The Complete Course from ExamCollection industry leading experts to help you prepare and provides the full 360 solution for self prep including 350-401: Implementing Cisco Enterprise Network Core Technologies (ENCOR) Certification Video Training Course, Practice Test Questions and Answers, Study Guide & Exam Dumps.

Architecture

27. Qos Design & Implementation Part01

Let us discuss QS design and implementation for this hardware or for these platforms. So we are going to start with 2960, and we'll see the QS design later on. We are going to discuss the remaining hardware, platforms, and design. Before we begin, we should understand this chart because you will be receiving various types of code words all the time. CS, DSCP toss, if, IPP, IP precedence, CS class selected, drop ofability, and ECN explicit congestion notification are examples. Now, I already told you that you have to focus on eight bits. So 123-4567, eight. The last two you can exclude because they are used for ECN explicit congestion notification. But you can consider these six. So, if you see this table on your right side, you can easily figure out what the cost value is and how you're going to map the cost to the DSV. Again, this particular diagram and slide are actually important. First of all, we have a category of applications, from high priority to low priority. So here you have high priority traffic, and then you have low priority traffic, correct? So you have routing and controls, for example, control information or a control packet. For that, you are giving class selector seven and class selector six. Now the first column that you are seeing is cost of service, or IPpresence, starting from zero and going to seven. That means you have eight options. So here is option number eight, that is, seven, and option number seven, that is, six. The seven and option eight is for the control traffic. You might think that this point in time is reserved for the system. Is that correct? Then you have high-priority real-time traffic. Let me clean this up. So we have high-priority real-time traffic. And for that high-priority real-time traffic, we should go and classify and do the marking. So let me go back. So again, you can see in this particular chart that you have class selective five, EF, that is for voice. Similarly, class elect four and then forward 41, 42, and 43. So 41, 42, and 43 mean these are the high priorities. So actually, this chart is from high-priority to low-priority traffic. Likewise, you have the streaming video transmission critical call signalling class Lecture Three, starting from 31 to 33. Likewise, you can read this. Now if you want to convert this in terms of bits, So now you must go and check the bit. What you can do here so that we can see that you have a DSCP So obviously the first one is the default. So everything is zero. But the second one is the Scavenger Traffic CS1 class-selected one. And you can see that one is a class-selective one, a 3810, a 1214. I don't know how you are getting this number. So you get this number starting with bit three. So 1234 is two, which means this is two. It will start like this: two to the power zero, two to the power one, two to the power two, two to the power three. And since these bits are zero, that's the reason you have two to the power of three, meaning two into two, four into two, which means eight. So that's why this is eight. Now, how much is ten? 1010. We know this binary will be ten, and then again, eight plus four will be twelve. Similarly, suppose you want to calculate this EF, how if this EF is 46, so that calculation here, we can see that you have 10110. So that means you have to leave this So you have two 48s, 16 and 32, and you are not taking this 16. So if you add this, 32 plus eight is 40.40 plus 444 plus 240. It's like that. You can check this. This is the math. You can pause the recording here and complete this math. Otherwise, whenever you want to write the code, you can go and take a snapshot of this particular slide, and you can refer to it all the time. Or you can make your own slide like this as well. All right, what is the best practice? The best practise for QS is that if you have it in both hardware and software, you prefer it in hardware first thing.Second, classify and mark applications as close to their source. So do the classification and marking of areas near the source. Because it is easy to classify and mark traffic nearby the source, you police unwanted traffic and, if it is unwanted traffic, put it into the default queue. Then enable a queueing policy at every note with the potential for congestion. If you see that condition is there, you can go and enable the queueing. Right? So, it's very straightforward. If you notice congestion on all three switches, you can enable that queue. On this slide, you can see that you have trusted ports. As a result, all of these yellow things can be trusted. Then you have conditionally trusted endpoints. So you have an IP camera or IP phone, et cetera. Those are the conditionally trusted endpoints. And then you have printers you don't want to trust, okay? So you can categorise and mark things as trusted or untrustworthy. You can see that you can have yellow ones that are trusted. Then black one is nottrusted, then green is conditional. Then there's this red one, which is the classification and marking. So what is the summary for this thing? That you should know where you have to trust, where you have to do classification and marking, and where you have to apply the QS policy at either the ingress or the egress, correct? Suppose