Unified Edge Computing: Extending the Data Center to the Edge

Oh, another important thing, please make sure to join our Slido for this specific webinar.00:00You can join it through the browser, going to slide.com and putting the hash code and the passcode shown on the screen. You can join it by scanning the QR code from the mobile or launching it from the Webex application if you are using it. It is important because we are going to use it for answering your question.00:18Questions throughout the webinar. So if you have any kind of question, you can put it on the Slido and we are going to answer you.00:37If there are like a big volume of questions, we are going to answer you throughout the day, so we are going to leave it open till the end of the day. So we pushed the 1st question to know more about you, where are you joining from? We see people joining from.00:46Germany and Yuki.01:06Welcome. And thank you for joining. So we're going to leave it a little bit to make sure that more people are joining our Slido and and importantly using it also for the for putting your questions.01:08Well, we have people joining from Switzerlands, Slovakia France, and more joining from UK. Welcome everyone. So we put the Also Mark put the Slido link in the chat.01:29In case that you, someone joined later, he can join through the the link and the passcode on our chat and the Webex webinar, the details are there. One more thing before we start our agenda for today, it's for Cisco IQ.01:45If you are not aware of Cisco IQ, this is a platform, a unified platform to manage your Cisco assets from coming from different Cisco solutions that you have. It's like to have visibility on the different solutions that you have.02:06It's like moving from the reactive approach to the proactive approach because it's built upon the AI intelligence, so it provides you with the anomalies, the vulnerabilities, the monitoring of your environment overall coming from different solutions that you have. So in case that you are familiar with the.02:25X cloud, it's like the evolution of the CX cloud. So in case that you already have CX cloud, you can migrate your account to Cisco IQ, in case that you are not using CX cloud, you can start by creating your account on Cisco IQ, going to iQ.cisco.com using your CCOI.02:45ID and then you will find itself guided account configuration as well.03:05And to know more also about Cisco IQ you can check the getting started guide. So, let's start our agenda for today. So, there are two main parts in our agenda. The 1st part is regarding the the concept itself.03:11Itself, what is the AI edge? And why you are introducing the unified edge and its differentiation from the legacy edge infrastructure, and then the Cisco unified Edge components, the main components regarding the chassis, the compute node, the networking, and the management of the.03:31Unified edge, how, how the management of it is different, what is the inter site fleet management and what it provides different as a different from the management or the normal management of inter site to the UCS servers. 2nd main part is led by Mark where it's.03:51The demo part is it will this part will be showing you the fleet management itself on inter site, how you can claim your unified to inter site how you can configure the workload definition, the deployment, the instance, which are the new concepts introduced.04:11Used by the fleet management and the workload versioning as well.04:31And at the end, if there are still unanswered questions, we are going to answer it to you whether live in case that we have still time. If not, we are going to answer you on Slido itself, and there will be a survey after you end the webinar. Please share your feedback with us to.04:35Improve the quality of the sessions.04:55So, starting with the AI at the edge definition, what's happening here? So, there is a, this survey was, was run that by 20 or at 2021, the percentage of the AI data processing was nine.04:5780 % at the traditional data center leaving only 10 % of the data processed at the edge. And then a data shift happened like a near complete inversion happened by 2027, it became 75 % of the AI data processing happening at the end and.05:17And 25 % at the traditional data center. So this shift or near complete inversion of the percentage of the AI data processing between the edge and the traditional data center didn't happen by coincidence or it's not just a trend.05:37It happened due to many technical requirements for the AI data processing. The main requirements that that resulted in this shift 1st the latency. So the AI inferencing requires sub milliseconds of response times which cannot be guaranteed.05:56By backhauling the data to the traditional data center or the cloud, but it can be guaranteed by processing the data at the edge. This is one main point. Another point which is the bandwidth optimization. So moving massive data sets from like coming from the.06:16The high definition video streaming, the sensor telemetry to the traditional data center over the WAN is very closed inefficient. But if you processed it locally, this would be or would call cut the cost big time.06:36The 3rd or the main technical requirement that led to this shift, it is as well the data sovereignty. So many customers now are looking to process the data locally to be aligned with the regulatory compliance and the security requirements as well. So.06:56So this data shaft happened due to those three main reasons. Latency, bandwidth optimization, and data sovereignty.07:15So what is changing? So indeed, we already have edge infrastructure. We already have the legacy edge infrastructure, but this current situation is not aligned with the changing that's happening with