Cisco UCS and Unified Edge: Scaling AI, HPC, and bringing DC to the Edge

Okay, so hello everyone and welcome to this session on Cisco UCS.00:12This is 2nd session of the expert insights series. This is a three part series.00:20We started this last Thursday where we covered UCS, some of UCS fundamentals templates and workflows. Today we will be talking about Cisco UCS and unified edge. We'll talk about some AI workloads, high performance computing, and how to bring the data center to the edge.00:27With that being said, I have with me Ronak, who is my teammate from the customer success team. I am Prativa. Both of us work in the CAEI, which is the Cloud and AI or data center team of customer success in Cisco.00:48Let's dive in just before we get started. We'd like to have this session as interactive as possible. We'll get to learn about you, your environments, and use that to you know customize the future.01:06Sessions as well and also understand how can we support you better in adopting Cisco UCS. So please scan the QR code log into Slido, the password is Cisco UCS. We'll have a few questions along the way. Again, like I said.01:26That the reason is to make it interactive, to get to know you better and to also see how we can support you better in your adoption journey.01:46I'll pause here for, maybe 20 s to give you some time to log in, in case you're in having any issues logging in, please let us know.01:55Either on the chat or we also have an icebreaker question. So let me move to the icebreaker question. What role do you play in your organization? We'll use this icebreaker to also check if you're able to log into Slido.02:16And access the poll successfully, you could either scan the QR code or go to Slido.com, type in the number of Slido 2189301 and the passcode is Cisco UCS. Okay, looks like Slido is working fine. I see network engineers.02:34Solution architects, systems engineer, instructor, post sale service delivery. Okay, more systems engineers.02:54Okay, let me pause here for maybe a few more seconds. I see more folks typing in. Okay, so we have a lot of engineers and architects. So please for the others as well, please log in and you know be ready so that you can take.03:03More questions as they come in on Slido. Okay, with that, without further ado, let's look at, again, this is a three part series as I was telling you. Mastering Cisco UCS from four to edge. Session one was completed last week on traditional enterprise.03:22Is workloads. We spoke about inter site templates and how can you use work.03:42Flows to pre provision bare metal servers without having to do the entire process manually. So that was about last week. Now today we will talk about AI high performance computing and some specific UCS models that could help you with your AI.03:51Workloads and also unified edge.04:11Now the most interesting one we also have another session next week. Same time, it's a 60 min session and that session is dedicated for all of you. We'll do a quick recap of the entire series and also give you an opportunity to unmute yourselves and.04:13Ask questions live. So if you have some takeaways from these sessions and you try doing this in your environment and let's say you have some follow up questions for us, then this would be the opportunity for you to join in and to ask us those questions. So to register for the.04:33Session, please scan the QR code. You will see the registration link to register for next week's.04:53Live Q and A session, but for today, we're still not stopping you from Q and A please do use the chat to ask your questions in there and I will take them up. In case we do not answer any of the questions live, we'll make sure that we take them up offline.05:00And get back to you one on one. But if you want to unmute yourselves and ask questions live, we'll give you an opportunity for that next week.05:20So before we get started with today's agenda, for those of you who did attend last week's session, did you try using inter site workflows after the session last week? You can give us an honest response. Yes, not yet or hey I didn't join last week's session, but I'll watch the.05:29Recording and see.05:48Okay, ok. Honest responses I see. Anybody who's tried it has used inter site workflows before. I see it's more of a, ok, 5050, not yet but I'll check the recording, I'll plan to do it soon. Okay, I urge you to do that at least before.05:50For next Thursday so that if you have any questions from your, you know, testing, we'll still be right there to address them for you. Of course you could reach out to us one on one. We'll share our email address later, but, hey, who would want to miss an opportunity to just ask questions live? So do try them out and let me know.06:10Know what you think there are many many workflows that are pre designed for you, you know, that can take away all the time that you spend in manually configuring it one by one.06:30Okay, with that being said, what are we gonna do today? So this is what the flow of the session looks like. 1st we'll, you know, do a sneak peek of the AI environment, high performance computing. We look at some of the UCS server.06:42And the UCS portfolio, we'll then talk about what unified Edge is and inter site fleet management.07:02So the 1st one will be explore with us where we have a few slides and we'll walk you through what the concepts are. Then we have show me how, which means we've always heard from all of you saying, hey, don't tell us just what these things are. Show us. How do you do it? Tell us how can we do it? So we'll spend a majority of the session.07:09With live demos, Ronak will be driving those demos. He'll show us inter site fleet management and workload versioning across the fleet, and the last one, help me succeed. So again, as I reiterate, we do not just want you to come here, you know, learn a few concepts and then go back. We want you to go back and act.07:29On it, how.07:49Is it relevant to your environment? How can we make it better? Tell us all of that because this is not just the 90 min session and then, you know, we just continue with our day to day lives, but we want you to go ahead, act on it and give us feedback based on your journey.07:51So just one last question before we jump into the 1st piece on AI and high performance computing. Which of the following do you have in your environment? Do you have UCSX servers, unified edge, other UCS servers?08:11Let us know this so we could also, you know, make the content relevant for you in this session and maybe in the future sessions as well. Okay, I see zero, ok, 13 % for unified edge, not bad I was just about to call out that nobody has taken up unified edge, but maybe towards the end.08:27Of the session when you learn about unified edge, you'll be curious to go back and check the data sheet and see how it can help you better.08:47Good, thank you for answering this. Now let's dive in. So 1st one, explore with us. We'll do a bit of a quick, you know, concepts that we'll cover and then we'll get into the demo. So 1st one.08:55A little bit of history, right? So we all know that AI is NO longer something that only some organizations use. AI impacts every industry, whether you're in healthcare, you're in IT, you're in financial services, it applies everywhere. But the only difference is that.09:14Each industry will have their own unique use cases. E.g., healthcare, they may focus on diagnostics, financial services may focus on fraud detection, risk analysis, maybe somebody on retail would be focusing on how to make customer engagements better, how to make them personalized.09:34So the use cases might be different, but the underlying infrastructure needs to run these use cases, AI use cases, they pretty much remain the same. So regardless of what industry we are in, we want more efficiency. We want to automate a lot of processes.09:54We want to enhance even the decision making, right? Where we have a human in the loop, but we want somebody else to tell us what would be the best decision in a certain scenario. And we also want to create good customer experiences. So this is the baseline for regardless of which industry you come from. But what is.10:14The challenges. Now when we look at AI from an operational point of view, the challenge lies in.10:34Converting an AI use case to a production use case. How do you bring it into production? Because most of the AI efforts are still at a pilot or a proof of concept stage. What does it take to bring it into production? Why is it so hard? Because you'll have to 1st.10:41Identify the use case. Then you will have to select a model which meets the requirements of your use case. Then you have to train the model or customize the model with your data. And once the model is ready, then you integrate it with applications either for your users or for your customers, and then you.11:01Scale the application. Now according to a study that was done by Deloitte, as you can see on the screen, more than 80 % of IT decision makers, they say that it takes six months to two years to move an AI project from pilot to full production, which is quite a lot of time. And even.11:21Even among these.11:4130 % of fewer of their experiments actually turn into production use cases. And the helloms, they are evolving so quickly the number of parameters that they're using. You know, you'll see Helims, you know, maybe you're using a specific version.11:43In quite a few weeks or months you'll see one more version being rolled out. So