if you're dealing with the hardware type ofQS, and that QS type with that particular switchmodel is one P, one P is stand forpriority three Q and one threshold. So in that case, you have one priority queue. and you have three queues. So priority queue. Obviously, you are going to use voice traffic or real-time traffic. And then you have three queues. In these three queues, you can go and classify your traffic. We'll see that in the upcoming three to four slides. You will see that it can be utilized. So then here you can see it's just next to this; after this we have, for example, one priority queue, three Q, and three T. Now, this example is for one threshold. But here we have the three threshold values. And that's why you can see t, one, two, and three. So all the queues have three thresholds. Anyway, it makes no difference how many thresholds you have with the queue. how you can apply this. Here you can see the classification of traffic. So that's a fixed thing; you have classifications of traffic for Escavenger, for bulk, for network management, and for transactional data. Remember this thing, mapped as it is with QS. So here you can see that Escavenger bulk data has CS 11213, low priority to high priority, and related DSCPmarking or DSCP table numbers 810, twelve, and 14. Okay. So while you are doing this, you should know the calculation. So, for example, this is a low-priority to high-priority task in this diagram. And in the last diagram, that is up. This is actually low priority up to high priority down. And here it is, high priority up to low priority down. Okay? All right. So now here, you can see that you have different classes. Obviously, you have one priority queue. So let me show you that. Which particular category is falling inside the priority queue? EF, CS 5, CS 4. Now here you can see that where is your EF is view IP. CS five is broadcast traffic and then CS four. Here you can see CS 4 for real-time traffic. So this traffic is falling inside of this. Then you have three queues. So you have Q number two. I can see Q number two. Here you have Q number four. And suppose you are using one Q as a default queue. Okay. So this is your cue number three. So I have Q two, Q three and Q four. Now, Q Two.I'm using CS Two up to CS Seven. Again, it depends on whether you put threshold three as CS seven, CS six, and then CS three as a threshold two inside Q2. And the rest of the weddings two, three, and four. Inside Q two, t one, then Q three,you are using 35% as a default. And then finally, you have Q four.You are using it as a four and a CS one. So it's a f one and a CS one. You can go and check. This is used for escaving. So I'm giving 5% of traffic to Q for default traffic or to Escavinger if one exists. Here you can see the bulk data. So bulk data and escaping to we are using 5% only, which means it will do the trail drop after that, but what is in default queue? So here you can see that inside default queue we have the default traffic, default traffic I'm musing for default queue, so Q three and Q four are actually used for default and scavenger, and bulk data for that I reserve 40% of traffic remaining 60% of traffic we are using for Q two and priority queue. So we have a policy map marking a mapping table and a weighted tail drop how it lookslike the first step for 2960 series which is that enableQS with command MLS QS then the good thing that youcan go and use course DSP or directly can use costas well we know Cost we know DSCP both have mappingas well so here you can see that cost mapping ifyou have any doubt to understand this mapping what you willdo you can go and check the reference slide so herewe have the reference slide The first one is here, so you can see the cost to the SCP of mapping column one and column three, which are actually column two and column 4810. 12. 14. 16. 18. 20. 22. If I look here, I can clearly see that 16.243-2464. Eight and 56 that's the marking if you areusing the DSCP model we can simply go anduse MLS QS trust DSP you can go andtrust different type of Cisco devices like Cisco phone. CTS. IP camera. a media player, so this is the way that you can go and enable the QS and do the DACP marking again. Here you can see that you can go and enable the GSP marketing. Now the next phase is how you will go and enable the policy. You have to create the class map and policy map classmap will be called the inside policy map. So, as you can see, I have a class map called VoIP media signalling transactional bulk and excavator. Once I have all of these class maps, I am calling each class map inside the policy map one by one, so policy map view IP set. Again, you can see that we are setting the DSCP bit, which means that we are setting the DHCP EF for Vape; we are setting for multimedia AF 41; and we are setting for class signalling CS 3 transactional data AF 2; and when you are doing this mapping, this diagram will appear, so according to your classification. You must go to your marketing and apply the policy, so basically, you have the traffic class classification and then you are doing the DACP marking while putting the policy. Now, once I have a complete policy, I can go to the interface, and then I can apply it, like service policy or input marking policy, and that's it. Okay, now it seems a little bit complex if you are new to QS, but if you go step by step with whatever we are following here, you will find that it's logical, and after a few practises or if you do a little bit of practice, you'll find that QS is an interesting topic to learn and do the mastery for.