the AI data processing.07:24And due to many challenges that we have within the legacy edge infrastructure. The 1st challenge, it is the deployment in consistency, and this is because the legacy edge infrastructure usually the deployment or the configuration happening box by box or a single deployment. So whenever.07:41If you look at hundreds or thousands of location, if you want to scale, you will find each one is different from the another which make the management of those large scale number of deployments very hard in terms of management and in terms of configuration as well.08:01This is one challenge. Another challenge which is the product incompatibility because in the legacy, in the legacy edge you usually get the hardware from different vendors. So you can get the compute from one vendor, you get the switches from another vendor, and this.08:17At the end of results in compatibility issues and performance bottlenecks. 3rd challenge in the legacy edge infrastructure, it is the limited technical expertise because usually if we're talking about edge deployment, it is like in retail back office or small industrial office.08:33So usually NO IT experts are there so in case of any issues it require manual intervention for troubleshooting and so on.08:53The 4th challenge it is for the environmental constraints, and this is when you are talking about edge infrastructure, so we are deploying servers outside the traditional data center environment. So there are, rack space constraint. There is a power and cooling.09:04A Constraint as well. There is physical security constraints in the legacy edge infrastructure. So all of these challenges that we have it in the legacy edge is resulting in the operational complexity and security risk which cannot withstand with the high percentage of the AI data processing. This is the main reason Cisco started to deliver.09:24Over the unified edge to resolve all of these challenges by having unified approach for the configuration and for the management for hundreds and thousands of edge deployments.09:44So before going into the Cisco unified Edge, we have one question for you if you can share your feedback. What is the biggest concern that you have regarding the edge computing deployments? So from out of the challenges that we have discussed, whether we're talking about operational complexity deployment.09:56Moment inconsistency, a lack of technical expertise or security risk. What would be your like the main or the biggest concern?10:16Yeah, we can see the the bigger percentage till now for the operational complexity and for the security risk, now the security risk and the deployment inconsistency are kind of equal.10:43And then we have answered on the chat from Frank, which is for the deployment itself yes? Okay, so let's see how the unified edge address those challenges.10:59So, let me get back here to the presentation. What are the main components of the Cisco unified edge? 1st thing if you are going to talk about the chassis, it is 18 " deep, three rack unit, ok? Which the traditional data center server.11:15This cannot get into. So this like this parameters can get into shelf, can be on a wall, so it's very, it's very friendly and easily to be racked in retail back offices or small offices. As well.11:34All this chassis is built to be acoustic environment friendly because it runs in the range of the forties visible. So you can put it in the office and then people are working beside it normally which cannot happen in the traditional data center. And the temperature range, it can.11:54Can handle the temperature range from five tenses to 45 tenses.12:14Okay, so this is regarding the, the chassis itself. Regarding the compute node, we have the compute node already have the processing for the CPU and the GPU required for the AI inferencing with different use cases and different options for the GPU as well. And the storage is already part.12:20Part of the compute node, whether you can get the storage ioptimized, whether you can get I/O optimized or GPU optimized options for the compute node, and we are going to show you that in detail. The software, we have a ecosystem of the commonly used virtualization options already validated.12:40With the compute node, whether you're talking about VMware, and Nutanix, Nutanix Edge, Red Hat, Linux, so all of this are supported.13:00And then we have the ECMC or the edge chassis management controller, and this is mainly for the, the internal networking, the East west, and also the connection for the core networking and the top of track. We have it redundant, so to avoid the single point of failure, so.13:10For each chassis there are two ecmcs and the networking is integrated. So you will have or the chassis has 25 gig five internal switches already inside the chassis with NO need of external switches to handle the node to node connectivity and the ECMC to the individual.13:30Will IMC communication as well.13:50And then we have the 4th block which is the secure router node. This is the latest addition to the Cisco unified edge. And this is you can consider it not only an router, it's also a security device due to the different security feature that it has with the next generation firewall and the sassy features and.13:53And then also this is your like gateway to connect to the one. So those are the four main blocks of the Cisco unified edge going to them into details starting with the chassis.14:13So the unified edge chassis, it is the XE9 thousand 305 chassis. So as you can see here, this is the front of you and then the rear view. The front view, we can 1st talk about the five identical slots. You have five identical slots for the compute or the networking, so it can be mixed of compute and network.14:26Nodes