the platforms also have to be flexible, and the last thing, data preparation, right? So you would want to maybe customize the model with your data. So how do you prepare the data in a way that can be used to fine tune a model and AI expertise.12:03Is is also another roadblocker. So usually the model is not the infrastructure is, you know.12:23Also an additional factor to the problem, but there are a lot of other things that also prevent our use cases from going into production, which is what we all desire. Now, with that challenge in mind, we also need to understand that not all AI workloads need the same.12:32Infrastructure, right? Like how many of us actually have LMs that we are training from scratch? I'm sure most of us don't do I'm sure hyperscalers do, but most of us, we don't want to train something from scratch because models they have grown trillions of parameters, very large, large data sets.12:52Too many GPUs, the training time is also large. So overall it's become very expensive. We don't want to train a model from scratch. But instead, what do we do? We.13:13Use an existing model. Now there are many foundational models, right? For natural language processing, maybe computer vision, audio models etc. So we pick a pre trained model and then we fine tune it for tasks that we want. E.g., answering some questions, summarizing, you know, maybe.13:25Emails or summarizing some data that is there on a SharePoint, you know, you analyze data or you generate data, so you pick a predefined model and then you fine tune it for specific tasks because all of us want to make sure that we see value very quickly and we don't have the time to build everything.13:45From scratch.14:05Or we use techniques like rag. So what is rag? So in rag, you feed, let's say, documents or company data as chunks to an embedding model and this data gets stored in vector databases. What does that mean essentially is when a user wants to retrieve.14:07Your data because these chunks of information are stored in a vector database, the responses that you get will be more accurate. You know, the, the model will be aware of the context because you've already fed specific documents or specific data as chunks to the model, which you later use.14:27As an inferencing. So you either do that or lastly, we have something called as HPC, which is what is called as high performance computing. Now what is this high performance computing? This is for workloads which are very latency sensitive. So even if you have a bit of a latency, it can.14:47Significantly impact your overall performance. So you need, let's say GPUs and compute nodes to be talking together, working together in real time, the.15:07Connectivity between nodes should be very good. There should be good throughput, very low latency, and overall performance should not have any impact. So for such latency sensitive workloads, we use the infrastructure which can support high performance computing.15:18Now, Cisco UCS comes in a portfolio, right? So if you want to use a specific model for just training, maybe a small data set or something else for training using a large data set or if you want to do high performance computing, then you have different models in the portfolio that you can use. Now.15:36When we share this with you, we also have some links that you could use if you want to calculate, hey, what will be my GPU VRAM requirements for my use case for an LLM, so there are some calculations that are available. We'll share these links with you if you want to get into that.15:56But, now with that being said, we've understood the AI, you know, kind of, the AI environment that is there. We've understood that, hey, not all the workloads need the same kind of infrastructure. I need personalized infrastructure, right? So based on my requirements, I need.16:13A different kind of infrastructure. So this is the Cisco UCS compute portfolio. Again, I'll not be going into the depths of it. You could always refer a data sheet and get back to us if you have any questions, but just understand that the Cisco's compute portfolio is not just for high performance compu.16:31Thing, but it's for the entire spectrum, training, inferencing, and so on. And each of the system in the portfolio is designed for a specific category, but what do they all have in common? They all have a single management plane, which is what we covered in session one last week. So regardless of which UCS.16:51Server or a system you make in the portfolio, you could manage it through inter site, which is going to be your single management plane.17:11Now if you look at, let's say UCSC 885A, right? The 1st one that you see on the screen, so that is the training powerhouse, which means if you have some large scale training that you want to do or high performance computing, you can use 885A now that is the default compute node that comes as part of.17:20What does Cisco hyper fabric. Now, what is hyper fabric? Has anybody heard of what hyper fabric is? Maybe a rise of hands or yes or NO on the chat.17:39We do plan to have a session exclusively on hyper fabric in the coming days. But anyways, hyper fabric is a way through which you can, you know, design, model, and also deploy your AI fabrics. But there are more use cases to it. We'll discuss that probably in one of the other sessions.17:50But this UCS C 8085A is the default compute node that comes in Cisco hyper fabric AI and this delivers the, you know, we were talking about GPUs need to be able to speak to each other efficiently. There should be very little latency or NO latency. So all of those.18:10Those requirements that is needed for high performance computing this model is able to provide you with that.18:30Now a little bit of the other models like if you look at 880A, which whenever you have very large data sets that you want to use to train the model with, you can use 880A, you have 845A, so this is more of a middle ground, middle sized training, fine tuning, medium to large model inferencing. So.18:37You could use 845A?18:57And the last one, UCSX. So this is modular. What do you mean by modular? You pay for what you need and then you expand as you grow. So which means you can keep adding more and more modules as your requirements expand. Maybe you want to add more GPUs, you know, maybe more compute nodes, so you.18:59Could do that on a modular basis.19:19So this is what the Cisco UCS portfolio broadly looks like for AI workloads and like we were just talking about single management plane, which is your inter site, so inter site, as you can see, right, you have a lot of components in inter site, one that we spoke about was provisioning, right?19:24If you want to provision a server right from scratch, maybe install some operating system on it, create templates, standard templates that you can reuse to deploy profiles to multiple servers instead of doing it one by one, then you could do that. So Cisco Intersight is a unified.19:44Management platform, which will help you simplify your infrastructure operations. So what do we want? We don't want complexity. We want to reduce the complexity. Now we want to reuse as many, you know, profiles and policies as possible without having to create from scratch. Maybe we want some best practices in there based on which.20:04Model of server I have, you tell me, you're the experts, so you tell me what is the best practice configuration that I have to apply for these servers, you know, based on the on field experience, based on the engineering experience that we carry. So inter site is able to provide all of these for your servers. So with that being said, if any of you.20:24Who is still using UCS manager to manage your infrastructure. I would still strongly recommend you to consider moving to inter site managed mode, but this is going to be a single platform. Now in this session, we'll not cover Intersight because we've covered that extensively in the last sessions. We'll share the recording of the last session.20:44So if you want to know more about inter site you can.21:04You know, watch that session. Now, again, we've looked at AI environment, we've looked at how, not for every workload, we need the same AI infrastructure. We've looked at Cisco UCS portfolio. How do we use Intersight to manage it? Now you have to understand that, you know.21:08So the compute portfolio is one part of it, but how can you get an AI ready infrastructure? UCS having compute is just one part of it, right? On top of that, maybe networking. E.g., you need your fabric to be able to speak with high throughput, low latency. So you have Nexus silicon one switches.21:28Before 400 G 800 silicon one switches, which perform high performance switching. So then you would need storage. You can either use the local storage that comes with the, you know, hardware or you could use storage with storage partners like NetApp or Vast. Compute we've already discussed on top of that.21:48Is where you install your OS and AI software. Again.22:08On the software stack to deploy your fabric, you can either use hyper fabric or Nexus dashboard. Nexus dashboard. For those of you who have it, you can use Nexus dashboard as well to deploy your AI fabrics. And on top of that security, now you have hypershield integration, your hybrid mesh firewall, Cisco.22:13Smart switches. So we NO longer want you know the switching piece and the security piece to be separate. We want security to be integrated with our switches itself. So you have Cisco smart