28. Qos Design & Implementation Part02

In this section, we are going to discuss the 365-03850 and 9300 platforms. We can see that we have the trust model meanswhich particular interface is connected with what type of deviceyou want to trust and you don't want to trustor you have have any conditional trust. Now in the case of 3850 or 3650, we have these steps to follow. So when we are doing the ingress QS model, we have trust DHCP, a conditional trust service policy model, and finally, for the egress queuing that is the hardware queue, we have two Ps, two priorities, six Qs, and three thresholds. So let's understand that ingress trust model. We can go and go to the interface, and then we have the CLI command trust device, a Cisco phone, a CTS IP camera, etc. Or we can use the conditional trust example here. You can see that you can go and create the class map, say voice, whose cost value is five, while signal's cost value is three, and then you can go inside the policy map called Cisco IPphone where you can call the class whose cost is five, and then you mark the DHCP bit as EF. Likewise, you can call it class signaling, and then you can mark the DSCP as CS 3. Clearly, you have class defaults DSCP as a default. Then you can go to the interface, and you can apply this policy. So we have the option of choosing which particular methodology we want to use for the classification, or at least for the trust model. Now in the upcoming slides, I'm going to show you how you're going to create the policy and apply it over the hardware queue or over the queue model that we have: two P, six Q, three. We know that the NBAR is supported in this hardware, and Nbars can have 1500+ application recognition when the NBAR is supported; this is how we can create the class. So I can go and use say for exampleclass map voice and then I can match theprotocol that's the keyword match the protocol cisco phone,Cisco Jabber, link, audio, citizen et cetera. If you go there and type "match protocol" and a question mark, you will get a long list of applications that are supported on that particular platform that you can match it.Apart from that, we can go and create other classes as well. This protocol is related to NBAR. So again, you can see that you have the broadcast video match protocols with the square IP phone; you have the real-time interactive; the protocol is telepresence call signaling; the protocol is skinny telepresence control; the protocol is Citrix, et cetera. Okay, so we can do the classification of the traffic with respect to NBAR, and then the next phase is to call those class maps inside the policy map, so I can go to the policy map and then I can call the class voice. Now voice will call all these four protocols, and I'm setting the GSCP as EF. Likewise, we can create all these policies, and then we can apply them over the interface, correct? Now NBA is supporting 1,000 or more applications, and it is supporting the 12-class model, and this is one of the references we have. So here you can see that we have class voice, broadcast video, interactive multimedia conferencing, multimedia streaming, signaling, control management, transactional data, bulk data, and scavenger. So 123-4567, 8910, eleven, and let me show you the last one as well. So, while applying within this policy, you can see that you have class voice, video, and all the way up to eleven, with the class default being the last. Okay? So you can have protocol matching with all of these classes, and then you can go and do the ASTP marking, correct? So this is the way that we can efficiently use this N-bar feature inside the three K platforms. Again marking and policy. How we can do so is here; you can see that the configuration is related to marketing and policy. So, for example, if I have class VLAN view IP, setDSP as EF, and then police one two, eight k confirmaction, transmit, exceed action, and drop, what does it all mean? Again, in the policy concept, you'll find at each and every point what you can do here that you have, for example, B sub C conforming burst and then B sub E excessive burst. So what you want to do at this point in time, if you have confirmed action, is to transmit. So you transmit it, and then you have a certain buffer (again, exceed action) and you have the option to transmit or drop, but in this case, say for example, I'm dropping, so BSUPC transmits and excessive bursts are dropped because we are using the policy. Likewise, do you have all the policies correct? So, for all traffic classes, you can see all of these policies listed here. Then you can see conform action, exceed action, set DHCP, transmit DSCP, and table is table map for transactional data, then classify bulk data and set DACP transmit DACP table as a table map. Then you have the table map, which is a map of the DAP table, and here we can see that you are mapping zero to 810 to 818 to eight, etc. Now if you see this diagram with this slide, you'll find out what it is called. So here we have the table map for 18 