1. 2017 session 1 TELE3118: Network Technologies Week 7: Network Layer Control Plane: Intra-Domain Routing
Some slides have been adapted from:
Computer Networking: A Top Down Approach, 7th (global) edition. Jim Kurose, Keith Ross. Pearson, April All material copyright J.F Kurose and K.W. Ross, All Rights Reserved.
Computer Networks, 5th edition. Andrew S. Tanenbaum, David J. Wetherall, Pearson, 2010.
Network Layer

2. Network-layer functions
Recall: two network-layer functions:
forwarding: move packets from router’s input to appropriate router output
data plane
routing: determine route taken by packets from source to destination
control plane
Two approaches to structuring network control plane:
per-router control (traditional)
logically centralized control (software defined networking)
Network Layer: Control Plane

3. Per-router distributed control plane
Individual routing algorithm components in each and every router interact in the distributed control plane
Routing
Algorithm
data
plane
control
values in arriving
packet header
0111 1 2 3
Network Layer: Control Plane

4. Logically centralized control plane
A distinct (typically remote) controller interacts centrally with local control agents (CAs)
Remote Controller CA
data
plane
control
values in arriving
packet header 1 2
0111 3
Network Layer: Control Plane

5. Routing protocols
Routing protocol goal: determine “good” paths (equivalently, routes), from sending hosts to receiving host, through network of routers
path: sequence of routers packets will traverse in going from given initial source host to given final destination host
“good”: least “cost”, “fastest”, “least congested”
routing: a “top-10” networking challenge!
Network Layer: Control Plane

6. Graph abstraction of the networku y x w v z 2 1 3 5
graph: G = (N,E)
N = set of routers = { u, v, w, x, y, z }
E = set of links ={ (u,v), (u,x), (v,x), (v,w), (x,w), (x,y), (w,y), (w,z), (y,z) }
aside: graph abstraction is useful in other network contexts, e.g.,
P2P, where N is set of peers and E is set of TCP connections
Network Layer: Control Plane

7. Graph abstraction: costsu y x w v z 2 1 3 5
c(x,x’) = cost of link (x,x’)
e.g., c(w,z) = 5
cost could always be 1, or
inversely related to bandwidth,
or inversely related to
congestion
cost of path (x1, x2, x3,…, xp) = c(x1,x2) + c(x2,x3) + … + c(xp-1,xp)
key question: what is the least-cost path between u and z ?
routing algorithm: algorithm that finds that least cost path
Network Layer: Control Plane

8. Routing algorithm classification
Q: global or decentralized information?
global:
all routers have complete topology, link cost info
“link state” algorithms
decentralized:
router knows physically-connected neighbors, link costs to neighbors
iterative process of computation, exchange of info with neighbors
“distance vector” algorithms
Q: static or dynamic?
static:
routes change slowly over time
dynamic:
routes change more quickly
periodic update
in response to link cost changes
Network Layer: Control Plane

9. Distance vector - RIP Each node maintains a table of triples.
table at B:
Destination
Cost
Next-hop A 1 C D 2 E F G 3
Network Layer

10. RIP: overview Iterative, asynchronous, distributed
Directly connected neighbors exchange updates
periodically (on the order of several seconds)
whenever table changes (called triggered update)
Each update is a vector of distances:
(Destination, Cost)
Update local table if receive a “better” route
smaller cost
came from next-hop
Refresh existing routes; delete if they time out
Network Layer

11. RIP: example Initial table at A: After receiving update from F:
Destination
Cost
Next-hop B 1 C D ∞ - E F G
After receiving update from F:
After receiving update from C:
Destination
Cost
Next-hop B 1 C D 2 E F G
Destination
Cost
Next-hop B 1 C D 2 E F G ∞ -
Network Layer

12. RIP: recovering from link failure
At F:
Dest
Cost Nh A 1 B 2 C D ∞ - E G
F receives update from A:
A receives update from C:
At A:
Dest
Cost Nh B 1 C D 2 E F G 3
Dest
Cost Nh A 1 B 2 C D 3 E G 4
Dest
Cost Nh B 1 C D 2 E F G ∞ -
Network Layer