switches. Observability, Splunk, and then we have partnership, you know, through again NetApp masters for storage partnership.22:32But let's say Red Hat OpenShift AI, right? If for the software stack if you want to manage or if you mlops, which means machine language operations for your AI workloads, so you have Red Hat openshift. Then we have partnerships with Red Hat and other organizations through which we have built CBDs, which are your validated.22:52Design. So you could follow what is mentioned in the validated design to get your entire AI infrastructure ready, whether you want to use it for training, whether you want to optimize the model or whether you want to use it for inferencing. So this is what your entire stack would, look like.23:12If you need, by the way, if any of you need access to the cvds that we are referring to, do let us know. We could share that on the space with you for your reference. The Cisco validated design, which encompasses not just compute, but all the other you know like storage, networking, security, observability.23:32All of that put it put together in one place.23:51Okay. Now with that being said, we have looked at AI high performance computing, Cisco UCS portfolio, along with UCS, what are the other things that I need to build a fully full fledged AI infrastructure, right? So this is what we've discussed so far. Now.23:55Of this is at the data center, but I want to take that to the edge. The workloads that create the most value, often run where the data is created. I'll read that again, so it says the workloads that create the most value often run where the data is created.24:15Which means data's created in the branches, right? Where we are all presents stores factories, etc. So it's not in the data center.24:35And one of the surveys that was done in 2021, that time 90 % of the AI data processing was occurring in the data center. Only 10 % of this AI data processing was happening in the edge. But by 2027, it's almost going to be a reverse. 75.24:426 % of the AI workloads are moving to the on premise edge.25:02No, why is that? So this is not just the trend, ok? There's a lot of technical requirements which call for this to move the AI workload to the edge. So what are these? So 1st thing if you look at it, bandwidth optimization, right? So if you want to move a large.25:08Massive data set. It could be maybe video streams or telemetry from different sensors. You want to move these data sets across the van. This is costly, it is inefficient, right? So if you can process data locally, then you can reduce the cost, you know, maybe congestions or all of that.25:25Can be better managed. So bandwidth optimization is one thing, but latency, as you know, right? You don't want data to go to the data center, come back from the data center, and especially if you have used cases where you need real time AI inferencing, your response time should be maybe sub million.25:45Second, you want very low latency, so at that time you can't be sending data from the branch to the data center processing, then getting your response to the branch again. So because of this, you want to bring the AI processing to the edge.26:05And of course, some of our customers also talk about data sovereignty and privacy, right? Where you want to make sure that all the processing, all the sensitive information, they remain in your own branch or in your own edge to satisfy your regulatory compliance. So for these use cases, or for the because of these.26:20Conditions a lot of customers, they want AI, the data processing to be at the edge and not in the data center.26:40Now taking this into account, Cisco unified Edge was built. I know from the survey, you know, we had very few responses saying that, hey, I have unified edge. But for the others, please continue to listen in because unified Edge comes with a lot of characteristics.26:49Around which Ronak will also do a demo to show you.27:09Which might actually be useful for your use cases as well. So we will look at Cisco unified edge. So, yeah, the image that you see on the screen, so that's probably how the unified Edge looks like. Three rack unit high and 18 ". So which.27:13Means this is the form factor is specifically designed for the edge. It's only 18 " deep and three rack units, so you can you can visualize it, right? So if you have any shallow depth cabinets you know in your edge locations, because data centers traditionally would have wider cabinets, but if you have something that's a little shallow.27:33To in your edge locations, this can still fit in right there. So it was optimized exactly for the edge or the branch form factor. Mounts anywhere, poor post rails, two posts, center mount brackets, so all of that easy to mount. Converged, so which means.27:53The compute node is the engine of the edge, but it also integrates compute storage, your software stack, so all of that is integrated in the edge. So which means you have the processing power that is required for local AI inference and data analytics, right?28:13Where the data is generated, so data is generated on the edge and you have all this info, all this processing power converged into one box, which is your Cisco unified edge. And lastly redundant control.28:33So there's a component called as ECMC, which is edge chassis management controller. So this ensures that even if some individual components have some issues, you can still access the system. So there is redundant control.28:45So this is what your edge or Cisco Unified Edge looks like. But again, as I was talking about the AI infrastructure, right? It's not just one piece. It's a lot of pieces tied together so that you have entire stack built on it. Like right now if you see on the screen, we have partnered.29:01With different partners, Nutanix, Red Hat VMware, Intel etc. So we partner with them to come up with, let's say best practices, what is the validated solution to deploy the full stack solution? Intercite, again, is a vital component, so you could do zero touch.29:21Deployment, global visualization. So where are my unified edge servers present globally? Let's say I have many sites, I have many servers deployed, I want to visualize and know where each server is present and that is possible through inter site.29:413rd party software integrations as well. So based on the use cases we were talking to right or the industry verticals, let's say retail managed manufacturing, healthcare. So based on those specific solutions, we integrate with corresponding 3rd party software so that we could you know validate.29:56With the design and those validated designs are available for you to use in your environment. And finally unified edge. Again, as we spoke about, it has everything in the same box which is compute, networks, storage, security, all of that that is built in together. Now, so this is what the entire stack looks.30:15Looks like with unified edge.30:35And, fleet management. So again, let's say you have just got a unified edge, you've just purchased a unified edge, right? And it will not be one unified edge, right? So you would have purchased a lot of unified edges. And how do I set them up? Now imagine previously racking and stacking, all of that used to be a bit of a painful.30:38Experience, right? You would do it one by one or maybe one server by another server, but here you have a zero touch provisioning as you can see in day zero, which means there are blueprints that are already available that you can use based on, you know, let's say what is the kind of deployment that you're going to do. You have specific.30:58Blueprints that are available that you can use and there is also NFC near field communication, right? So what happens is using this, even staff who do not have technical expertise, they can claim these devices from their mobile. So we'll show that to you in the demo, so you can just take your mobile scan a specific.31:19Code, you know, and then the device gets claimed automatically, so device claiming can be done by somebody with NO technical expertise, then deployment at scale. So once you've decided what is the blueprint that you want to do, you can create configuration that is you know reusable and repeatable.31:38So even if you have multiple sites, multiple infrastructure or multiple edges that you want to configure, you can deploy them at scale with just a few clicks and config versioning for conformance. So let's say you have used a specific.31:58Version of configuration, let's say V one. But later you want to make a change to V two with some updates, then it's very easy for you to update it and then roll it out to your entire fleet. So all of this, including how can you map it to different geolocations etc. So we will show that to you in the live.32:13Demo. But just before that, here are the tasks I'll just give you a brief of the tasks that you follow in fleet management and then Ronak will exactly you know show you these tasks in demo. So 1st one is the prerequisites are being met. So advantage license, do you have the admin or operator role?32:33Then you create the workload definition. So these are the blueprints that we were referring to, whichever suits your environment. Then you derive the deployment. Let's say you want to select a specific resource, right? Maybe based on serial number, maybe based on the asset tag, you want to select certain compu.32:53Put a certain nodes which qualify this criteria and then you want to deploy those configuration to those specific you know