Optimum Path Computation Algorithms for Multimedia and Real
Time Services over IP-Networks
Rajiv kumar, Yuvraj Singh, Varun Uppal, Mandeep Rawat and Ankit Bhatt Department of Electronics and Communication
Jaypee University of Informations and Technology, Waknaghat, Solan, HP
ABSTRACT: Traditional Routing Techniques over internet involves shortest path based on minimum distance, delay, etc. for various applications including E- mail, instant messaging, voice- over-ip, etc. Such a technique is based upon connection- less approach where multiple paths are formed between source and destination. Although such an approach supports faster and less delayed path along with less traffic congestion but, with rapid growth of multimedia applications such as telesurgery, live video conferencing and various other “mission critical” areas such connection- less approach dreadfully fails as time delay i.e. start- up delay, increases considerably reducing real time factor of multimedia applications. In this paper, a routing algorithm is proposed which draw parallels with QoS requirements of multimedia applications i.e. high bandwidth, minimum start-up delay. Comparative analysis of routing algorithm based upon less approach against proposed routing algorithm based on connection-oriented approach is also done via a simulation taking an example of a particular network.
Keywords— Traditional Routing Techniques, Multimedia Applications , Mission Critical, Routing Algorithm.
INTRODUCTION
ultimedia applications such as digital video and audio require a connection-oriented approach with high QoS requirements such as high bandwidth, minimum start -up delay, maximum throughput. A network has to pass simultaneously several such constraint in order to facilitate QoS requirements. In a traditional routing method e.g ARPANET (provides only connectivity based on distance), using connection-less approach a path is computed only on the basis of distance and delay. Such a connection less oriented approach is substituted by ours connection-oriented method where maximum bandwidth allocation is the prior factor in determining the route along with minimum number of hops [1, 2, 3, 8].
Connectionless means that no effort is made to set up a dedicated end-to-end connection. Connectionless communi-cation is usually achieved by transmitting information in one direction, from source to destination without checking to see if the destination is still there, or if it is prepared to receive the information. When there is little interference, and plenty of speed available, these systems work fine. In environments where there is difficulty transmitting to the destination, information may have to be re-transmitted several times before the complete message is received. It reduces delay time and effective cost. But causes a lot more problem when a safe and secure data transmission is required. Although data packets may get transferred from source to destination but always there are some packet loses at the receiver end. Various applications like radio might not need such a high accuracy and hence can be applied in connection less media. Walkie-talkies, or citizens band radios are a good examples of connectionless communication. You speak into the mike, and the radio
transmitter sends out your signal. IP, UDP, ICMP, DNS, TFTP and SNMP are examples of connectionless protocols in use on the Internet [8].
Connection-Oriented means that when devices communicate, they perform handshaking to set up an end-to-end connection. Such a technique is the most secure and guarantees 100% transmission of data. This kind of accuracy and guarantee is required by various applications termed as “mission critical”, where dependency on the system is very high and margin of error reduces to near zero. This includes along with various real time and multimedia applications involving space communication, telesurgery, and teleconferencing [8].
PATH COMPUTATION ALGORITHMS
Routing algorithm is a logic a router uses to decide, for each incoming packet, which output link, the packet should be transmitted on. Desirable properties of a routing algorithm: Correctness, Simplicity, Efficiency, Robustness Stability, Fairness , Optimality. The two most common routing algorithms are distance-vector and link-state.
In a distance-vector routing each router exchanges information about the entire network with neighboring routers at regular intervals . Neighboring routers are connected by a direct link (e.g. a LAN) which exchanges in formation at regular intervals e.g. every 30 seconds.
In a Link -state, each router exchanges information about its neighborhood with all routers in the network when there is a change . Here the n eighbourhood of a router is a set of neighbor routers n, where each router’s neighborhood information is flooded through the network.
To calculate its routing table, a router uses Dijkstra’s Shortest-Path algorithm [6, 9].