based on the use case or based on the deployment that you want to configure. This is one main thing. Another thing which is the NFC tag, and this is like a game changer when you are talking about the unified edge, because this is something that enabled the technical experts to manage the unified edge. So they can just.14:46Have their inter site mobile application tapping the NFC on the chassis, start claiming the unified edge, pushing the already existing configuration doing the provisioning and also checking the inventory and the system health. So it is like a very.15:07The great feature for the non technical experts specifically. And then when we are talking about the redundancy, we have redundancy again on the level of the ECMC, we have two ecmcs, this is for the remote management, and we have redundant power supplies that can be configured.15:26For the grid power mode, and then the usual power button status lids that are used for monitoring the system health on the chassis. On the rear view, you can see here that there's the passive backplane, and this is supports the intra chassis communication.15:43This is in case that you are looking for, expansion in the future as well, so like you can do the communication between the, between the chassis. And then we have the acoustically optimized 80 mm fans, and this is the reason behind the acoustic environments that I was talking.16:00About about the range of forties visible because the small fans have been replaced with the larger 80 mm fans, so resulting in acoustic and friendly environment.16:20Another thing that is like a differentiator in the chassis of the Cisco unified edge, it is the mounting options that it have. So it has already the full post track and the two post track that are the traditional mounting option, but the, the one that really.16:36We make the XE 9305 chassis shines is the wall mount bracket. So let's say you don't have a sufficient rack space on on the floor for the chassis, you can mount the chassis on the wall and then the wall turns into server room. So this is like.16:54A very good feature that the chassis has been introducing. And also in case that you are having a shelf environment or desk environment or you want to put it on a counter, we have the horizontal positioning and the vertical positioning for the chassis as well. So those like flexible options make it.17:14Much easier for you to put it in your like retail offices, in your small industrial offices and so on.17:34So now that was like the 1st component regarding the chassis regarding the ECMC or the edge chassis management controller, we said that it is redundant and this is handling the data plane and the, the management plane as well. So you can see here that there is.17:44The dual 01:10 or 25 gig uplinks, ok? So it can be connected to the core, networking. If in case that you are going to use both uplinks, they must be a port channel. And then we have the serial console access for the recovery console. We have the USB C port for the local file transfer and supports the.18:04Bluetooth Ethernet as well. And we have the one or 10 gb management port for the KVM access. While the back end of the ECMC.18:24It has the 20 the 05:25 gig internal network internal switches that are used for the ECMC to the IMC or the individual IMC of the compute or the networking sled based that you have compute node or networking.18:33Node and then.18:54Also the ECMC to the ECMC communication, there is NO data traffic, ok? There is NO data traffic between the ECMC and the the another redundant ECMC inside the chassis. So as you can see here in this slide, the screenshot it shows the communication between the ECMC and the individual.18:55IMC at each slot, ok? And the connection of the ECMC to the external networking or to the core networking, so it handles both the data and the management as well.19:16Regarding the compute node options itself, so the compute it it integrates the compute, the storage and it has already the supportability or the validated blueprints with the different software ecosystem, as you can see at the bottom of the slide, whether we are talking about Nutanix red Hat VMware, ubuntu.19:33So those this wide range of virtualization options is already validated and we have blueprints for it.19:52The, the CPU itself in the node, it is in this is it has like many great features. One of them is the CPU native AI inferencing, the AMX, which allows the matrix multiplication on the CPU, and this is for the light.20:00AI inferencing workload, so maybe you don't need a GPU and this intel Zek sock can can be sufficient for your AI inferencing if you are talking about light workload. So this is one thing. Another thing which is the confidential compute, the TDX and SGX, this is for the security and the hardware based isol.20:21Features for your compute node as well. The integrated ethernet, and this is reducing the need for the external NICs and the scalable multi threaded cores options for 1220, and 32 cores for this Intel day on six sock. For the GPU options, we have two compute nodes currently for.20:41The Cisco unified edge. We have the XE 130 CM eight and the XE 150 CM eight.21:00The 130 CM eight, this supports half height, half length GPU, like the Nvidia L four. This is the one supported, and for the XA 150 C, this is for like denser AI workload. It supports full height, full full width GPU like Nvidia RTX proof 4500.21:076000 and the L 40 S for nvidium. And then there are additional GPUs upcoming on the roadmap.21:27So we have also node level personalization for the compute node. So you you will find that we have the storage optimized configuration, the I/O optimized and the GPU optimized. So this is mainly based on the use case that you