hardware. So you can do that. You can derive the deployment based on the resource qualification criteria, and then you can deploy it at a scheduled maintenance window. You monitor the deployment, you will see templates, profiles and.33:13Getting created, OS getting installed for your entire fleet. And if at all you want to create a new version of the definition, let's say, then you could just edit that, create a new derive a new deployment, and then the same process can be repeated. Now this flow chart has maybe 678 boxes, but in.33:34In reality it's just a few clicks. You go to one place, decide your blueprint, you derive your deployment, you push it. So it's as simple as that.33:53But again, having, said this, this is just the explore section where we walked through a few concepts, AI, high performance computing, Cisco UCS portfolio, inter site, what does the AI ready infrastructure stack look like? How can I bring UCS or my AI process.34:01Processing capabilities to the edge. What are some of the, you know, features of this unified edge to manage my fleet? So we've discussed all of this, but enough of me talking now, I will pass it on to Ronak to actually show you in action how all of this is being done. So with that, let me stop.34:21Sharing my screen and pass it on to.34:41And again, if you any of you have any questions, please do use the chat to ask that. Thank you very much Pratibha for such a wonderful explanation of the theory aspect of both X series and unified edge. Now we will move on to the demo part, but before.34:44So I will quickly share my screen for the demo. I hope it's visible can see your screen. Now, when we come to inter site page, we have a new tab here that you can see, which is called.35:04The workload tab. We will go to it and how do you define your workloads there, what are these tabs? What is the definition, what is the deployment, what is an instance? We'll try to cover that as much as we can in the demo. But before that, I would like to show you this. When you start using or when you claim your unified edge two.35:21Your inter site account, you get to see this tab unified edge. This is again a new tab that gets shown to you when you onboard your 1st unified edge. I'll just quickly go through the unified edge and how it looks like in that various tabs and then we'll show you how you can get started with your journey.35:42The unified edge. Now this is again a very similar page what you can see e.g. for your chassis or your servers, this is the details page on inter site. Here on the left hand side you get the details of it on the middle you have the diagram showing how the chassis or the unified edge looks like.36:01Now this unified edge has six slots here. This bottom left slot is fixed where we have the ecmcs, the pair of ecmcs and two pair of power supply units. Apart from this slot, the five other slots are completely customizable. We can customize it according to.36:20To the different use cases that we have. We can have compute nodes here, we can have networking nodes, we can have security nodes. So according to our use case, we can customize it.36:40Now if you want to see how does it look from the backside, you can also see that from inter site, so this is the fan modules that are there. You can see the inventory. Now in inventory it will have the, of course, the chassis management controllers, and if we haven't configured any.36:51Thermal profiles power profiles that will also get shown up and we also see the servers. Yep, now it's coming up. So in this unified edge, we have four servers, three of them are currently powered on and one is not.37:09This is the make of the server that is UCS XE 01:30 CM 08:32. If you want to see each particular server, this is what one particular server looks like. It will come up in a few seconds.37:29Yep, so this is the server that we are talking about, and if you want to see the top view, we can also see that. Now, this is what the unified edge is and how it will look like. And how do we get started with using it?37:45To get started with it, we come to this new tab that is called as workloads. In workloads, we have three different sections, the definition, the deployment and instances. Now each has its own unique categories or characteristics.38:00In which we start 1st with deploying the definition. Now what a definition is also mentioning about something called as blueprints where Cisco comes up with these best practices and configurations according to the type of workload that you will be running or according to the type of environment you want.38:16To have through these servers. So the definition is just some of the parameters from your end and the blueprint and it doesn't have to be one blueprint only. As we saw that unified edge, a single unified edge can have up to five different.38:35Nodes or five different slots, so you can have up to five different blueprints and you can maybe define five different blueprints for each slot and that can be done for all the unified edges that you have. We'll come to the details of it so that it becomes a little more understandable, but we will see how it will do.38:50Find these definitions as well. Now, when we get started with the definition, we need to have some configurations. Here we can see we have configured three servers in this definition. So there are four servers total in the unified list that we have physically present. Out of them we have configured the.39:10Three servers, that's why we saw that one server is powered off. So we have configured the three servers, and once that is done, then we assign this to a particular unified edge or a collection of unified edges. So we define the best practices and configuration on.39:30The definition and to actually push this configuration to the various unified edges and to which unified edge and to which slot do we want to push it on that we configure in the deployments. And once the deployment is configured, we also get to define the.39:48Maintenance window in which we want the deployment to happen. So the configurations has to be pushed to the servers itself. So when do we want the servers to be pushed or the when do we want the configuration to be pushed that also can be done through the deployments. So once all those configurations are done.40:06Then the deployments are ready. So here we can see that we have all these deployments, but only one of them is showing a status ok. So currently in this particular lab setup, we only have one unified edge. We have configured that to be part of US East store, so we are considering that we have a store in the US East and the unified edge is part of that particular.40:22Store and we have deployed the configuration on this particular one. But these fixed of them.40:42Are still available. So as soon as we have a unified edge in this US specific store and we claim it onto inter site, it will automatically get the configuration that we have defined in this deployment. So that is the magic of zero touch provisioning.40:49You configure this all at once and then as soon as you claim it on inter site, the configurations get pushed according to the schedule that you have made. So NO need to do the configurations one by one on the boxes or you have you don't even have to be part.41:04Off the place where the configuration is happening, you do not need to be near the box. You can just configure this here as soon as the box comes up onto inter site, the configurations can get pushed, so you do not need to do anything and this is very simple. We'll see how it looks like, and this is just the instance. Each instance is just.41:22Corresponding to one unified edge, so every unified edge will have one particular instance. So that is what instances, this is the actual running configuration on the particular unified edge. And we can also see the inventory again the same thing the servers and the unified edge can be seen.41:40Now let us move on to the actual demo of how do we configure definition. We'll start with the definition and then we'll move along with deployments and we'll see what comes along. So to create a definition, we have this tab.41:57Over here that is to create definition, this is a guided demo that we have created so that it becomes easier for everyone to notice what is the step that we are going to do. So we'll click on create definition. In creation of definition it is very similar to profile creation, the similar kind of configurations are there in the general tab as well.42:13And one thing that we have introduced new is this setting of tags path tags or even locations can be set. So in path tag you can configure under which path do you want this definition to be part of. So do you want it to be part of.42:34Of global america. So you can kind of do a geo tagging through the path tag itself. Where do you want this definition to have the scope? Do you want it to be, do we having the scope of the entire America or a particular region, you can set that through the path tag. Once that is done.42:54Then we have these five blueprints. So these blueprints are available to you. These will depend on the type of operating system or environment that you want to create. So if you want, e.g., if you have a unified edge and you have five different servers in that particular unified edge.43:12You can go ahead with one bare metal, one ESXi, one red hat, one Ubuntu, and one windows. You can also have, e.g., three ESXi and two Ubuntu's. According to your use case, you can configure