Dijkstra Algorithm
T his algorithm builds a shortest -path spanning tree for the router: such a tree has a route to all possible destinations, and no loops. First it identifies all link costs in the network: either from the link-state database or using the fact that the cost of any link from a network to a router is 0. The router running the algorithm is the root of its shortest-path spanning tree . Even if all routers’ link-state databases are identical, the trees determined by the routers are different (since the root of each tree is different). A node is either a network or a router; nodes are connected by arcs. The algorithm keeps track of 2 sets of nodes and arcs— Temporary and Permanent . Initially, the Temporary set contains all neighbor nodes of the router itself, and the arcs connecting them to the router; only the router is initially permanent . When all nodes and arcs are in the permanent set, the algorithm has terminated. Then it identifies the temporary node whose arc has the lowest cumulative cost from the root: this node and arc are moved into the permanent set ; any nodes which are connected to the new permanent node and are not already in the t emporary set, along with the connecting arcs, are made t emporary. Also, if any node already in the temporary set has a lower cumulative cost from the root by using a route passing through the new permanent node, then this new route replaces the existing one; repeat until all nodes and arcs are permanent.
The Dijkstra’s shortest path algorithm is the most commonly used to solve the single source shortest path problem today. This algorithm is the basis of computations of shortest path according to minimum distance, minimum delay and maximum bandwidth.
For a graph G (V, E), where V is the set of vertices and E is the set of edges,
Dijkstra’s shortest path algorithm for each u belongs to G:
Start
1. d[u] = infinity; 2. parent[u] = NIL; 3. End for
4. d[s] = 0; // s is the start point 5. H = {s}; // the heap
6. while NotEmpty(H) and targetNotFound: 7. u = Extract_Min(H);
8. label u as examined; 9. for each v adjacent to u: 10. if d[v] > d[u] + w[u, v]: 11. d[v] = d[u] + w[u, v]; 12. parent[v] = u; 13. DecreaseKey[v, H]; End
Using the above dijkstra algorithm, three programs were developed in the C++ environment and used for simulation to compute best suitable path according to certain requirements like bandwidth, delay , no. of hops etc. Snapshots of the output of these programs were taken and are shown in this paper.
A GIVEN UNIDIRECTIONAL NETWORK
TOPOLOGY WITH VARIOUS PARAMETERS
Given in the figure is a network topology with 12 nodes and 24 links between the nodes. Source is the node no. 9 and the destination is node no. 2. It is required to compute the optimum path between the source and the destination. Simulation results are computed and shown with the help of programs. Fig. 1 2 1 3 4 6 5 7 8 9 10 11 12 (Destination) (Source)Table: Network Link Metrics
S. No. Link [u,v] Distance [D(u,v)] Bandwidth(in Mbps) [c] Delay(in secs) [d(u,v)]
1. 12,1 108 5 10 2. 11,1 50 4 13 3. 5,1 30 5 13 4. 1,2 30 6 4 5. 11,2 108 4 20 6. 5,2 50 5 14 7. 3,2 50 6 14 8. 10,2 50 7 15 9. 7,2 30 4 17 10. 5,3 50 3 9 11. 4,3 108 4 4 12. 6,4 50 5 34 13. 4,5 108 6 21 14. 11,5 50 8 19 15. 12,5 50 9 18 16. 8,5 30 2 3 17. 7,6 50 4 34 18. 9,8 50 5 16 19. 8,10 50 6 15 20. 9,10 50 3 13 21. 6,10 50 4 13 22. 10,11 50 5 12 23. 6,11 50 3 15 24. 11,12 50 3 19
Computing Parameters
Various QoS parameters such as delay, bandwidth, number of hops, capacity, etc are considered in this paper while computing the optimum path between two nodes. For the connectionles s approach, we have used minimum distance, minimum delay & bandwidth as the criteria for optimization. For the connection oriented approach, we have used maximum bandwidth criteria along with minimum number of hops.
A ssociated weights of delay
w
d, bandwidthw
c, no. of hopsw
h are computed for each case. For the case of connection oriented approachw
cis the is the minimum capacity in the whole path. The capacity of link less thanc
w
is not used for computing the path. For the case of connection lessw
c can be derived as follows: An equivalent of whole path is assumed to beC
eq.As we know that, eq
C
σ
= Total Delay. Hence the net value of [ ( , ) ] ( , ) = +
∑
eq C d u v c u v σ σ (1)Connection-Less Routing Algorithm Shortest Path According to Distance Only
Below is the snap shot of the program based on Dijkstra Algorithm for computing minimum distance path between two points in network taken above taking a single parameter of distance [D(u,v)] between two nodes. a shortest len gth is hence calculated between source node no. 9 to node no. 2. The path is 9..10..2 and minimum distance is 100. This route is often changed as connectionless approach might require. Depending on traffic, the shortest path is computed taking new path where distance parameter is minimum.
