EFFECT OF VARYING TIMERS OF OSPF ON NETWORK TRAFFIC

Supervisor Dr. S. C. Gupta

BY ANIMESH SRIVASTAVA SHUBHAM BARANWAL SNEH BHARATI

Saturday, August 11, 2012

AGENDA
Routing Introduction to OSPF Understanding business flows Build consensus with Global IT and OPS Team Use of New Technology

Saturday, August 11, 2012

ROUTING
Routing is the process of forwarding data packets from source network to the destination network which is done through routers. Routers determine the path to a network through the use of routing tables which list known networks. Routing protocols running on different routers exchange information about the network and automatically populate routing tables. Routing metric defines the cost of a route in terms of how long it takes to deliver packets through a particular path. Algorithms that the routers run among them in order to make the correct routing decisions are called routing protocols
Saturday, August 11, 2012 3

ROUTING PROTOCOLS
Distance Vector Routing: Distance between network segments is measured as number of routers a packet must cross to travel from one segment to anotherhop count Router does not know the complete network topology just the direction to which packets have to be forwarded and the distance to the destination network Routers send their routing tables to every neighboring router Link State Routing : Every router has its own map of complete network Each router sends status messages to every other router of the network periodically Status messages are used to build network topology

Saturday, August 11, 2012

OSPF (Open shortest path first)

OSPF PACKETS
To exchange information OSPF uses five packets types Type 1 - Hello Type 2 Database Description Packet - DBD Type 3 Link State Request Type 4 Link State Update - LSU Type 5 Link State Acknowledgement - LSACK
Saturday, August 11, 2012 6

FINDING NEIGHBORS
Each router establishes a relationship (adjacency) with its neighbors Each router generates link state advertisements (LSAs) which are distributed to all routers LSA = (link id, state of the link, cost, neighbors of the link) Each router maintains a database of all received LSAs (link state database), which describes the network as a graph with weighted edges Each router uses its link state database to run the shortest path algorithm (Dijikstras algorithm) to produce the shortest path to each network
Saturday, August 11, 2012 7

OBJECTIVE
Our work is a trade-off between faster failure detection and bandwidth usage.

Subnet LAN_1
Saturday, August 11, 2012 8

OBJECTIVE
Links: Within a subnet, a100-T LAN is connected to its router via 100-T link. The router in a subnet is connected to its backbone via 100-T link. Backbone routers are connected to each other via PP_DS0_int link. Backbone 2 is connected to server via 10Gbps_Ethernet line. Parameters to change: Hello interval time Router dead interval time Simulation Output: Global Http traffic Global OSPF data traffic Application Definition: HTTP Heavy Browsing. The Server caters to HTTP requests from the 3 LANs. All IP addresses are auto-assigned.
9

Saturday, August 11, 2012

ANALYSIS
Scenario 1a:Hello Interval = 10 sec Dead Interval =40 sec The starting time for the HTTP profile is constant and set to 100 seconds.

This is the time required by OSPF to set up the routes and do its computations. These 100 seconds is the duration of the peak as seen in the OSPF traffic graph.

Saturday, August 11, 2012

10

ANALYSIS
Scenario 1b:Hello Interval = 5 sec Dead Interval =40 sec Http traffic Nearly same as Http traffic of Scenario 1a (10,40) OSPF traffic First 100 seconds are nearly same The number of updates has increased significantly Traffic usage has approximately doubled from 1250 bits/sec to 2500 bits/sec
Saturday, August 11, 2012 11

ANALYSIS
Scenario 1c: Hello interval=5 sec Dead interval=40 sec
First 100 seconds are same
Slight increase in OSPF traffic from 1250 in Sc. 1a to 1300 OSPF updates are not regular as in Sc.1a No. of updates have increased

Saturday, August 11, 2012

12

ANALYSIS
Scenario 1d: Hello interval=20 sec Dead interval=40 sec No. of updates have decreased significantly in comparison to previous results. OSPF traffic has decreased to around 1000 bits/sec

Saturday, August 11, 2012

13

ANALYSIS
Scenario 1e: Hello interval=30 sec Dead interval=15 sec Hello Interval < Dead Interval No Http traffic As the Hello Interval is increased to 30 seconds , OSPF traffic has decreased to 800 bits/sec

Saturday, August 11, 2012

14

ANALYSIS
Scenario 2a: Link failure Hello interval=10 sec Dead Interval =40 sec
Backbone 1 Backbone 3 400s Fail Backbone 1 Backbone 3 800s Recover Backbone 2 Backbone 4 600s Fail Backbone 2 Backbone 4 1000s Recover for each change in the link state, the bits per second of routing updates increases gradually. These updates are sent at intervals which are multiples of the Hello Interval.
15

Saturday, August 11, 2012

ANALYSIS
Scenario 2b: link Failure (5,40)
Scenario 2c: link Failure (20,40)

Saturday, August 11, 2012

16

ANALYSIS
Scenario 2d: link Failure (30,15)
Same as Sc.1e : No link failure(30,15)
Only2 updates whereas there are 4 link-state changes.

Saturday, August 11, 2012

17

CONCLUSION
Considering the dead time interval constant, the data sent over the network decreases with increase in hello-interval. The increase in hello-interval time delays the detection of the network failure. Hence depending on the topology and bandwidth available, a suitable value of hello interval and dead interval can be used. Future work: To find out how topology plays a role in convergence time.

Saturday, August 11, 2012

18