it. And you do not need to configure it separate for all the unified edges. E.g., you want hundreds of unified edges.43:30Is to have.43:50Two ESXi server nodes and maybe three Ubuntu server nodes. You just configure it once here and that can be used to create deployments and those deployments can be used to create instances automatically once we get there. So we'll come to that, but for this example we will have two server.43:51Was configured, one will be dead hat, one will be Windows server. So once that is there.44:11We select that and we click on next. Now, once the blueprint is selected, each blueprint will have its own user defined configuration that you will need to do. So e.g., for the red Hat one, you of course need to select which firmware version do you want the server to run. So under.44:18It is still a UCS server, it needs to run a server firmware.44:38So we need to select which firmware version do we want to run. Once we select that, we move on to the next tab, and then we need to add VLANs, which VLANs do we want this server to be accessible with? So for this example, we'll just add one VLAN that is 1355 you can add multiple as well, and you can.44:41Add them as ranges as well. You can make it-and then add the starting and.45:01Ending values. So for this example, we'll just put it here. Once that is done, we'll next move on to the OS configurations. Now in OS configurations, of course you need to select which version of OS are you planning to use.45:07So all the versions that will be available will be listed here. We'll select nine.six and you also need to upload the OS image. We can do it on the software repository which you can see as an option here. You can go to software repository and upload a OS image directly or you can also give a file.45:22Location through a file server and that can point to that location. Either way, once that is done, you click on OS image and whatever is available currently on inter site or on a file location, you can select that and then select.45:43Then we also need to select the SU image. We will be using the unified for this. Once that is done, we of course need to give the host name for the OS. What username will the OS be using? What will be the password for that, and the network configuration. You can configure it in a static way. You can also.46:00Go ahead with the DHCP server or you can also configure through an IP pool. You can create IP pools on inter site and you can let that IP pool define.46:19What IP address will this server be using or which which IP address will this OS be using? The OS will be having its own IP address. So for this example, we will select just the server for our labs example and then we select the constraint.46:28Now here you can see that it is asking the slot ID. So we have one to five because there are five slots that can be configured according to our needs. We can have our own servers on the five slots. We can configure which slot do we want this particular blueprint to take place in.46:45So our example we will select the slot 1st itself. So the 1st slot will be the Red Hat server, 2nd slot will be the Windows server. The rest slots we are not doing for now, this example will only have an example in which we have.47:04Multiple unified edges and the 1st slot in those unified edges will be a data server and 2nd slot will be a Windows server. So we'll select one here and we will go to the next option. Similarly, we need to do the same. We need to do similar configuration for Windows as well.47:20We select the firmware version of the server. Once that is done, we need to select which VLANs do we want to allow to the server. Once that is done, then we move on to the OS configuration, the OS version that we want to use.47:38The OS image again and the SU again. Once that is done, we configure the host name for the OS, the password, and in windows, we also have the option to select which addition of Windows server do you want to use so we can use standard data center.47:54Or V according to your use case. This example will put it as standard. We also provide the product key so that the windows can be activated, and then we select DHCP again for this particular installation and we select which slot do we want the Windows server to be at. So in our use case we want the window.48:13The servers to be at slot two for all the unified edges.48:33So we will just select it as two. Once that is done, then we go to the system configuration. So this is mostly for the chassis itself. What IP pool do the chassis or will the chassis use? So we can either create a pool or if it is already created, we can use the.48:37IP pool from here. Once that is done, we select.48:57That we also provide the in band management VLAN ID, which ID will you be using for that? You can mention it here. Now, for management, this ECMC is very similar to the fabric interconnect or UCS direct that we have in other chassis or as a switch itself.49:01So we also need to select which infrastructure former bundle do we want to use? This is nothing but those former bundle similar to those FI ones. And one more thing that I would like to highlight here is because this is completely done through inter site, it is only supported in in inter site managed mode.49:20And not supported in UCS managed mode. So this will again be a firmware version for inter site managed mode. So we will select the former version for IMM domain. Once that is done, we will, we can create a local.49:39User policy as well where we create or define a local user for login and access to it and define the password and we can also select which time zone do you want to use.49:54For this particular chassis, and then we need to define few more things like the DNS, what is the primary and secondary DNS and which VLANs do we want to allow to this particular C ECMC. So we want to allow.50:11135-52-1360. Once that is done, we need to select whether we want to allow auto allow it on uplinks or not or do we want to auto allow it on the cluster links that we can select and then we click on add. Once that is done, we have an interesting feature called as.50:26The allow overrides option here. So the allow override option, if you enable, what that does is there would be few settings in definitions which will differ for different locations. E.g., you are defining something which you will use all across India or all across you.50:45But there will be specific.51:05Options that will be specific to a particular region. E.g., the IMC access, the IP pool, the VLAN IDs, those could, those could be and probably would be different for different locations. So if you allow this, allow override, then you will be able to override these settings when you create your deployments.51:08So the definition will be a broader thing on the top and when you create different deployments out of it, in each deployment, if you want to override this configuration, you select the allow override so that you can make different configurations for these different specific options. So if you want to use different time zones for different deployments.51:27You can select that. So once that override is selected once all the configurations are done.51:47Then the system configuration is completed. Then we'll move on to the next option. Now here are a few optional settings you can configure your NTP, so you can configure your QoS, you can configure SNMP, serial overland, syslog.51:52Fan control, do you want it balanced, do you want it to have aggressive control of the fan where you want to have maximum performance? You can configure all these, but these are optional According to your use case, you can configure them. Then once that is done.52:09We are done with the definition creation, so you can see the various options here. You can go to each tab and see what it looks like. And once you are satisfied with the definition, you can just create and validate. It will create the definition itself and validate whether all the.52:24Configurations are done properly, and once that is done, then you can create your deployment directly from here or you can close it and then create it in the next option. So this is what the definition is looking like. So we have these.52:44Definition that we just created and we can go to the option here and it was a similar option that we saw on the bottom right, but we can do it from the three.menu as well. So once we go to the deployments.53:01This will give you option to create a single deployment or multiple deployments as well. If you want to have three, four kinds of deployments in which maybe one deployment is for one particular kind of workload or one particular subsidiary of your organization, another deployment for another subsidiary of your organization and.53:17So on, you can create multiple as well or if you just want to create one deployment, that is also possible. Now one deployment can have multiple unified edges. It doesn't mean that one deployment is only for one unified edge.53:36Our deployment can be used to create or configure multiple unified edges as well with the similar configuration. So we'll go with single deployment for now, we will click on start. We need to give the name what name, what is name of this deployment. So in this example we are creating a deployment for the CS.53:49Region of US and we can also again set the tag. Previously for the definition we used global/America because it was for the entire America region, but this is this deployment is specifically called America