3. CBWRED – Configuration
Now in this video we’ll see the configuration of class based weighted trade based on the IP presence values. Now I have designed a small lap where we are going to see how to configure the weighted rate based on the IP presence values here. Now here I’m going to assume that I got a traffic which is coming with a marking of zero and one and I want to ensure that any traffic coming with a zero and one should get a guaranteed bandwidth of 20%. And then after that I want to enable the weighted red for that particular traffic and I’m going to define the minimum and the maximum threshold values based on the default parameters.
Again, the similar way, I got another traffic with two n three with a marking of IP presence value of two n three and it should get guaranteed bandwidth of 25% and and the presence value with four should get a guaranteed bandwidth of 30%. And after that we need to enable the random detect for all the traffic coming with the marking of zero and one. Now, based on the default parameters, we are going to use simply default weighted rate parameters and also the drop probability denominator as well for individual specific markings. Now let’s see how the configuration goes. So the first configuration is we need to create three different class maps and one class map is going to match the class zero and the one that is a presence value of zero and one.
And another class maps is going to match the presence value of two and three. And the third class map is going to match the presence value of four. So let’s create some class maps. First the class maps I’m going to use presence value of zero one. So let me use something like this classmap zero and one, let me remove the previous one. So we’re going to create a class map of zero and one and I’m going to match the IP presents value of zero comma one. And then I’m going to create a second class map with two and three where I’m going to match the presence value of two comma three. And the third class map, I’m going to use this name as president four and I’m going to create the presence value of four.
So here the name of the class map can be anything. I’m going to create three different class maps. If you verify show run class map, you can see there are three class map created, presidents one, two and three. Now the next thing we need to create a policy map and in that policy map we need to define, let’s say the name of the policy cci. I’m going to call the first class map. And in the first class map, first thing I’m going to do the bandwidth reservation. And as per the requirement, the bandwidth reservation should be 20% and I’m going to define bandwidth percent 20. And after that I’m going to enable the random detect, random detect for all the remain once it reaches the limit.
Now the next thing I’m going to define the random detect. Again, if you use question mark you’ll find some multiple options whether you want to do the random detection based on the dcp values or based on the presence values. If I do not define anything, it’s going to do based on the president’s values. So the default presidents, you can see presidents. Now we can say presidents based, like, if you want to change, if I don’t give this command, it’s going to do any way. IP presidents based values. And then the next thing we can define what should be the president’s value here according to our requirement for zero and the one for the zero president’s value, what should be the minimum threshold value now, as per the default president’s values here.
Now the major difference here is like if you take the default president’s values here, if you’re using the president’s value of two, the default president’s value is 20 and for president’s value one it is 22 20. It will go on something like this. Now the maximum facehold value will be always 40 for all the different classes here. Now as the question as the requirement here is we are going to use the default values. So I’m going to say the default president’s value, the minimum threshold value for the president zero is something 20, it’s 20 default and I’m using the default value 20 packets and the maximum default value it’s 40. And then it’s going to ask the mark denominity denominator.
Now mark probability denominator defines how many packets should be dropped out of ten and the default is ten. If I say ten out of ten packets is going to drop one packet randomly. Okay, once it reaches the limit. Now I’m just going with the default values, we can still change this values as per our requirement and I’m going to change for president’s value of one, as the quotient says, we are going to use the default values. So in order to understand the default values, I’m just writing this for president’s value of one, it is going to be 22. Now we just verify the commands. I have created a class for the presence value of zero and one.
We have reserved 20% of the bandwidth and in case if there is any tail drop to happen, it’s going to random detect that and before it reaches the limit, once it once if a marking is coming based on president’s value of 20, it’s going to be 20 means 20 packets. And once it goes between 20 and 40, it starts dropping randomly. For the president’s value of one, it will be 22 to 40 randomly, one packet will be dropped out of ten a similar way. Let’s do the same confirmation for the second class. Second class, I think I used two, three and I’m going to define the bandwidth percent 25 and then we are going to say random detect presidents. Now the president’s marking will be two and three.