13. RIP: link cost decreases1 Y 4 1 X Z 12 X 4 Z 1 X 1 Z X 1 Z
At Y: X 2 Y 1 X 5 Y 1 X 5 Y 1
At Z:
Good news travels fast
Network Layer

14. RIP: link cost increases14 Y 4 1 X Z 12 X 4 Z 1 X 6 Z 1 X 6 Z 1 X 8 Z 1
At Y: X 5 Y 1 X 5 Y 1 X 7 Y 1 X 7 Y 1
At Z:
and so on
Bad news travels slow
“count to infinity” problem  loops!
Network Layer

15. Breaking the loop … If next-hop to D is R:
Split Horizon: do not include D in update to R
Split Horizon with Poison Reverse: include D, but with metric = ∞ 14 Y 4 1 X Z 12 X 4 Z 1 X 14 Z 1 X 14 Z 1 X 13 Z 1
At Y: X 5 Y 1 X 5 Y 1 X 12 Y 1 X 12 Y 1
At Z:
Does this solve the “count to infinity” problem?
Network Layer

16. … is not always easy At A: C receives update from A:
Dest
Cost Nh B 1 C D 2 E ∞ - F G
A receives update from B:
C receives update from A:
B receives update from C:
Dest
Cost Nh B 1 C D 2 E 4 F G
Dest
Cost Nh A 1 B C D E 5 F 2 G
Dest
Cost Nh A 1 C D 2 E 3 F G
Network Layer

17. RIPv2 (RFC 2453) details Included in BSD-UNIX Distribution in 1982
Distance metric: # of hops (∞ = 16): why?
Distance vectors only exchanged among neighbors
Up to 25 destinations per RIP update message
Update-interval is 30 sec:
If too large, convergence is slow
If too small, too much traffic
Triggered update whenever change in routing table
Split horizon mandatory, poison reverse optional
Network Layer

18. RIPv2 details (contd.) Updates sent every 30 (+/- 5) seconds
Address of net 2
Distance to net 2
Command
Must be zero
Family of net 2
Family of net 1
Address of net 1
Distance to net 1
Version 8 16 31
Updates sent every 30 (+/- 5) seconds
Route not refreshed for 180 sec is timed-out
Still included in update messages
Timed-out route is deleted (garbage-collected) after 120 sec
Triggered update timer set for 1-5 sec
Includes only changed routes
Suppressed if regular update due
subnet mask of net 1
next hop of net 1
subnet mask of net 2
next hop of net 2
Network Layer

19. RIP: where does it run?
RIP runs as application-level process (route-d)
Updates sent as UDP message (port 520)
Multicast IP address (with TTL=1)
routed
routed
Transprt
(UDP)
Transprt
(UDP)
network forwarding
(IP) table
network
(IP)
forwarding
table
link
link
physical
physical
Network Layer

20. Link State - OSPF Strategy: each node learns complete topology
send information about directly connected links (not entire routing table) to entire network (not just neighbors)
Link State Advertisement (LSA) include
Nodes (routers) and links (networks)
Sequence number and age
Reliable flooding
Store most recent LSA for each node
Send LSA to all nodes except one that sent it
Generate LSA periodically (with higher sequence number)
Age out each stored LSA
Network Layer

21. A link-state routing algorithm
Dijkstra’s algorithm
net topology, link costs known to all nodes
accomplished via “link state broadcast”
all nodes have same info
computes least cost paths from one node (‘source”) to all other nodes
gives forwarding table for that node
iterative: after k iterations, know least cost path to k dest.’s
notation:
c(x,y): link cost from node x to y; = ∞ if not direct neighbors
D(v): current value of cost of path from source to dest. v
p(v): predecessor node along path from source to v
N': set of nodes whose least cost path definitively known
Network Layer: Control Plane