Central.54:09So we'll select that. And once that is selected, we'll go on to the next tab. Now here we can see the configurations that we have already configured in the definition. A lot of those are already here. We can see those.54:24There is the Windows server configuration which we can see the system configuration. Now in system configuration, as we gave the option to override for IMC access, we get the option to delete it from here. We can remove this and select a new IP pool according to the IP pool.54:39That is present in this particular region. So for entire America or for one particular region, the IP pool will be different.54:59For this particular region, the IP pool is different so we just select that and again again the time zone is different for this particular region, so we'll just update the time zone to the correct one. And other settings we are keeping it constant. We're not changing it only the settings that we wanted to change, we configured them as.55:06Allow override and then we are configuring them here.55:26Now once that is done, we click on next. Then we go to schedule maintenance. So when do we want these configurations to take place on the unified edges? So we can create a new configuration on a new schedule here or we can select the one that we have already.55:30Created. So here we can see that we want to have the maintenance.55:50On weekly basis on Sundays, the start time is this, the duration will be 4 h and recurring. So why is it a recurring one? Because this is taking an example in which you have a lot of unified edges and slowly and steadily, those unified edges are getting shipped to you. So as soon as they get shipped to you, you.55:55Add them to inter site on every Sunday it will check whether it meets the criteria of the unified edge, whether the unified edge is in in this particular location, is it set with that particular tag if it is and if it is available, then we move ahead and deploy the unified edge.56:15Again, if there are any changes that we want to make and we make a change in the definition and the deployment, and we want to ensure that the change gets added to the actual deployment, these maintenance windows will be used to make the changes on the unified edge as well. So that's why it is recurring, that's.56:32Why we have this particular schedule that we have here. So once we configure the schedule, we go on to the next tab. Then we add a rule group.56:52Now this rule group defines which location do we want this deployment to happen. So we select the location, we can either select it by location, serial number or path tag. We have been configuring path tag from the 1st demo itself.57:02And while creating the deployment, we also selected that. So we will select global America central. So once a unified edge comes up with this path tank, it will get the configuration that is deployed.57:17That is configured in this deployment. So we'll select that and we'll go to next. And then we move on to instance name prefix. This is just the name of the instance that will be defined on the particular unified. So e.g. unified edge one, you want a particular prefix to be there.57:33In our case, we just want it to be US specific and how many digits do you want to it to have as a suffix? Just 123 or four. So do you want it as one or zero one or 001 and so on. So similarly, you can configure it.57:52And once that is done, this is just for naming convention of the instance, the instance will have the actual configuration of the deployment and that will get assigned to a particular unified edge. So just for the naming of that, this is there. So once that is done, then we get a summary page and then once we go to the next option.58:08We are ready to deploy. We click on prepare to deploy and then we see that the deployments are ready. We have configured few deployments here. So this is the one that we showed currently, but apart from that, we have few more which are ready to deploy as well. So next step will be to actually.58:28Okay, now let us move back to once those configurations are done, once you have configured your definition, once you have configured your deployment, what do you do next? The next step as also mentioning is very simple. You do not need to do anything. You do not need to have any knowledge, you just need to have a mobile phone. So on your.58:49Mobile phone, you have the inter site account or inter site application. You can just get started with using it.59:08And it will save a lot of your time. You can have a visibility of your infrastructure from your phone itself. You do not have any management options because of course for security concerns, we do not want to give management access to the mobile app, but if you want to have monitoring and a lot of monitoring is possible right from your mobile application.59:15But apart from that, there is a new feature that we have introduced recently that is the target claiming of devices. So once you go to targets, we have a new option here. So here we can see all the targets that are there on this particular inter site account. Our inter site account is already logged into this mobile application.59:35Now if you want to add a new target, you click on the plus icon. Normal target claiming will be through that claim ID or claim code and device ID, and the new option is to claim with NFC. So every phone these days have NFC and we tend to use a lot of NFC tags in a lot of our day to day usage field.59:55Just. And it does something for us. So those kind of tasks can be configured. So taking the same.01:00:14Functionality we have introduced this new way of target claiming, you do not need to copy all those device ID claim code. You do not need to do that. You take your phone, install the inter site application login with your account, and then claim with NFC.01:00:23Once that is done, then we get to see that we get started with claiming the device. So we get started. Then this is the important part where we set the location. We need to select which location do we want to use.01:00:38Now this location will be used to do two things. 1st of all, it will show a global view in a map format where this unified edge is located so that you have a world map of the different locations where the unified edge is located and it will also help us in the provisioning.01:00:54So we will go to the next part. We can also assign it to some resource group again similar topic of inter site, if you want, we can do that. If not, we'll just go to claim. Once we select claim and we tap the phone on the NFC type of the device, then the unified edge is claimed. We can see it here.01:01:14And this is all the options that we get to see once it is claimed. We get to see every configuration that is there on this particular device. So that is the entire process of claiming. You do not need to go to your internet account, NO need to copy paste anything. You just tap your phone on the device and it gets.01:01:34I mean, that's all.01:01:55Now, once that particular configuration or that particular device is claimed onto inter site, this is how it will look like on the maps as we were seeing. So these numbers define how many unified edges are there in this particular region. If it is single, it will be highlighted like this. If it is in with the number it will.01:01:57Now the blue one mentions that it is newly onboarded and it is not currently having any configurations so we'll click on this and we see that the configuration is in progress. So as soon as you claim the particular device or the unified edge and it matches the path.01:02:18ID, it will automatically go ahead with the configuration according to the maintenance window that you scheduled. So we'll select this and we see that this is the name of the unified edge and we will see that it is already in progress. So this is the unified edge that we are looking at.01:02:35So what is currently happening on this? We can see in the request tab, we can also click here, but we can also do it from here. So we see that the blueprints are getting deployed, which are the blueprints, the Windows server and the server that we were defining. So once that is done.01:02:55We closed this and then we can see the configurations that we had made. So the server configuration, the OS configuration, the constraints, similarly for the windows as well.01:03:14Similar to what we already configured, we can see it here. Now the same configuration that we deployed or we created in the deployments are getting pushed in the instances as well. So in inventory we can see that these are the two servers that are getting configured. So once that is done.01:03:29We'll just refresh it and let it be complete. So we can see that it is now in green colors, so that means that it is in an ok state and the conformance is also all good. So this is what it will look like to get started and pushing the configuration.01:03:45To any device without knowing anything, so you just need to onboard and if it matches the criteria, it will get the configuration done. You do not need to do anything at the edge location so.01:04:03We can see what all are getting configured, we get to see that similar concepts as we saw in previous.01:04:23Sessions as well, that is UCS server profile. So at the end of the day, the deployments, the definition, the instances are going to create server profiles and everything that is there on inter site, but that will be created automatically for you. You just need to configure some things