propagate updates and react fast enough to
  short time-scale changes in link cost, so
  sluggish response is OK
• Perlman complex.algorithmsbad idea, little
  difference in network capacity between fixed
  cost assignment and dynamic metric

9
Scalability in Internet Routing

• Neither Distance Vector (RIP) nor Link-State
  (OSPF) scale well to many nodes
• DV slow to propagate and converge
• LS overhead of flooding all nodes to all nodes
• DV 50 million nodes gt long distance vectors
• Solution use hierarchy
• Group local routers into a domain, can be a
  subnet on local area or Autonomous System (AS) on
  wide area
• All routers within an AS use an intra-domain
  routing protocol, e.g. RIP and OSPF
• Use another inter-domain routing protocol to
  route between ASs

10
Scalability in Internet Routing (2)
Inter-Domain Routing
AS 1
AS 2
Border/ Gateway Router
Border/ Gateway Router
RIP
OSPF
Intra-Domain Routing
11
Inter-Domain Routing

• Routers within an AS share a common prefix in
  their IP address, i.e. the top 16 bits are all
  the same
• aggregation of routes
• Why Inter-Domain Routing is hard
• 50000 CIDR prefixes to store in each router
• Assigning a cost to the AS between two border
  routers is controlled by owner of each AS
• Advertising a cost of 1000 is fast for one AS1,
  slow for AS2
• Inter-Domain only advertises a reachable path
  across multiple ASs, not shortest path
• Each ASs owner wants per-route QOS/tariffs
• Policy-based routing over performance-based
  routing

12
Border Gateway Protocol (BGP)

• Interdomain Routing
• Similar to Distance Vector, but called Path
  Vector instead
• BGP router advertises only reachability info in
  its vector, not costs/hop counts
• E.g. networks 128.96, 192.4.153, and 192.4.3 can
  be reached from AS2
• BGP router advertises its path to each
  destination in its vector
• Avoids loops

13
BGP (2)

• Each routing update carries the entire path
• Loops are detected as follows
• When AS gets route, check if AS already in path
• If yes, reject route
• If no, add self and (possibly) advertise route
  further

R1 to AS 1
R3 to AS 3
R2 To AS 2
14
BGP (3)

• How do intra-domain routers learn about routes to
  other ASs?
• Option 1 Border router injects a default route
  into intra-domain routing protocol for all
  addresses not advertised within AS
• Option 2 Multiple border routers inject a
  specific network prefix into intra-domain routing
  protocol
• E.g. I, BGP 1, have a link to 192.4.54/24 of
  cost X, and I, BGP 2, have a link to
  192.4.72/24 of cost Y

15
BGP (4)

• Perlman Although were probably stuck with BGP
  forever, Ive never been convinced it is the
  right approach.
• Perlman I think the only way to solve the
  general case of policy-based routing is with a
  link state protocol plus source-specified routes.

16
Interior Border Gateway Protocol

• Each AS may have many border routers
• Each border routers could inject 10000 prefixes
  from neighboring AS
• LSPs too large
• Shortest path calculations too expensive
• Border routers use interior BGP (IBGP) to limit
  routing info received by internal AS routers
• IBGP routers determine best route to each
  destination
• Only the best interior BGP router injects info
  into AS
• Any router in AS learns one best border router to
  use when sending a packet externally

17
Hierarchy In Addition To BGP

• OSPF has its own hierarchy group OSPF routers
  into areas
• Hierarchy AS gt OSPF area -gt OSPF network
• Subnets
• Fixed Classes A,B,C inefficient - Class B
  exhaustion
• Subdivide a Class B IP address 128.96.34.15 into
  ltNetwork ID, Subnet ID, Host IDgt
• IP address is ANDed with subnet mask to extract
  subnet address
• Subnet mask 255.255.255.0 ANDed with IP address
  128.96.34.15 gives subnet address 128.96.34
• Subnet mask 255.255.255.128 ANDed with IP
  address 128.96.34.15 gives subnet address
  128.96.34.0