to eight, then you have the table map for ten to eight, and finally you have the table map from zero to eight. Okay? And obviously you have the class default, youhave classes given bulk data for those youare using the table map as well. So this is the way that we can use the policy as well. Let me quickly go over the full policy configuration in the case of per port per VLAN policy. For example, we have a VLAN that matches VLAN one10 classmap, which is Dvland, which is 10. Again, you can go and call these class maps inside the policy. Now what we can do is that rather than using all the features that we have, we should strengthen the configuration, like creating the class map like this and then the policy map like this, and then we should apply it. Okay, obviously, if you want to do the policy, you can do the policy as well, and that's very much required as well. So let me quickly show you that how this queuelook like and what will be the final configuration. Because there are two priorities here, you can see that. So you have priority Q-1 and priority Q two.Then you have 123-4566-Q with three thresholds, and how can we map the traffic? This slide can be your reference slide. That's how you're mapping yourtwo priority and six queue. So for example, I have priority queue one and priority queue two. Here you can see—let me try to highlight this. So, for view IP, you have EFand priority queue label EF, and then you have CS5 and CS4 for broadcast video and real-time interactive. So if you are using your priority queues (10% and 20% here), then you should use your AF 4. So Afore is used for multimedia conferencing, where we use the bandwidth remaining at 10% and then the weighted tail drop in conjunction with both. Then we have AF three for multi meterstreaming where bandwidth remaining person DHCP based awaretail drop you are using the trail dropmeans whatever leftover it will start dropping. Then finally, you have the transactional data. Then you have AF one; we see a bulkhead and an escalator. So like that, you can use all your queues in this manner, and if you draw the configuration, that is what it will look like. So here you can see that you have voice priority Q 1 matching DSP EF. So when you're creating the class map, you are matching the DSP value. So you're in class 123-4567. By default, there are seven classes plus one. So you have to tell eight classnow how you are utilising those classes. So for example policy map, the name of policy mapis this for example then class priority Q one, prioritylevel one, percentage of bandwidth ten then Q two, prioritylevel two, percentage of bandwidth is 20. This is very much like this, correct? So once you do that, then you have the other class, which is the control management queue. If you go and check, you have a control management queue where you're mapping this high-level class electro. What you're doing for that you're givingbandwidth percent remaining ten and you're givingthe buffer as well 10%. If you go to this point in time, and if you check the SDWAN quality of service configuration, you'll find they've structured it in this manner, which means even better than this because you're getting all those things in the same place, and you can even do the configuration with a GUI as well. Then you have the conferencing queue, 10% buffer queue limit, DSCP remaining percent of 80%, 90%, et cetera. So the configuration, although here it isusing the strict priority, the remaining bandwidthpercent, the tail dropping, et cetera. So whatever you're seeing here, you can drill down to all the configurations and check the same day we have the configuration here. So finally, for wealth data, we have bandwidth-remaining percent five; the Q buffer ratio (Q limit DSP) is 80%. If 11% equals 90, it refers to Escaveventure traffic. So when you have the Escaveventure traffic, you're using the DSCP-based weighted kill drop. So that's why you have the DHCP-based weighted tail drops of 80 and 90. So after that particular limit or point in time, they will start doing the trail dropping, and finally you have this class default where you have the bandwidth percent remaining at 25%. So here you can see thisbandwidth remaining percent is 25%. So this is the way you can design the policy. So first of all do the classification, thendesign the policy and you can apply it. Correct? So do the classification, create the policy, and before applying, suppose if you want to use the shipping as well, so you can do the nesting of policies as well. I can go ahead and create one parent policy for 50Mbps traffic, and then I can map inside class, class defaultship average, and give this specific subline C IR for shipping, and then I can call my child policy inside the parent policy, and then I can apply the parent policy over the interface. Okay, so this is the standard way of doing the MQC modular QS CLI configuration, or even mostly on the iOS X platform. We are creating the policy and applying it like this one. All right, so let's stop here.