and all these policies, all these profiles get created automatically.01:04:30Quickly for you. So your job becomes a lot easier and you do not need to do a lot of manual work. So you can also go to the unified edge and see the profile for that. This unified edge profile is very similar to the profile that we see on inter site for domains, the IMM domains that we were saying.01:04:50So this is again very similar, the similar kind of configurations you can see over here. We have EMC A, you have ECMC B similar to fabric interconnects, and we also have switching configuration, again very similar profiles or policies as we see in.01:05:08Fabric Interconnect as well. So once this is done, then we look at automatic upgrades. How do we define upgrade of firmware? Okay, so this is what the workloads look like. These are all the unified edges that we have in.01:05:26Our.01:05:46Environment currently we can see a list of all them. We can also view it in a tabular view. So this is the definition. We will again go here, we will manage version. Now there is version management in the definition as well. As we were talking about that you might have to make some changes in future, you might want to upgrade.01:05:47Something you might want to make some configuration change in the definitions that is possible.01:06:07And that is also done through versioning. In case you want to have any change, you can make it a new version and then create it so that you can have different versions and you can log what version you're creating and have a visibility of all the versions. So we'll click on create a new version.01:06:12So the source will be one of course and the next version that we are creating is two. So we'll create that. We'll need to give so we can create the version here and then whatever changes we want to.01:06:29Make whether we want to make it on the OS itself.01:06:50Or some other tab, we can select that accordingly, but in our case, we are just going to make it on the system configuration. So in the system configuration, we want to make a change on the infrastructure firmware bundle. So the ECMC firmware. So we'll select this and we see.01:06:53That a new version is available, so we'll select that and then we'll click on save. So this is the only change that we're going to make for this demo, but you can make changes here as well apart from this. And then we'll click on next. And again, any optional settings if you want to include, you can do that. Then you see a summary of.01:07:11Of what you have currently.01:07:31Created and then we save and validate. Now, this is again another option that you get to see. Do you want it to be saved as a preferred version or not? If you want it to be saved as a preferred version, that version will be taking precedence and those configurations will get pushed automatically. If you don't want, you just want to have.01:07:33Another version of it that you can do else you can push the configuration for that. In push out or in rollout as well, you have two kind of strategies that you can use. You can use batch or canary. Batch will update all the systems in this particular.01:07:52Definition and Canary will let you choose a small group of system 1st so that you can make the changes on boom, on those and then see if that is impacting something. Maybe you can do it 1st for your lab environment and then for your production, that option is also available through the canary option.01:08:08For this example, we'll just keep it to batch and then we'll go to the save option and that will update it. And we'll go back and we will see that the changes are happening in all these options.01:08:28So these are getting blue, so the changes are happening, we can also see the status as the changes are getting pushed, and when the changes are completed, we also get the status of that. So once all of it turns green, we get to see that the changes are successful. So all the unified edges got upgraded with the.01:08:44Firmware for the ECMC controller. We can see it in the table view as well. So all the deployments are at good state now are all at ok and all the instances also under those deployments are also in an ok state.01:09:04Okay, now we will see if you want to add a new region as well to a definition, how do you do that? This is very similar to the creation of definition or creation of deployment.01:09:22So we'll just skim through it and we will select the options, ok, schedule policy, we'll select that.01:09:38We can see ok so we can also have multiple schedule policies which you can see on this. You can the path tag for specific this we give the prefix name.01:09:54And finally save and validate, and this will be ready to deploy. Now we have one particular unified edge that just came up, so we'll click on this and then we set the location. So we'll set the location as this which is relevant to the tag that we set.01:10:09Then we go to workloads, then we see here, and we go to instances. So the instance automatically got created as soon as it saw the unified edge that is claimed. So we go to inventory and we see these are the two servers that are part of this particular unified edge instance.01:10:29Now we can just click on deploy and the deployment will happen either in that or canary according to your need. Okay, so we can see it here and go back to workloads and it is in progress.01:10:47And we can see that once it is done, it will again down to green. So that brings us to the end of this particular demonstration, but I will share one more thing quickly on the versioning as well. So we saw a versioning of.01:11:11Firmware version, but there is one more versioning that is supported. I will share my screen once, ok? So we'll see the version.01:11:30For this as well in terms of if you want to make some configuration changes, how will the versioning look like? So in this we will see if you want to add a VLAN to a particular definition, how will it, how will it look like? So you create a new version here.01:11:48Oh Similar to that, what will be the original one and what is the one that you want to create? Then we want to make changes here and add a few more VLANs, currently only VLAN ten is there, but we want to add a few more.01:12:07Okay, we have added the VLAN that we needed and then again in the other bare metal server as well, we want to add another VLAN, so in this case also we have two blueprints for both the servers, one both are bare metal.01:12:25And for the system configuration or the VLANs on the ECMC, we have added two more VLANs that is 40 and 50 so that both gets passed to the service. Now we'll select it as.01:12:43Preferred version, then we select it as batch, and then we click on save, and then we go back. Then we'll see that 2nd one is active now.01:13:00And the one is inactive and both the versions are visible, so you can see both the versions and what changes were made and some changes are scheduled. So according to the scheduling that we have done, we will see it. And according to that we can see that the status is ok but the conformance is not because.01:13:19The configuration is still not fully pushed. Once it is completely pushed, the status will update to conformance ok as well. We can also see it happening on the instances as well. We have three kinds of deployments but eleven instances up.01:13:37Unified edge.01:13:57So all the status is getting pushed ok once it is done, we can see both conformance and status are ok and we will also be able to see the wheels are also added.01:14:01To the unified edges all the profiles will also get updated accordingly. The server profiles.01:14:20And everything will be updated in the backend by just updating it on the definition. So that's the other demonstration that I wanted to show for versioning as well. With that I would hand the stage to Pratibha so that she can.01:14:35Move ahead with the session. Thank you. Thank you for that. I hope the demos have helped all of us understand how can we, you know, configure our unified edge, manage the fleet, do, you know, version.01:14:51Change to the configuration that we push, monitor the geolocations where they're configured, use asset tags and so on. I know there's one question on the chat from terms of deployment guide. I'll check and I'll get back to you on that week I've noted down your email address. So let me check. Ideally.01:15:11See Nutanix is just the overlay on top of it. So if you have an image server with the corresponding GPU drives installed, as long as it's supported Nutanix is the overlay so it should auto detect the GPU drive. But in case you you're saying that you already have a node or UCSX node deployed using Nutanix.01:15:32But you just want to add a GPU drive to it and you're looking for, some sort of steps that could help just clarify that so that I can look for it accordingly.01:15:51Okay ok so you say you have an inmate with the GPU. Is it already installed or are you just looking for an installation of it you know using Nutanix? But anyways while that's being answered, let's continue with the remainder of the session. Let me quickly share my screen.01:16:02Okay, I hope you're able to see my screen or if it's loading, let's just give it a moment.01:16:31Okay, so, just a quick check before we go to the last section which is help me succeed. Which feature of Cisco unified Edge is the most important to you or did you find most interesting? Is it the simplified onboarding of your fleet? Is it the automated life cycle management?01:16:39And zero touch provisioning, is it the device claim claiming process which leads which needs NO technical expertise or is it the global visualization?01:16:58Okay, zero touch provisioning. Secure device claiming. Okay, see 50 50 % of responses.01:17:09Okay, I see one for simplified onboarding as well. Okay, do let us know if you already have unified edge, and if you have any specific use cases around it or if you have any questions, you could always reach out to us. But with that, let's move to the last segment, which is help me.01:17:28Succeed. Again, like I was mentioning, the session is not just 90 min you come you learn and you know something new, but it's about how can you take it back to your environment and.01:17:48You know, use it for your needs. So there are key points to remember. 