1. Classification & Marking
Now, in this video we’ll see some basic introduction to the classification and the marketing. Classification and a marking is a basic feature of the quality of service. Now, if there is classification is a method of differentiating the traffic in different classes. Like let’s say you have some voice traffic coming in your network, voice and video traffic, which has to be a high priority traffic. And then there are some other low priority traffic like ftp and Http as well, and maybe have some financial transactions or some database servers which need some medium kind of priority. Now, we need to ensure that we classify this traffic in different categories and that is what we call as classification. And this can be done with some class maps where we are going to use some classmaps where we differentiate each and every traffic by using some different parameters.
Like we can either match the traffic based on the input interface, like we have two interfaces, we have F zero by zero and F zero by one. And you want to match all the traffic from you from F zero by zero. That is something we can do. Or we can match based on some marking values. We’ll talk about markings in our next section more in detail. Or we can also match based on the source and destination addresses where we can write some external seals. Or we can also match based on application as well. Like we have an option of match protocol Http or match protocol ftp like that. So classification is a method where we are going to differentiate the traffic based on our priority levels.
Now, if there is no classification, then all the traffic will be treated equally, which means it’s going to forward based on normal forwarding, where all the traffic will be treated equally if there is no classification. So, to classify the traffic, we can use something called class maps. If you remember in the previous section we have seen with a basic example where we have created an icmp traffic which is categorized in one class map and then we got some Http traffic in a separate category. We can create a class map where we can match, we can also create an acl actually to match any specific traffic based on the source and destination. And we can call this particular classmap acl inside this classmap here.Now, the next thing is marking.
Marking is a method like once we define different classes, like we have matched Http traffic, also we have matched some icmp traffic. Now we need to do some marking to that. So marking is something like turning to the packets where we are doing some marking as it goes over the network, so that it can be identified on the remote devices. What we can do is we can match a specific traffic and we can apply some marking value. So let’s say the marking is ready here something, and as it moves on the network on the remote routers we can match. Now, based on that particular matching marking values, either we can match that particular marking value or each and every device is going to treat the priority of that particular traffic based on the marking.
Now, we have some different markings like layer two marking and layer three marking values. So based on the marking value, every device is going to treat differently based on the markings. Now, we got something called layer to marking and layer three markings. Now in the layer to marking majorly used in your layer two networks maybe if you are using a layer two ethernet network. Now, inside your Ethernetwork we use either an isl or dot one can encapsulation. Now, inside the frame you have some three bits reserved for these marking values, like in case of isl header which is going to add some 26 bytes of header. We have some three bits of cost value, class of service value which is used to define the marking value.
Now, the same thing if you are using dot one give frame, probably again you have a separate field called tag and in that it’s going to be three bits again, which is defined as a cost value. Now, if you’re using any layer three packets, like in case of layer three packets, we have a separate field called type of service. And inside the type of service there are two different types of marking values we can use IP presence or dscp. Now, dscp is somewhat advanced marking values which can which provides some more options when compared with IP presence. So more on this layer two and layer three markings like how they are differentiated. We’ll be discussing much more in detail in our next section.
2. Layer 2 – Layer 3 Marking
Now, in this video we’ll discuss some of the layer two and layer three marking which are available in the cisco iOS. So, before we go ahead with the marking options, let us quickly revise what we have discussed in the previous sections. Now, we have seen some basic introduction to classification. Classification is a method of differentiating the traffic into different groups by using classmaps. Now, using classmaps we are differentiating the video traffic, maybe a separate classification and we are differentiating the voice traffic and also your data traffic, something like that. Now, once we differentiate the traffic by using the class map, the next method we have seen something called marking. Now, marking is a method of writing the information over a packet.
It can be a layer to marking or layer three. Marking depends whether it is going in your layer to technology like in the switching or in the routing route. It’s going via routers or not, where this particular marking helps the remote devices to get identified. So when you are sending the information, it’s going to be sent with some marking and the remote device is going to identify based on that marking and it’s going to give the treatment like whether it is a high credit traffic or low period traffic based on the marking option. Now, one of the major advantage we get with the marking is we don’t need to when the packet comes, it doesn’t need to check the complete packet and it always allows you to save the processing on the particular lavender.
Now, the marking can be done either in the layer two or layer three. Like if we’re using some layer two technologies like especially in the switching concepts, we use something called cbos class of service which is used in the layer two lan. And we use something called mpls experimental bits. If you’re running some mpls core networks or Frame Delete discard eligibility bits we use in case of Frame Delete. Now we refer this as layer to markings or it can be a layer three markings as it goes from router to router, IP presence and dscp or the two different marking options we have in the layer three. Now, this marking information will be added as inside your packet here, like if it is in a layer two header, maybe your switch networks, you have three bits reserved for class of service.