29. Qos Design & Implementation Part03

Now we are going to discuss the rest of the platform. We have a 450-06500, seven K, and seven 7700 platform. Whatever knowledge that we have gained so far, thoseknowledge we are going to use in these platform. So we need not worry to relearn all those things. Now directly here, we can see that in the case of the 4500 platform, we have, for example, one priority seven Q and one threshold. So, in total, we have eight questions. The priority queue is derived from that one, and if you have a twelve-class model here, you have twelve classes. You can see that these classes are mapped. So the priority queue is mapped with EF, CS 5, and CS 4, and from here we can check what is five, CS 4, and EF. Likewise, you have queue numbers 12345, six, and seven. As you can see, Q one starts with the default Q of 25%, and Q two is the remaining bandwidth of 1%. Q three-bandwidth remaining 4% We have all of the Q to class mapping correct for the remaining 10%. So this is about the 4500 QS design. You can see in the 650-06800 QS design that you have a twelve-class model and two priorities for six Q, which means you have two plus six Q plus four thresholds. Here again, you can see that Q 1 through Q 6 and then 2 are in priority queue. So, first and foremost, you can determine how you are mapping priority Q levels one and two, ones and two. We are using EF CS Four and CS Five. The rest of the queues here have been used. Now, for example, a foursome is using multimedia conferencing. It is used for multimedia conferencingqueue where 20% is bandwidth remaining. And then we have DHCP-based weighted random early detection. So if you want to do congestion avoidance, you can use either tail drop or a random early detection method where you start discarding the packet before you reach a certain threshold value in three. Also we have 20% bandwidth remaining plus weightedrandom early detection in Aft also like same. However, because you do not want to give more bandwidth for bulk data, you can use 5% of the remaining bandwidth. or maybe in the case of a scavenger as well. Okay so this is the model, this isactually the standard format but if you wantto change anything you can change as well. Finally, we have the Nexus platform, and inside the Nexus platform, you can see four Q and one T. Now, with this twelve-class model again, you have the cost. So, for example, you have eight Qs in one Q one like that.Actually, if you go and log into the Nexus platform, you'll find that you have so many inbuilt QS that it is recommended that you not change them until they are required. So how the mapping is here you can see that you have Qone, you have Q two, you have Q three and Q four. Q 2 is used as a Q default. Correct. So, for Q 1, it is cost five, six, seven, and Q 0. Obviously it is for default. Q three is mapping cost two, three, four. So if I start with cost one, then cost one four is Q four. And again, you can check the subsequent mapping. Scavenger is thus mapped to cost one. The best default is mapped to cost zero. That is your default Q. Then here you can see that you have Qthree mapping with cost two, three and four. Two is transactional data network management. Three is you can see thethree is multimedia and signaling. And likewise, all these queues have been mapped. So once we know this mapping and this queue methodology, obviously we can then go and write the syntax and write the software commands. That is again, it's easy. Again here you can see that you have four Q, one T. How it is mapped here, apart from this mapping, is that we are giving the bandwidth percent and the queue limit as well. bandwidth percent and queue limit as well, because we have one threshold as well. In 7700, where you have eight Q and one threshold, it's very important that we have a six-key design where you have two priorities and six Q. But anyways, you have eight queues, and you're mapping the eight queues. So first of all, you can go and check your priority traffic. Priority traffic is priced at five, six, and seven useful priorities. And here are the price values. Then you can see the Q 23456seven and then you have one default. All these twelve classes model how they aremapped with Nexus 7700 F series line card. So this is the way that all thesewired keywords can be used and mapped. And we've gone over each and every component and thing inside the QS in great detail. The next section onward, we have to study keys for wireless networks or wireless infrastructure. You.