want to run. So the storage optimized configuration, you can see that we have 4E3.21:40P.S drives up to 120 tb, ok? And then we have the one half half height half lens PCI gen, five that can take the L four or the LVD L four class GPU. Before talking about I/O optimize optimized configuration, we are going to replace one of the drive base.21:59With one half height half less PC identifier that can take Nick or DPU to make it eye optimized, leaving three mve e3.S drives for the storage up to 90 tb.22:19And then the 3rd option, which is the GP optimized configuration, this is like for the full height full length PC identified that can take the nvidia RTX proof of thousand 506000 and N 40 F This is for the denser, inferencing AI workload.22:33So we talked about the chassis, we talked about the ECMC, we talked about the compute node. Now it's talk it's time to talk about the networking. So the the left side we already talked about it, that we have like the 05:25 gig internal switches. This is like for the node to node connectivity, it provides you with high speed network.22:52Working within the chassis that eliminates the need for external switches. And when we say is eliminate the need for external external switches, this means like easier, and more simpler cabling and deployment aspect. So for the people that are saying that the challenge or the most biggest concern is in the deployment.23:12This is make the racking and the cabling much easier as well. Also it keeps the traffic local and fast leading to better performance for the node to node and the cluster environment as well. This is for the node to node connectivity, and then we have the secure router node, and the secure router node.23:32You already mentioned that you can consider it as like a security device due to the different security features that it has for the next generation firewall, the IPS IDS, the advanced malware protection, the UR filtering and the secure access service. It has 14, one gig ports and four 2.5 multi.23:52Gig ports. It is power over ethernet enabled, so there is NO need to connect the peripheral devices to another power resource because it it will provide the powering for it which is reducing the cost and more flexible. And then we have the 02:10 gig asset.24:12FP ports that can be connected to your WAN or the larger regional other regional network as well.24:32So, going, going out of the different components of the chassis looking at the bigger picture, what is the ecosystem of the unified edge and the different layers whether you are talking about the horizontal and the vertical layer of the ecosystem itself?24:45The Cisco unified edge, it has like the compute networking, storage, security, and it supports already the the common used virtualization options as we've mentioned earlier, and also vertically it supports the different use cases, whether you are talking about retail, manufacturing, healthcare. Those are the main use cases when you.25:05We're talking about the unified edge or four generic AI use cases. And the management of all of this ecosystem is done through inter site and inter site provides you with the zero zero touch deployment and the global visualization. This is through the fleet management that we are going to talk about it.25:25In the next slide.25:45So, after we now understand the different components of the Cisco Unified edge, we need to understand how to manage it and what, what is different, ok? So the fleet management came to resolve the challenge of having that, different deployment and different configuration.25:47Across a big number of edge locations. So the fleet management came to make it more simplified starting from the onboarding step. So here we can talk that starting from the NFC tag that we had a look on it on the chassis itself, you can just use your mobile application.26:07To claim your Cisco unified edge deployment into your inter site account and then start to push the configuration. In the inter site fleet management, we have blueprints delivering consistent repeatable deployments across sites and then you can push it across hundreds of sites at the same.26:27In time. So we have like different blueprints across different, vendors that you can already create a workload definition and deployment out of them so that you can push repeatable units of deployments across large number of unified edge deployments.26:47So this leads at the end to zero touch provisioning deployments. And then the automated life cycle management, and this is because in case that now I need to do changes to the deployments that I have it for another, set of edge deployments. How can I do this? We have like the configuration versioning, how you.27:05Can create versions out of the workload definition that you have created in order to do changes to your configuration and apply to different set of deployments. This is also supported into the inter site fleet management, and then at the end you will see the global visualization. As you can see here in the slide, this is a screenshot taken out of the.27:25Intersite fleet management, how you can see the different edge deployments that you have across the map. This is because you are mapping it to the locations, the geolocations that you have it and the path tags that is created for each eh edge deployment. And all of these details.27:45As we are going to show it to you or Marco is going to show it to you on the demo.28:05So how the inter site fleet management help me from help me from day zero to till day end. So from day zero again the zero touch provisioning, the claiming through the mobile application option for the non technical experts and then the blueprints and we can pause here because this is very important. So.28:11In order