Or it can be a dot one Q. In case of dot one Q, you have a priority tag which is going to use the three bits of class of service again. And if you’re using a layer three packet, the ipv four packet, you have a field called type of service which is going to be eight bits which is used to mark that particular information. So we’ll be seeing those. What are the different marking options we have, like what are the specific categories in the marking? First thing, we’ll start up with a layer to marking. Now, in the layer two marking. Now if you’re using a dot one Q encapsulation, you have a field called priority bit which is going to be three bits. And that three bits is going to define a class of service, that’s what we call as a class of service cos value. And based on this particular three bits we get eight different combination values.
Like you can either use 00001, that is the equivalent conversion, it can be 123456 and seven. Out of these seven values, rake zero is the best effort data. It’s like no marking. So all your traffic which you don’t want marking, it will be referred as a best effort data where the device will try its best to forward this particular information. Now whereas if you have some other traffic like low priority data like ftp traffic, you can mark with the value of one which will be considered as a low priority value, then two. Now here the higher number is more preferable, means nothing, but it’s a high priority traffic. So when any marking is coming with the value of zero, which means it is the least priority traffic, and any traffic is coming with a layer to marking with a value of seven is considered as a high priority traffic.
Now higher the number, it’s going to be higher the priority. That’s all depends upon our requirement. We can use any one of these markings, but typically we use something called medium priority traffic like ftp, we generally consider to be marked as one value. You can use something called high priority traffic like maybe your time sensitive application between the servers or some citric servers. We prefer to use some high priority traffic and the priority value of three. Three is something we can use for control signals on your vyp like hold option or the controlling signals inside your voice. And for the video conferencing, like real time broadcast video applications, you can use a marking of four and five is something we can use for voice VoIP traffic.
So specifically we’ll be using any one of this marking from zero to five based on the cost values because the last two values are reserved for specific control traffic. Like in case of layer two, it can be reserved for spanning tree or vdp based messages or it can be for routing protocols, updates, network traffic, so it’s going to be reserved for that. So in general, when we define any cost value, we give the marking from zero to five. Okay? Whereas phi is referred as very high priority traffic, whereas zero represents a very low priority traffic. Now the values will depend. Again, it depends, let’s say if you are using a frame relay based network, you have a frame lay header and inside that frame layer you have an option of discard eligibility.
And that discard eligibility will define again the priority of the traffic. If you’re using a famplus frames, again inside that mpls frame you have an experimental bits and again it’s going to be three bits again, which gives the same values what we have in the class of service again. Now, when you talk about any layer to marking, it can be on a switch network or it can be your frame related discard eligibility, or it can be mpls expandable bits. You have seven, eight different values we get. And in that we’ll be using something around zero to phi. Phi is referred as a very high priority traffic, whereas zero is referred as the least priority traffic. Now the next thing we have something called layer three marking. Now inside the layer three marking, we have two different categories again.
Now, layer three marking is done from router to router, especially router based on the layer three information. And the router is responsible again to convert the layer two marking into a layer three marking as it goes over the router network. Like, take an example, you have a lan, it’s receiving a marking with the layer two marking. Probably the router is responsible for converting into equivalent layer three marking in. Now, the original tcp IP standard defines eight bits in the form of type of service. We have something called a type of service field inside your tcp IP packet and it’s going to define some around eight bits. And that eight bits we can use to mark the specific traffic. And based on that eight bits, we have two categories of marking we can use.
Either we can use an IP presence value or we can use something called dsvp. Now, IP presence value, out of eight bits it’s going to use, only the first three bits will be used for marking. And it’s going to generate the same eight numbers, like from 0123-4567, same as your class of service, layer markings, the values, what we get with the IP presence value, it’s going to be the same, same combinations like what we discussed in the layer two. Like if you use a value of zero, zero is the least preferred traffic, whereas phi is preferred as a very high priority traffic. The naming will slightly change when you come with the layer two or layer three markings. But at the end, again, the six and seven are used for network control traffic, like routing protocol messages, ehr, pose messages, something like that.
Now, to verify these values, we can always go to the command prompt and either we can match based on the access list, I can say acl permit, IP nene. And then if you use an extended acl, probably here you’ll find some dscp values where I can say dscp value. If you use question mark here, probably you’ll find all the possible options. Now, either you can define the dscp value in the form of decimal values here, AF eleven issued forwarding options with the equivalent binary class selector, the backward compatible with IP president’s values, or expired forwarding option here. Now, even you can match like, let’s say if I give access list one, one, two permit IP nene we can also match based on the IP presence values. I think we need to use IP presence. And then if you are using your presence value, you’ll get seven different options.