This routing algorithm uses Dijkstra algorithm to find best path depending on minimum delay (d(u,v) in sec). This program computes minimum delayed path according to bandwidth (c(u,v)) and delay (d(u,v)) linked with each node. The best path is calculated on the basis of traffic to data to be sent (
σ
= 5 mbps).Here the for mula used for minimum delay is = ( , )[ ( , )] ( , )+
∑
u v d u v c u v σ (2)As said above, the path would change according to the traffic of data to be send. Here the best path is calculated between node no. 9 to node no. 2 and minimum delay is 31 seconds.
Simulation no. 2: optimum path accor ding to delay and bandwidth
Connection-Oriented Routing Algorithm Shortest Path According to Bandwidth and Number of Hops
As the path is dedicated one there is no loss of data packets here. Only criteria for above program algorithm is the maximum bandwidth. This is calculated on the basis of traffic of data to be sent. If under any circumstances any node is witnessed with bandwidth less than traffic of data (σ) then a condition called bottle neck is reached. Under such a condition the node (u, v) with less bandwidth than the traffic (σ) would be truncated or the different path would be chosen removing the node (u, v) from the decision tree. After that path with minimum no. of hops is selected from source to destination. As shown below for traffic (σ) 4 mbps shortest path is selected from nod e no. 9 to node no. 2 in minimum of three hops.
Simulation no. 3: optimum path according to bandwidth Above is the comparative analysis of the various parameters taken separately in the above three different cases. The following inference were drawn:
1. High Bandwidth requirement for the multimedia applications is only satisfied by the last row where the value of equivalent bandwidth is highest among the three values.
2. Although the value of distance, no. of hops and delay is far less in case of first two cases still for dedicated applications only bandwidth value reigns more than all other parameters to give a secure and guarantee stable communication.
3. Also the path was changed between node no. 9 and node no. 10 as “bottle neck” was encountered as the bandwidth of between 9 and 10 is less as compared to traffic requirement.
c(9, 10) = 3 <
σ
= 4CONCLUSION
By using the simulation programs it is shown that for the multimedia and real-time application over ip-networks, it was mandatory to use the bandwidth efficiently. The traditional routing method which depended only on delay and distance has proven incompetent for high bandwidth requirement applications. Also shown is the adaptive routing, where route is changed in accordance with the traffic load.
A COMPARISON
CASE NO. ROUTE TRAFFIC (σ
) IN Mbps DISTANCE (WD) DELAY (Wd) IN Secs BANDWIDTH (WC) IN Mbps NO. OF HOPS 1. 9..10..2 3 100 30 0.102 2 2. 9..10..2 5 100 31 0.164 2 3. 9..8..10..2 4 150 49 5 3REFERENCES
[1] Mohamed, F. Mokbel, Wafaa, A. Haweet and Nazih El-Derini, M., “A Delay-Constrained Shortest Path Algorithm for Multicast Routing in Multimedia Applications” Computer Networks 47 (2005) 87–104, 11 September 2004. [2] Wei Liu, Wenjing Lou and Yuguang Fang, “An efficient
quality of service routing algorithm for delay- sensitive applications”.
[3] Zheng Wang and Jon Crowcroft, “Quality of Service Routing for Supporting Multimedia Applications”.
[4] Samphel Norden and Jonathan Turner, “Inter-Domain QoS Routing Algo rithms”.
[5] Hussein, F., Salama Douglas, S. and Reeves Yannis Viniotis, “Evaluation of Multicast Routing Algorithms for Real-Time Communication on High - Speed Networks”
[6] Carleton University School of Computer Science, “Fast Shortest Path Algorithm for Road Network and Implementation” HONOURS PROJECTFall Term, 2005. [7] Ciro A. Jr and Fouad Tobagi “Evaluation of multicast
routing algorithms for multimedia streams” Technical Report No.CSL-T R-94-619, April 1994.\
[8] Tanenbaum ”Computer Networks”.
[9] Jain, V.K. and Gopal, K., An Algorithm for Determining the Most Reliable path of a Network, Microelectronics Reliability, Vol. 26, pp. 883–884, 1986.