30. Wireless Qos

The last topic in Section 1-6 is WLAN QS. Now WLAN QS may become the biggest linkin the chain because there are some youcan see that actually there are three primaryreasons for that means there are some issues. These W lands are not that efficientthat we have the wired land. So that's one of the reasons. So as you can see the primary reason are typicaldown sift in speed and throughput here you don't haveany dedicated pipe, you have some sort of shared media. Here you can see in the diagram that all the mobile devices or devices that are using WLC, SSID, etc. are using the shared network. One thing you can say is that they have shifted from full duplex to half duplex. So these are the issues; these are the constraints we have that may be creating the weakest link in the chain. The other important thing here to understand is that wireless land supports a wide range of services. So if you have wired land, you have to map your, for example, eight classes of wired land to four classes of wireless land. So that is also one of the limitations you have now because the wireless media is going to use CSMA-CA, which is a collision avoidance mechanism. So there are chances we'll see that in upcoming slides: that you have to wait for some period of time or a short duration of time, but that is a small fraction of time that can also cause trouble. For example, if we have a 15 millisecond cheater versus a five millisecond cheater, the improvement is 300%. For example, transactional data with a 14 millisecond latency versus a two millisecond latency can have a 700% performance issue or an enhancement in performance, both ways you can think of it. So these are the things that are considered for wireless media, and for that, obviously, we want to apply keys as well. What are the tools we have? We have the user priorities, and later we'll see how we have to map DSCP to the user priority. We have access categories; we have arbitration interframespacing; we have continuous windows; we'll see what AIFs and CW mean on the next slide, but they are associated with some sort of collision avoidance mechanism. Then finally, you have an enhanced distributed coordination function. We have to map the DSP to the user profile. You can define the trust boundaries. We have discussed policy innovation enforcement points earlier, so you have PS, and obviously in the WLCS they are supporting ABC as well as application, visibility, and control. Because we are using CSMACA, which is a collision avoidance technique, you will now understand about an if-sand contention video. If you compare it with the land technique, that is CSMA CD collision detection. So if collisions happen, the algorithm will start recalculating the refresh interval, and then next time they will avoid them. But here, by default, they will wait until they find that the link is okay or the transmission is okay, which means I can do the transmission if it's ready to send, right? So something like collision avoidance is the reason that the sender had to wait for a fixed amount of time, and that is nothing but AIFs. The other waiting time you have is a random amount of time, which is the contents of the window. So you're not only waiting for a set amount of time, but you also have the contents and window. And then finally, this AI phase and its contents and window timers vary by access category. So here you can see that you have access to categories defined by voice, video, best effort, and background. Now, in this access category, you know that you have high priority for the low-priority data, right? And according to that, your access category in conjunction with an "if" in conjunction with "contention," "window," "minimum," and "maximum," everything will work together. So, in the diagram below, you can see that the background has the longest contents in window plus the longest AI, as opposed to the voice, which has less contents in window plus the et cetera. So voice and video are the same, but the contents and video for voice and video are different. Let me quickly show you how these models are going to map. So you have an 8211 access category. You have access to categories for voice, video, best effort, and background. Likewise, wireless media mapping You can see that you have voice, video, and the best preferred background. If you are using Cisco Airbase, we will see that you have this category like Platinum, Gold, Silver, and Branch. Let me show you much more information related to that. So we know, and we have a study about this model. This is our RFC 4594-based DSCP model. Now, how you are going to map all these classes—or traffic classes—with the It Pelle 80 to 11 model can be a reference for this particular slide. So in the future, if you are designing the wireless QS, you can map in the same way that you are seeing here. So, for example, CS one is mapped with upon, and AF one is mapped with up, up, up. User priority zero is mapped with best efforts, so you can go and check everything. Then we have the high-priority data as well. So, for example, multimedia conferencing or voice plus TSCP admin As you can see, voiceplus TSCP Admit is user priority number six, and multimedia is user priority number one. Like that, the mapping should be happening. Now, the next question here is how the three-bit user priority can map to the six-bit DSCP. In that case, when you have user priority for DSCP default mapping, you can see in the diagram that you will map three bits and zero the remaining three bits. So they will mark it as a zero. So in that way, the frame will be processed, and obviously your actual header or actual DSPmarking will be kept inside the IP packet. So whenever you are marking your user priorities, that is going inside the Cap app channel from your access point to the WLC. So here you can see the marking will be like this, and they will mark only the last three bits. So user priority has three bits that will be marked with the three last bits of the JSP. Okay? And in this way, the flow of frames will happen. If you want to learn more about the QS related to WLAN, you can go check out RFC 4594, and you also have the Cisco validated design as well. So you can go and refer to the design document from Cisco as well. Yeah.