So for two for two the default value will be 24 24 the minimum chisol value and the maximum chisel will be 40 at ten. Now we need to enable random detect first before I enable that command. And then for the presence value of three the default minimum chisel will be 26 and the maximum disorder will be 41. Packet out of ten packets will be drawn, that’s what we call as mark probability denominator. And then for the third class presents value of four and we are going to say bandwidth percent, it’s going to be 30%. And then we are going to say random detect and then I’m going to say the presence value is four and the minimum switch forward will be default 28 and then all the remaining values are same.
So if you just verify Show run policy Map to verify my configurations. Now we have created three different class maps and we have configured the bandwidth guaranteed. And if the queue is full then it’s going to do the random lead, it’s going to drop the packets based on the president’s values and this is the minimum and the maximum choice for different different markings. Now we can even change this marking as per the requirement but even if you define the default parameters that will be sufficient. Now once you are done with this, now if you verify Show policy map now I can see the configuration here. If I use Show policy map now here you can see the default values.
We are going to mark the class presidents of one. It’s going to mark these two values and this is the one out of ten packets here. This is a minimum and the maximum result values for the second class two and three. It’s going to do for this two, one. And for the third class class presents value of four. It’s going to mark the presence value of four with a minimum 28 and maximum threshold value of 40. Now this is how the configuration goes in general if you want to enable the class based weighted randomly detection. And then finally if you want to apply this on the interface, either it is inbound or outbound. We need to say simply service policy output and then the name of the policy on that.
If you verify Show policy map interface s one by zero. Now you’ll see almost a similar kind of output here. We just maximize the screen here. You can see the options here. If I give Show policy map here, show Policy Map interface is going to show you how many packets or bytes have been transferred. Random drop and tail drop all the information here, minimum and the maximum result value for each and every class. Now here we don’t have a real time traffic going. If you find any real time traffic going. You will see the packets will be matched and you’ll see how many packets have been sent or randomly dropped. More on that. Okay? Now if you want to change the default parameters, we can still go ahead and change these parameters.
Like if you see here, I got a task here where you can specifically define with a presence value of zero, we can even define the queue size, the default for the presence value of zero, we have defined the minimum syslode is 20 and the maximum result is 40. If you want, we can change those parameters. Now that is something what I did in the next task, changing is very simple. We just need to go to the command line and we are going to remove the default values, whatever we have configured in the previous section, adding no command. And then we are going to change the values based on the requirement. Now, this is something more applicable if you have some questions tested in your cci exams where you will be asked to do the random drop based on the minimum.
And the maximum threshold values. So if not, you can just simply go with the default values based on the default presence markings. Now, if you want, we can even change this marking values based on these cp values as well, like in this section. Again, you can simply remove the policy. What we have configured in the previous lab, we can also configure the same thing exactly. Instead of matching based on the IP presence values, we can use the dhcp based values as well. Now, the configuration goes exactly the same, there’s no much difference except we are going to create a class map which is going to match the dscp values instead of matching the president’s values.
Like here, I took an example where I’m going to match the dscp marking of af two three, af three two and af 41. And then I have defined some different parameters for different maps here. Like if we just come down here where I’m going to call the first class map and the bandwidth guarantee and this command is a little bit mandatory, we need to tell because if you don’t say random detect, dhcp based, the default will be IP presence based random detect. And then we need to define what is the minimum threshold,maximum threshold and the mark probability denominator for this class. And similarly for the remaining classes, we can change and we can simply go to the interface and implement.
Now in this actually we have seen two different configurations where we can do the class based weighted fairping, based on the IP presence values or based on the dsp values. It’s all about your requirement. You need to know the exact traffic and the marking values and we are going to match those marking values inside the class maps and we are going to define the guaranteed bandwidth. That is what we are going to do. And along with that in case it reaches the limit if there is a congestion and if your Q is full in that case we are defining to drop where we are going to define the minimum and the maximum result values and the mark probability denominator.
4. Traffic Policing-Shaping
Now, in this video we’ll talk about traffic policing and shipping. Now, traffic polishing is a method of limiting the number of packets rate. And we can define the upper limit of the packet rate. And once it reaches that limit, whatever the exceeding traffic, we can either drop traffic or we can send the traffic. Or we can mark that particular excess traffic with some low priority marking. Something like that. Policing is a method which is where we are going to define the rate limit and if it exceeds, probably we either can drop the traffic or we can do the remarking of that particular traffic and the shaping is slightly different from the policing, whereas we can define a limit and anything exceeding will be stored in the buffer rather than getting dropped.