22. Dijsktra’s algorithm 1 Initialization: 2 N' = {u} 3 for all nodes v
if v adjacent to u
then D(v) = c(u,v)
else D(v) = ∞ 7
8 Loop
9 find w not in N' such that D(w) is a minimum
10 add w to N'
11 update D(v) for all v adjacent to w and not in N' :
D(v) = min( D(v), D(w) + c(w,v) )
13 /* new cost to v is either old cost to v or known
shortest path cost to w plus cost from w to v */
15 until all nodes in N'
Network Layer: Control Plane

23. Dijkstra’s algorithm: example
D(v)
p(v)
D(w)
p(w)
D(x)
p(x)
D(y)
p(y)
D(z)
p(z)
Step N' u ∞
7,u
3,u
5,u 1 uw ∞
11,w
6,w
5,u 2
uwx
14,x
11,w
6,w 3
uwxv
14,x
10,v 4
uwxvy
12,y w 3 4 v x u 5 7 y 8 z 2 9 5
uwxvyz
notes:
construct shortest path tree by tracing predecessor nodes
ties can exist (can be broken arbitrarily)
Network Layer: Control Plane

24. Dijkstra’s algorithm: another example
Step 1 2 3 4 5 N' u ux
uxy
uxyv
uxyvw
uxyvwz
D(v),p(v)
2,u
D(w),p(w)
5,u
4,x
3,y
D(x),p(x)
1,u
D(y),p(y) ∞
2,x
D(z),p(z) ∞
4,y u y x w v z 2 1 3 5
* Check out the online interactive exercises for more examples:
Network Layer: Control Plane

25. Dijkstra’s algorithm, discussion
algorithm complexity: n nodes
each iteration: need to check all nodes, w, not in N
n(n+1)/2 comparisons: O(n2)
more efficient implementations possible: O(nlogn)
oscillations possible:
e.g., support link cost equals amount of carried traffic: A A D C B
given these costs,
find new routing….
resulting in new costs A D C B
given these costs,
find new routing….
resulting in new costs A D C B
given these costs,
find new routing….
resulting in new costs 1
1+e
2+e
1+e 1
2+e
1+e 1
2+e
1+e 1 D B e C 1 1 e
initially
Network Layer: Control Plane

26. OSPF details RFC 2328 (244 pages long!)
Neighbor up/down detected using “hello” packets
LSA reliable flooding over entire AS
LSA includes sequence number and age
LSA integrity using checksum (excludes age)
OSPF messages directly over IP (no UDP or TCP)
Hierarchical OSPF: allow scaling to larger networks
5 types of LSAs:
Router LSA: set of nodes
Network LSA: set of links
Summary LSA: inter-area networks
Summary LSA: area-border-routers
External LSA: external to AS
Network Layer

27. OSPF “advanced” features
security: all OSPF messages authenticated (to prevent malicious intrusion)
multiple same-cost paths allowed (only one path in RIP)
for each link, multiple cost metrics for different TOS (e.g., satellite link cost set low for best effort ToS; high for real-time ToS)
integrated uni- and multi-cast support:
Multicast OSPF (MOSPF) uses same topology data base as OSPF
hierarchical OSPF in large domains.
Network Layer: Control Plane

28. Hierarchical OSPF backbone boundary router backbone router area border
routers
area 3
internal
routers
area 1
area 2
Network Layer: Control Plane

29. Hierarchical OSPF two-level hierarchy: local area, backbone.
link-state advertisements only in area
each nodes has detailed area topology; only know direction (shortest path) to nets in other areas.
area border routers: “summarize” distances to nets in own area, advertise to other Area Border routers.
backbone routers: run OSPF routing limited to backbone.
boundary routers: connect to other AS’es.
Network Layer: Control Plane

30. Comparison of LS and DV algorithms
message complexity
LS: with n nodes, E links, O(nE) msgs sent
DV: exchange between neighbors only
convergence time varies
speed of convergence
LS: O(n2) algorithm requires O(nE) msgs
may have oscillations
DV: convergence time varies
may be routing loops
count-to-infinity problem
robustness: what happens if router malfunctions?
LS:
node can advertise incorrect link cost
each node computes only its own table
DV:
DV node can advertise incorrect path cost
each node’s table used by others
error propagate thru network
Network Layer: Control Plane