31. 1.7 Catalyst 9300 Hardware walkthrough

Finally, we reach Section section one seven.This will be the last subsection, actually, of Section 1. And what are the topics we have to understand in one seven about hardware and software switching mechanisms, how safe squareexpress forwarding is, what TCAM, Mac addresses, and FIB versus RIB mean? So, all these things will be studied step by step. And before coming to One Seven A, B, and C, I just wanted to explain to you the hardware architecture of the 9300. Although we have a study in our QS section that shows Cisco has a long list of switching platforms, Obviously, there is a routing platform, but we have a long list of products now because of this core exam and upcoming exams, and Cisco is keeping track or they are thinking that in the future people will go and use DNA or SDWAN in enterprise networks. So, when it comes to DNA in particular, we should use Dnaready hardware. And that's why, in the hardware section, I just wanted to focus on this. The 9300 calculus switch is security integrated, which means you can enable maximum and other security-related things, whatever they are in the switch generation related to security, and you'll find inside the 9300 that it's mobility ready, it's IoT ready, because we know that IoT is a big topic nowadays, and it's an even bigger topic in Cisco DNA, and it's ready for the cloud as well. So here you can see that you can connect with STA, you have a web UI, you have patchability, et cetera. So let's understand this 9300 and at least three to four very important features related to it. Now, here, we can see the family. So we have one G-code, 9300. Specifically, you can go and refer to the dataset as well. So you have 24 plus 48 ports, then you have one G unified POE plus 24 plus 48 ports, and then you have MiGG up. So first of all, at least one thing we can understand here is that they are POE ready. And not only are they POE ready; here you can see that line rate on all ports. They will give you line rate and throughput on all the ports. All the ports are poe ready.So you can connect IP phones or Cisco phones to any port and they will receive power. Then you have 24 multi-gigabit Ethernet ports. Here you can see you have 10 Gig, 5 Gig, and 2.51 Gig ports available. So you can see that portability ability, POE line rate, and everything else are enhanced here. Now, which SFP models are they supporting? As you can see, you have online insertion and removal of SFPs supported, like four into one SFP or SFP SFP, plus you have QSFP as well, which is two and 40. Then finally, you have the copper support as well. So four X, one 2.55, and ten GB. So all sorts of SFP, SRP Plus, and copper interfaces are supported, and here you can see that we have the power capability, which means what type of power modules you have. You can choose between 350 and 1100 watts of alternating current. It's very much similar to Cisco 3850 switches, and then we can do the stacking as well. The stacking capability is also there when you are doing the stacking of the switches, and then you can stack up to eight members in a switch. And in that case, obviously, when we are doing this stacking, they have an active and a standby feature. So someone is active and is taking control of all the switches. You can then log into all of the switches, and from active, you must configure the active, with the rest of the members supporting the active configuration. So that means we have the centralised control plane; we have the distributed data plane; and that's the power of stacking. We have one-to-one stateful redundancy. And again, we have discussed this earlier: we have a stateful switchover (SSO), plus we have NSF nonstop forwarding as well. So you've got a state full switchover; you've got NSF.