18
Additional Hierarchy (2)

• Subnets
• When host 1 wants to send to host 2, AND the
  subnet mask with the destination IP address
• If result is same subnet as sending host 1, then
  send over local LAN subnet
• If result differs, then route to another subnet
  using subnet-to-subnet routing
• Forwarding table changes from ltdestination IP,
  next hopgt to ltdestination subnet, subnet mask,
  next hopgt
• For each entry, router ANDs subnet mask with
  dest. IP address and looks for match with
  destination subnet
• Longest match breaks a tie

19
Additional Hierarchy (3)

• CIDR (Classless Interdomain Routing) Subnets
• When subnet mask is top N bits, then have a CIDR
  network prefix,
• 192.4.16 with 20 bit prefix is written
  192.4.16/20
• Approaches for fast prefix matching
• How do nodes advertise their CIDR prefix/mask?
• IP header only has 32-bit address
• Where is subnet mask?
• BGP-4 path vectors and OSPF LSPs carry the CIDR
  prefix along with the IP address, e.g. 192.4.16/20

20
Additional Hierarchy (4)

• How do CIDR and non-CIDR routing stay compatible?
• OSPF and BGP support CIDR, RIP does not
• RIP builds a routing table by falling back to the
  old Class A,B, C network prefixes
• makes RIP more inefficient
• Packets are still routed correctly
• CIDR Bottom line
• Improves address assignment efficiency
• Helps aggregate routing to occur between networks
  rather than nodes

21
Fast Matching of Variable Prefixes

• Need to match CIDR network prefix with IP
  packets destination address
• Brute force for each destination router in list
• apply mask to match prefix with destination
  addresss prefix
• choose longest match

22
Fast Matching of Variable Prefixes (2)

• Speeding it up Organize prefixes into a Patricia
  tree
• If Nth bit is zero, go left, otherwise go right
• Automatically finds longest match
• Worst case 32 bit tests

Bit to test 0 left child,1 right child

1
0

default 0/0
0
1


0
1
1
0
128.2/16
252.32.150/24
192.3/20
163.32/16
23
Dynamic Host Configuration Protocol (DHCP)

• RARP A host knows a destinations MAC address,
  but not destinations IP address.
• If destinationitself, then same goal as DHCP
• BOOTP similar goal to RARP, devised same time
  (1985)
• DHCP a host knows its own MAC address, but
  doesnt have an IP address yet
• Due to hierarchical addressing on network, cant
  have manufacturer-preassigned IP addresses
• Manual configuration is time-consuming,
  inflexible to changes, wastes addresses on
  disconnected nodes

24
DHCP (2)

• Goal Automatic configuration of a hosts IP
  address
• A host queries a DHCP server to obtain an IP
  address
• How does a host find the address of a DHCP
  server?
• Host sends a DHCPDISCOVER limited IP broadcast
  packet, with destination address 255.255.255.255
• Routers never forward such a packet, so it stays
  within LAN

IP Router
LAN1
LAN2
DHCP Server
Requesting Host
25
DHCP (3)

• DHCP relays enable one DHCP server per
  administrative domain, rather than one server per
  network
• Requires a DHCP relay on each network
• DHCP relay sends a unicast IP packet to DHCP
  server when it hears a local IP broadcast packet
  with DHCPDISCOVER

IP Router
LAN1
LAN2
DHCP Relay
DHCP Server
Requesting Host
26
DHCP (4)

• DHCP server selects a dynamic IP addr. from pool
• maps hosts MAC address to the dynamic IP address
• Another advantage of relays enable DHCP
  responses to get back to requesting host
• Server cant send directly back using hosts MAC
  address
• DHCP server sends unicast to known IP address of
  DHCP relay, which sends to hosts local MAC
  address

IP Router
LAN1
LAN2
DHCP Relay
DHCP Server
Requesting Host
27
DHCP (5)

• Hosts cannot keep dynamic IP addresses
  indefinitely
• Timeout/lease by DHCP
• 3 days for Windows NT, 8 days for Windows 2000, 1
  day
• Configurable when starting DHCP server
• Host must periodically renew lease, otherwise IP
  address goes back into pool of available
  addresses
• DHCP is implemented as an application-level
  protocol on top of UDP and IP