Eight different options, zero to seven where you will find the marking from zero to five. Here you can see six and seven for the control traffic. Similar way either you can match them in the acls or inside the class map. Let’s say I’m going to create a class map match access group one, one, two inside the policy map. Let’s say I’m applying the policy map CCIE define the class class http. We can always say set president’s value. And if you want to specifically do the marking, we can define a class map which is going to differentiate the traffic. And then we can apply some marking value by using set presence value. And we can define what set of marking has to be applied for Http traffic. Or even we can define some dscp values based on our options here. Now this how the marking can be done by using a policy map. And inside the policy map we can use the set option set dscp or set presidents to do the marking.
3. Classification & Marking – Configuration
In this video we’ll see a basic example on classification in the marking. Now, in the previous sections we have seen classification is a method of defining traffic in different class maps. And then marking is a method of adding some extra information to the packets. So in this scenario, I’m going to take one small example where I have three different types of traffic. I got an Http traffic here and I want to ensure that htp traffic leaving the interface on the router one as one by zero should be marked with the presence value of two. And then I also got some voice traffic. We have IP traffic which is within the range of using the udp ports from 16 384 to 32767 should be marked with a dscp value of expedited forwarding. And all the remaining icmp packets should be marked with IP presence value of one.
Now, I got an EHRP pre configured on both routers and if you verify show IP EHRP enables. I have EHRP preconfer on both routers. So, the first thing we need to create three different class maps which is going to match three different categories of the traffic. So the first one is Http. We go. IP and icmp. Now, to match all the Http traffic, we can create one class map and we can define an option called Match protocol Http. Now, the good thing about the Sisquires is it can you can use something called network based application recognition where we can directly match a specific protocol, any specific protocol we can match. Once we give this command Match Protocol Http, it’s going to match all your Http traffic running on any port numbers.
Okay, it can be running on port number 80, or it can be any other port numbers. It’s going to match the Http traffic. Now, the next thing is we need to create another class map which is going to match my udp traffic, the Yvyp traffic. But before I go with the class map, I need to create an access list. So I’m going to create an access list with a name called Voice, which is going to be extended. And then I’m going to say Permit udp nene and I’m going to match the range of the port numbers. So all your vivo IP traffic uses the range of port numbers from 16 384 to 32,007, six, seven, and then I’m going to take it. So, if you show IP access list, so there is an access list which is created with a voice which matches the udp traffic from 16 384 to 32767. And then I need to create another classmap which is going to match my voice traffic.
So voice classmap. And then I can say match access group. Voice access group name. Voice It’s a name race field. So we need to say Match Access Group name and name and ACL name. And I need to create another class map. The third class map is icmp traffic, which is going to match protocol icmp. So it’s going to match my protocol icmp. Now if you verify the configuration, if I give you showrun classmap I can see there are three different classmaps created. One matches the voice traffic, another matches the icmp and the third classification is based on Http traffic. Now we need some different markings as they leave the interface. Now we classified and we need some markings. So to do the marking we need to get into the policy map, any name for the policy map. Now we need to define the class.
First we’ll start with Http and then what is the marking we need? We need to set IP presidents value of two. And then the next class is view IP. We need to say set dscp value dscp values. We can use expedited forwarding. And then the third category is the third class icmp. It’s going to mark with a presence value of one and then I’m just leaving it blank. So if we verify Showrun policy map you can see inside the policy map we had defined three different classes and we had defined different marking values for each and every class. Now the last step is to apply on the interface, leaving the outer one on the router one s one by zero is the interface. So we need to say service policy output and the name of the policy map. Now for verifying we can say show policy map interface s one by zero.
On this interface you can see it matches Http traffic and matches the protocol Http and it’s going to set the presence value of two. Right now there’s no real traffic which is coming and then voice traffic and you’ll see them packet marked with a dsc PX red forwarding leaving the interface and there is an icmp traffic matches the business value of one. So I’ll try to generate some traffic from router one. I’m going to generate some traffic for ping. I’m going to generate some ten prepaid packets. If I verify show policy map interface. Now you should see ten packets matched and you can see the ten packets have been mapped. So similar way if you have a real time traffic which is coming from voice, probably you’ll see the packets marked here and similarly htp traffic. If you have then probably you’ll see the packets marked here.