to be able to scale to a large number of deployments at the same time you need to have like a very structured base or invalidated designs for the deployments. So the fleet management of inter site provides you with validated design and this is across many levels. So across the hardware we have blueprints, we are talking about the policies, the service profi.28:31Files on the OS we have the different options for the vendors as you have seen, and we have blueprints for each of them for from the cluster which is a combination of the hardware part and the OS part validation as well. And from the application perspective, we have the blueprints whether you are going to deploy VM based application.28:51Applications or containerized applications.29:11And then the day one here, the NFC based claim by mobile app. So if in case that you already prepared your workload definition, the configuration, the blueprints that you want to push on your inter site and then you go to the non technical expert that is on site and tell him now you can start claiming the claim.29:14Claiming the chassis with the NFC. This is a very straightforward operation where you use your inter site mobile application to to tap the NFC tag, then you will be able to to add the target or claim the target and then the chassis is discoverable on your internet.29:33Site account.29:53And then when you are talking about day two, then when you're talking about yes, I have already a consistent repeatable AI ready infrastructure deployments that I can push across multiple sites, I can do workload versioning in case that I want to do changes on the existing workload definition and push it to.29:56Different regions or different sites so I can scale to a large number of deployments and push the configuration at the same time using the blueprints that have already used or configured.30:15How about the day in operations? The day in operations here, let's say that you want to to deploy or run new AI application that required GPU and the current compute nodes does not have any kind of GPU. There is NO need to forklift upgrade or getting a different chassis for this. You can just.30:30Just add a compute node that supports GPU into your current chassis and start using it and start running your AI application that require the GPU easily as I've mentioned. Another thing which is the proactive the proactivity of the inter site fleet management.30:50This is because it detects if there is a hardware failure and then it proceeds with the automated RME and then you receive the part that requires to be replaced before the failure happened which is minimizes the unscheduled downtime and provides with faster restoration of the service.31:10Another important pillar which is the integration of inter site that inter site can be integrated with Splunk and thousand eyes for having that complete visibility not only from the fleet management on inter site but from Splunk and thousand Eyes as well.31:29So one last thing to set your mind for the upcoming demo by Mark, which is the tasks involved in the fleet management. So let's say you want, you have your unified edge hardware ready and then you want to start managing it, configuring it and using the inter site fleet management. 1st, you need to make sure that you have.31:48The advantage license tier for inter site and then you have the sufficient privileges. We are talking about the workload admin or the operator role and then you have all the targets already part of inter site organization. You create the workload definition and we can like we can like do.32:08A simple analogy that you can consider the workload definition like this is the golden template. You, this is you where you pick which kind of Cisco validated blueprints that you want to use under that specific workload definition.32:28And then out of this workload definition, you start to derive deployments. So you start overriding overriding this workload definition with the parameters like e.g. the resource qualification criteria, putting the schedule for the maintenance window, when to push that workload to your hardware.32:42And then the instance gets created once you already have the configuration pushed to your hardware. The deployment occurs at the scheduled maintenance window that you have set in the deployment and then you monitor the deployments you will find the templates, profiles and policies are already created on inter site out of this workload deploy.33:02Definition and the deployment that you have derived. So the policies gets deployed, the operating system gets installed across the fleet, and then the loop starts again whenever you have a new hardware that you want to configure. And in case that you want to create a new version out of this definition, you can do the versioning of the workload.33:22So this is like the complete workflow to set your mind before going to the demo. One last question before Mark starts the demo.33:42Which feature you find out of the unified edge are most important to you? Is it the automated life cycle management or the zero touch provisioning with the blueprints, the global visualization, seeing the, the different unified edge mapped to the different geolocations and the path tabs that you have set for each.33:51Deployment or the secure device claiming using using the NFC tab. So, please let us know your feedback. We are interested to know what kind of features like is most important for you.34:12We can see that there is answer for the simplified onboarding. So this is actually a e game changer in the unified edge. So we are going to leave it.34:26Running for you to answer the question is answer for the automated life cycle management and this is very important as well in case that we are talking