1st one is of course in terms of AI sizing from the workload. So, you know, do you need infrastructure that.01:18:00That's just for, maybe a small sets of data or let's say large data sets or if you want infrastructure that supports high performance computing, so all that needs to be decided 1st because not everybody needs the same set of infrastructure. Then using the cvds, you can form a repeatable pattern.01:18:15Instead of having to redesign every time that there's a new requirement or for you to scale up your AI workloads and the corresponding AI infrastructure. We spoke about edge especially because the initial part was about the entire UCS compute portfolio that helps with AI workloads, different, systems.01:18:35For different needs and then decide to manage it all. But edge especially because the trend and the need of that is such that we want to do a processing of the AI data in the edge locations and not in the data center. So for that we have unified edge, which comes with zero touch provision.01:18:55Thing you have NFC claim blueprints that Ronak was explaining about versioning and so on and all of this managed from inter site, which is your single management plane regardless of which.01:19:15You know, type of UCS system that you are using or if you're using unified edge. So this, these are some of the key points to remember, but before we close and before I give you some resources to also look back at, I'll just reintroduce.01:19:29Cisco IQ, for those of you who have not heard about it already, this is something, this is a AI driven product that Cisco has launched which needs NO, you don't have to buy any hardware. You don't need any additional licensing. This is something that I'm sure all of you can use just with the environment that you currently have.01:19:46So Cisco IQ is gives you a personalized, proactive and predictive insights based on your assets and assessments based on all your Cisco assets, right? Your catalyst switches.01:20:07Data center, let's say Nexus switches, ACI devices, UCS, inter site, all the devices that are managed by inter site. So you can onboard all your devices in Cisco IQ and get insights regarding your assets, which ones are covered by contract, let's say which ones are nearing LDOS.01:20:19That I'll have to act on in the near future or assessments, security posturing, right? Let's say I have some security hardening that I can do. It'll give you recommendations based on your configuration, your hardware, your software. Once you implement those recommendations, you can read.01:20:39From the analysis to see if now your that specific security parameter has been.01:20:59Fixed and has been made secure. So you have and of course there's AI assistant, so if you want a specific report to pull out critical assessments, let's say or configuration hardening, security hardening that you can do or best practices, you can pull out a report, you have AI recommendations.01:21:07For it. These are.01:21:26Some, you know, modules that currently exist on Cisco IQ, but this will be enhanced and more and more modules will be released. So this is to give you a one stop place where you can monitor all your Cisco assets and be ready because in the age of meetors and AI security.01:21:28Is NO longer an option, right? So how can you proactively understand how secure your environment is and take actions based on recommendations? And like I said, you don't need to purchase any hardware, you don't need any licensing as long as you have an existing contract with Cisco, which all of you should be having.01:21:48Thing by now if you have Cisco devices, you can easily onboard your assets on IQ. All you have to do is go to iQ.cisco.com, create an account with your cisco ID or CCO ID and if you have any contracts associated with that ID, you have an option to add your contracts manually through it.01:22:08And the assets that are linked to those contracts will start showing up, but contracts is a very static piece of data, right? Contracts will be associated with assets, so you will have visibility in terms of what hardware do you have, you know, is it conformant, is it nearing end of life, et cetera?01:22:28But addition to that, if you have inter site.01:22:48You can connect your inter site so that all the UCS servers, maybe even if you have epic or Nexus dashboard claim to connect it to inter site, you can monitor all of those assets and get insights based on assessments as well, which is what I spoke about security posturing, best practices.01:22:53Configuration recommendations and so on. But let's say you tell me, you know, Pratibha we don't have an inter site let's say in the US region or let's say we don't really use intercite, but I have a lot of other switches and routers which are not controller managed. How about them? Then you can install some.01:23:12Thing called as a Cisco IQ link. This is a VM that will remain in your premise that will talk to your non controller managed switches and routers and it will pass on that information to IQ. So this link is going to be in your premise, but only the insights will be available to you at iQ.cisco.com.01:23:32For you to act on them.01:23:52You also have an option to monitor all your TAC SRs, rmas, status, is it pending on Cisco or is it pending on you? All of that as well. That's another module in addition to assessments. So the steps are very easy. Go to iQ.cisco.com, create an account, add your contracts, if you have intercite connect.01:23:55If you want to add an IQ link to monitor your non controller managed devices, you could do that as well.01:24:16So these are the simple steps and just before we wrap up, just a quick question because we keep bringing up Cisco IQ because you don't really need anything else. You have everything that you already do to use its benefits. So just let us know have you.01:24:22Do you already have an account on IQ? Have you added contracts connected into site or is it something that you're planning to do in the near future? If you can just take maybe 30 s to let us know, ok, NO, not yet, please give me details. Sure, we'll share this. We'll share additional information with you. Again, to emphasize, you don't have to buy.01:24:41See any hardware, this is not a sales pitch, this is just to tell you that there is a product out there that Cisco has launched that you can use already as long as you have an existing Cisco contract. So you don't need anything else.01:25:00You don't have to buy additional license or anything to use this. So, do try this out. We'll share links with you already. I think I'll skip the next question because most of you here say that please give me some details. The next question was around if you have installed high q link, but probably we'll bring that question now later.01:25:15So resources and what's next, we'll share some of these resources with you. We'll share, you know, details where you can find more details about unified edge, the cvds, steps to onboard IQ. Where can you register for the next session, which is next Thursday where you can unmute and ask a question.01:25:35Questions live or if you want to watch past recordings, you know, maybe last week's session or upcoming sessions, we'll share the link for all of that with you. Please do continue to join and share your feedback. If you've missed to answer any question or if you think some question was not completely answered and you would still like to get in.01:25:55In Touch with us. That's the email ID that you will have to reach out to us on. Please write to us and we will get back to you one on one.01:26:15Also, when you and when you leave this webinar, you'll get a survey. Do tell us what other topics do you want from the data center technology that would be more beneficial for you. We've been covering inter site, you know, UCS unified edge. We have hyper fabric, smart switches, all of that line.01:26:25End up in the future. But if you have something specific where you're, you, you have the product already, you have already invested in it, but you're not very sure how to use it to the fullest, please share your feedback and let us know we read through each of your feedback so we can customize our sessions accordingly. So tell us what is it? How can we help.01:26:45You better and you'll see sessions lined up.01:27:05Along those lines. So with that being said, we can drop this session. Thank you everyone for, joining this session today. We look forward to seeing you next Thursday for the quick recap and live Q and A Until then, please reach out over mail if.01:27:09You have any questions for us or if not, we'll see you next Thursday. Thank you everyone and thank you Ronak, for your time as well.01:27:29