 Hello, everyone. Welcome to the session on Flow-Based Routing. I am Mrs. Kanchan Khedikar, working as an assistant professor in Computer Science and Engineering Department, a Fulcrum Institute of Technology, Solapur. Let us see the learning outcome. At the end of this session, student will be able to illustrate the Flow-Based Routing algorithm and they will also compare it with Shortest Path Routing algorithm. So, let us start Flow-Based Routing algorithm. It is the non-adaptive routing algorithm which is also known as static algorithm. Static algorithm that means in that algorithm the router set the path and these path are not changed. That means we can say that the path are fixed from source to destination. It uses topology and load condition for deciding the route. So, load condition that means what? There is a traffic from source to destination. Now, let me explain with this subnet. Now, in this subnet there are total 8 routers A, B, C, D, E, F, G, H and there are communication link and all these routers are connected with the communication link. Now, suppose this is your home. So, the source is your home that is A and you want to reach to your college that is destination C. So, there are so many path from A to C. So, this is first path that is from A to B, B to C. This is another path from A to G, G to E, E to F and F to C and one more path is there. So, likewise there are some more path from source A to destination C. Now, suppose you selected a shortest path and let us consider the shortest path is from A to B and B to C. But something went wrong there or because of some accident or because of some road condition if the traffic is more in that path. So, rather selecting this path means from A to B, B to C and stuck into a traffic you can choose the another path so that you will reach to your college properly. So, in flow based routing algorithm instead of selecting the shortest path it chooses the another shortest path from source to destination. So, let us see this flow based routing algorithm in detail. So, it can estimate the flow between all pairs of routers. Now, the basic idea behind the analysis is that for a given line if the capacity and the average flow are known then it is possible to compute the mean packet delay online. So, here flow based routing algorithm calculates the mean packet delay. So, from the mean packet delay on all the lines it is straightforward to calculate the flow weighted average and to get the mean packet delay for the whole subnet. So, let's see how to calculate the mean packet delay. To use this technique certain information must be known in advance. So, there are three important point or we can say that there are three important information we must know. The first one is subnet topology. Another thing is traffic matrix which is represented by F ij, i is the source and j is the destination. Then the third one is line capacity matrix which is represented as C ij. So, again i is here source and j is destination. This line capacity is nothing but the bandwidth of each line. Now, look at this first point that is subnet topology. So, here consider the full duplex subnet. So, full duplex subnet that means what? The traffic or the flow of transmission or the flow of data from source to destination and from destination to source. So, the same load or the same traffic is there from A to B and B to A. So, that's why what we are saying the subnet is full duplex subnet. So, in this subnet again there are total six routers A, B, C, D, E and F and all these six routers are connected with the eight communication line and these communication lines are shown that are these communication line are nothing but the line capacity or this communication line shows the line capacity. So, here line capacity which is represented as what? C ij and these are measured in terms of Kbps kilobits per second. So, now here in this subnet the line capacity is 20 Kbps from A to B. So, likewise on another line also like from B to C, C to D. So, line capacities are given. Now, the next one is the next important point is here traffic matrix which is represented as Fij. The matrix has an entity for each source destination pair and all the traffic is symmetric. So, now look at this matrix. In this matrix here on this side there is a destination from A to F and on this side or we can say that on vertical side there are the source from A to F. So, A to A traffic is 0 that's why this blank is shown over here. Now, from B to A, B to A the traffic is shown and the line capacity or here the line capacity is already shown on what that subnet that is 20. Now, here the number of packets are transmitted from B to A are what 9. So, here 9 packets per second go from B to A and that is represented by what lambda i. So, similarly B to A, C to A, then D to A, then E to A, then F to A. So, likewise this matrix is formed here. Then what I said here all the traffic is symmetric. That means from B to A how many number of packets are transmitted? 9, 9 packets per second. Similarly from A to B also 9 packets are transmitted. So, that's why I am saying the traffic is symmetric ok. Now, let's see the next point. So, here to calculate the mean time delay right. So, here to calculate this mean time delay we are using queuing theory formula. So, need to use the formula for time delay and that formula is t equal to 1 divided by mu C minus lambda. So, let's see 1 by mu C equal to mean packet size in bit. Then lambda indicates the mean flow in packets per second. Then C equal to line capacity which is measured in terms of bits per second, t equal to delay time and assume mean packet size of 1 by mu equal to 800 bits. Because we are going to calculate the delay time for mean packet size that is 800 bits. Now pause the video and write down the answer of this question. Compare a flow based routing algorithm with shortest path routing algorithm and the answer is both are non-adaptive routing algorithm. That means sometimes it is known as static algorithm. That means the path are already fixed from router 1 to router 2, 2 to 3 likewise. Shortest path algorithm uses only topology for deciding the route. But flow based uses the topology as well as load condition for deciding the route. That means it consider the topology as well as it consider the traffic present on that route. By calculating the delay time of each line, second shortest route is selected in flow based routing algorithm, so that packet reaches to the destination properly. Now let's continue the another topic. Now look at this table. This table shows the analysis of the subnet. Just now we have seen the subnet in which total 6 routers are present and there are 8 communication lines and on that communication line load is given. That means what the line capacity is also given as well as lambda, the value of lambda that means how many number of packets are transmitted per second is also given. So what we have to calculate? We have to calculate the delay time. We have to calculate the delay time as well as the weight on that line. So now look at here in this line. So total 8 number of lines are there. Now the reverse traffic is also same. So suppose on line AB, the lambda is 9 and C that is line capacity is 20, then we have to calculate the mu C. So here the analysis of the subnet using the mean packet size of 800 bits we are taking and the reverse traffic that means what from BA or from CB etc is the same as the forward traffic. That means whatever we are calculating here time delay, the same time delay is for the reverse traffic. So let's see how that mu C comes as a 25 packets per second. Now look at this formula here mu C equal to 20 divided by 800. So that is what that is 0.025 kbps per second. Now convert it into packets per second. So that is what 25 packets per second. So likewise I have calculated all the values for all the lines that is the mu C value for all 8 lines and the reverse traffic is also same. Now after that calculate the time delay which is represented as Ti and it is measured in terms of millisecond. So how this 63 comes? Now look at this formula. So here Ti equal to 1 divided by mu C minus lambda. So it is 1 divided by 25 minus 9. So what is 25 minus 9? That is 16. So 1 divided by 16 which is nothing but this 0.0625 seconds. Now convert it into milliseconds. It will gives you 62.5 milliseconds. So approximately 63 milliseconds. So the time delay on the line AB or BA is 63 milliseconds. So likewise other time is also calculated right. Now look at the weight. So here weight equal to 9 divided by 43. So what is this 43? 43 is the addition of all these lambda value from all the lines that is 43. So which is nothing but this 0.29. So likewise all the values are calculated over here. Now look at the delay time the maximum delay time. So maximum delay time is here 118 which is on the line FD. Then the second maximum delay time is 105 which is on the line CD. So in this way we have to calculate the delay time on each line right. So the conclusion is this calculation can done offline in advance and the other than the shortest path is used for the transmission and finally the line having the maximum time delay will not use for the packet transmission and due to huge traffic present in the network or on that line. So because of that this line is not chosen by the router source router. So this is flow based routing algorithm. So in this video we have seen the flow based routing algorithm as well as the comparison between the flow base and the shortest path routing algorithm. These are the references. Thank you.