about large scale environments where because like for the edge locations, maybe we're talking about hundreds or maybe.34:42Be thousands of locations that needs to be configured.35:02So now let's start with the demo. Over to you Mark. Thank you, let's go to the demo parts. Give me just one sec. Okay, so you should be able to see my screen right now, yes?35:07So basically the 1st thing that we will do is actually what she will talk about. So let's say you are starting with a new deployment, you have shipped the new unified edge chassis to your shop to your edge office.35:24And you want a non a non technical person to to do the onboarding with you to help you with onboarding. So with the inter site app on the on the mobile phone. So the 1st thing that you do is, is to create a guest access. Then the the person on the smartphone can just go to top.35:40Targets just right there, go to targets, then you click on add on the top right and click on claim with NFC. Okay, so as we mentioned before, there is a NFC tag directly onto the chassis itself. So clicking on start.36:00Then it's asking me for location and for those are optional but I will I will still put a location. So let's e.g. put and click on done. So now the next in the next step.36:19Would be to put the smartphone next to the NFC tag. So just clicking on claim, putting it next to the NFC marker, and that's it. The target is now claimed. So what it's done, the person on site doesn't need to go back to.36:34To the to the hardware itself. Now we are just going to inter site. So this is your role as a technical person. So basically this is inter site, this is the the the management tool for our server. If I go to the my list of targets, I can see all the stores, all the unified edge that has been.36:51Added that has been claimed on my inter site account with those non technical person with the NFC tag, ok?37:11It's all of my all of my unified edge are there. If I go a bit up on the operate menu and click on unified edge, I can see a list of all my chassis just right there. E.g., let's click on the New York City one. If I click on it, I can see and I can have an overview of my of my chassis.37:18With the blade inside of it, with the server inside of it and we with the ECMC part also.37:38Right? So I have all of my targets, now I want to configure them, ok? So before I do that, what I want to do is to split them into multiple regions, ok? So I have stores all across the US, but let's say I want to split them into a.37:45U.38:06I want to split them into US east, US West and US central. So to do that, the 1st thing that I will do is go to the tag settings, ok? And I will actually create some tags. So what I will create is three tags, so as I mentioned before, I will create three tags related to using.38:07US east, US West and US central.38:26So what I will do is create those. So what I'm doing right now, as you can see there is multiple/between my my my tag. So basically I'm creating a structure, a tree structure where I will have, e.g., in global inside global I will have inside of it I will have US and then I will.38:29Have east.38:49So if I just click on create, just a bit for yes, it's there so now if I just click on global on US and east. So now I can see all of my, I can see my new tag. I will use automation to deploy the to create all the other tags in a quick in a quick way so now.38:50If I just refresh.39:10And open up again my US region. I can see my three main region that I want to use. So US, US central, east, and west. So I just created some tags, ok? That's the only thing we have done so far. Now what I want to do is create some locations.39:12Okay, so what I want to do is to assign some locations to my to my unified edge on each store. So I have put the city next to it on the name, but I need to I want to create some real location that will ease the process later. So clicking on add a location.39:31I can do each location one by one, but I can also use a file. Okay, so if I just click on multiple locations, I can browse on my computer to get a file. You can also download of an example file so you can just.39:50Use it as a, as a sample to 1st get it so I would just click on browse and use my CSV file that I've created before. So now what I've done is submitting my file, we can we can see that it's called location.CSV and it's showing me all the location that.40:05Will be created with my files, so we can see all of my, all of the cities, the the real address of my stores, the location, the GPS location and stuff like that. I will just click on add now. So all of my location have been created, right?40:25But what I want to do is now put the tag, the region tag on top of it, ok? So let's say e.g. New York, so I will just modify New York, right? And I can just put a tag just right there. So I will use my tag.40:43Global yes I will use my it's all new York is on the east coast. So I will use the US East and click on save. Now if I just scroll a bit on the right or maybe if I reduce that yes, we can see that on the right side of the screen, I can see my.41:00My tag US East just right there, ok? So.41:20Same as before, yeah, I I want to do it. I want to do that for all of my locations to quick the process. I will also use automation for that. So yes, now if I refresh and just yes, all of my locations are now the right the right time next to it. So yes, you can see the the one.41:24West east, central and and such. As I said before, now I have created my tags, I have created my locations, I have put my tags on top of my locations. What I want to do now is assign the locations to my real hardware, ok, so to my unified edge.41:44So let's go back to my inventory of my unified edge, and let's click e.g. on the New York City