Now, take an example. We have a link. Let’s say I have a router and this router is connecting to a service portal router. And on the service folder I have taken a five mvps line and the service folder guarantees that the link connecting to the router minimum you get five mbps. And that is what we call as cir committed information rate. And we are paying for the five mbps. And this limiting of this bit rate is either done by the service portal by using traffic policing or shipping. So maybe you are connecting to the router and you’re using some fast ethernet link and if I’m using fast ethernet link it is capable of sending and receiving 100 vps traffic. But the service portal will ensure that on this particular link you can only utilize only five mvps of traffic at one time.
So even though I’m going to take 100 mvps link but only I can send and receive at a speed of five MPs, that is something guaranteed by the service model and this is something done by either policing or shaping. Now, policing and shaping is going to define a limit of five EPS traffic and if it exceeds five EPS probably we can either define to drop it or to remark it or it can be configured as whatever exceeding can be delayed by using some buffers. So this is something done by two different mechanism. That’s what policing and shaping. Now, let’s talk about more in detail on the policing options here. Now, the policing is a method of limiting the packet where we can define the maximum upper limit of that particular packet, anything exceeding that will be automatically dropped.
And what traffic policing is going to do is it is going to measure the byte rate of the arriving packets and it’s going to see whether, let’s say the committed information, it is five aps. And if it exceeds the five epps, either we can define the parameters to allow the packet exceeding that or that’s what we can say. Confirm, transmit. Now there is something called confirm. Transmit. If it is less than five MPs, I can simply say transmit the traffic if it exceeds five MPs, probably I can define a policy to drop the traffic or even we can define instead of dropping we can simply say anything exceeding above five appears should be sent with a lower president’s values or lower preferred marking like zero or one, or it should be treated as a normal traffic than a priority traffic.
Now, we can define some three options like confirming. Confirming is something when we define five EPS. If it is within this five epps limit, it’s going to allow the particular traffic to forward as long as within this limit we can define something called exceeding traffic so exceeding its above five mvps and then we have an option of violet. Now, these are three different options. We have something called confirming, exceeding and violet. The next thing we’ll talk about token bucket. Now, each single token bucket refers to one byte. It’s something we can say single rate of transfer at one time. Now, each time when a packet arrives on the policer, it is going to put some tokens in the bucket. And this bucket size, it depends upon policing configured on that particular router.
Each time when a packet is policed, the policer will put some tokens into the bucket. Now, that is something calculated based on the packet arrival time and the previous arrival time multiplied by policing value divided by it. Now, let’s take an example. The packet arrival time is 1 second and it’s going to compare and previous arrival time. Let’s say this is the first time it has arrived and we are going to define the policing as 128 kbps. So it’s going to be 128 divided by eight total. It’s going to make it as 16,000 bytes. So 128,000. Now, this is the size of the token bucket. And when the packet arrives to the policer, if the number of packets is less than that particular number of tokens inside the router, it’s going to forward the traffic like here. Take an example.
The packet, the 700 bytes is the size of the token bucket and I’m sending some 500 bytes of information as it is within this limit. It’s going to say that it’s going to transmit the traffic. But whereas once it exceeds like in this scenario here, now there are 200 bytes of token remaining and if you’re receiving 300 bytes of information as it is exceeding, it’s going to automatically drop here. Now, I got a sample configuration here, a basic example, which is going to just give some idea on how the policing configuration goes. It’s going to be based on the classmass, what we are using here. So let’s take an example. I have a rule that I want to match this icmp and Http traffic and I want to ensure that my icmp and Http traffic should be policed up to 512 kbps.
And as long as within this 512 kbps it has to transmit. And if it exists, we can define an action called exceed action drop and the command starts with Police police file 12,000 you can see 5120 zero and confirm action is going to transmit and anything exceeding it is going to drop. Now similar way, maybe I have another rule where I want to ensure that all the tennet and snmp traffic smtp traffic should be police to 256 kbps. And if it is within the 256 kbps it has to transmit and anything exceeding tad has to be dropped. And we can apply either inbound or on the outbound of the inter a face again. So the configuration wise it’s going to be same. The only difference is inside the policy map the command starts with police and then you can always use the iOS help to figure out the next possible options.