one again. So if I just call down a bit on the general information about that unified edge chassis, if I go just a bit down I can see location and I can see there is NO location.42:02Set at the moment. So I just click on set.42:22And I will try to find my new York location. So it's exactly the one that I've created a few moments ago. So clicking on save, now if I scroll down, I can see the new York location I can see the address that I have put before.42:25I can see also a map of it and I can and I can, and I can see the tag that was added to that specific location. So by just doing that, I have actually put the tag also on top of my chassis. So yeah, same as before, I will also use automation to put all.42:43The other locations in a quick and easy way.43:03So e.g. now if I just click on Atlanta, there should be a location assigned. Yeah, exactly. Good. So now I have assigned all the locations. I have assigned all the tags to all of my unified edge.43:07Hardware, ok? So now what I want to create is configuration, right? So what I will do is go to configure and go to workloads. So if you are used to inter site for a usual configuration.43:24Maybe maybe we we you might have seen in the latest updates that a new menu was added that is called workloads and actually it's used for the fleet management of the unified edge. So in the workload menu, you can find three main.43:40Instance three main tab what so what basically what we will create is a definition, then we will create multiple deployment out of those definitions and from the from the deployments the instances will be created and deployed onto the physical resources.43:56So let's start with the definitions. So when we create a new definition, so basically this is my golden configuration, ok? So this is the this will be the golden configuration for all of my servers, for all of my fleet of systems, all of my chassis also not only your server.44:15So let's call it, e.g., definition, ok? Clicking on next, and then I can select the blueprint. So basically the blueprints are pre provided help on the configuration based on your.44:32Use case, ok? So if I just want to create a bare metal server, I won't have as much configuration to do. Quite an easy because we expect you to do the OS installation on your own. But e.g., if I choose the ESXi server installation as a blueprint, I will also be able to provide an OS.44:48I will be able to provide the configuration of the system itself so that actually it will deploy everything directly onto my servers and all the whole configuration will be set in in only a few.45:08A few minutes. So for that example, I will use a bare metal server in all of my in all of my chassis I have two servers so I want a blueprint that is enabling to bare metal server. Clicking on next. So now the 1st thing that I need that I need to set up is the configuration for the.45:21The 1st bare metal server. So basically the 1st bare metal server in each of the blueprint in each of the deployments. So I will select what is they want. So all of my server one will all have the same.45:41In that way. Same I need to add some VLANs, so let's say my my server one is only talking with the VLAN one. Let's use that one. Then click on next. Now it's the bare metal server two. So same as before.45:58Configuration for my second bare metal server in each of my unified edge chassis and all of those will talk with the VLAN 20. Now I need to put the configuration for the system itself. So basically for the chassis or the management for the network.46:13So the 1st thing that I need to do is to create and to assign an IP address pool. If I already have one, I can just use it as is, but I can also create a new one. So let's create a new one dynamically so that even even if you have NO configuration already done in inter site is quite easy to do, I will show you. So clicking on IP.46:31So let's call it IP pool demo and let's put that specific mask, let's say and put that gateway, then that.46:52DNS, of course it's just to make it quick and to show you just a demo. So I will just see I'm just creating an IP pool that will be fetched from the deployment to get an IP inside that block. Here you can see I'm putting I'm choosing what is the 1st.47:11The 1st IP in my block and what is the size of the block, the size of the block. So clicking on on next, I want I don't want IPv six and I will just click on create. So I have done dynamically created my new IP pool and you can just, you can see I've been assigned directly in there.47:31Let's say my inbound management ID is 30, then what is the firmware of my chassis itself? This will be the firmware. Now it's also asking me for local user policy. So yeah, this time I already have I already have created it. You could have also created.47:51One, but I will use the one that I've already created. So basically inside that local user policy, you will find.48:10What is the main users, the main user or the main users you want to use local connectivity on it? So let's use that one and let's scroll down a bit putting the DNS DNS for my.48:16For the for the chassis, then enabling the jumbo frame, let's say. And now inside the network configuration I just need to provide the free VLAN that I have set before. In that way.48:34It will enable the communication with those inside the chassis itself. Of course, there is even more optional settings. We won't go into the details of all of them. There is gamble settings, there is some power settings.48:48And such. So let's click on next. Now we have a summary of everything that I want to do what a summary of my definition, and then we just click on create and validate. So now my definition is validated, right? So what I will do.49:03Is create some deployments, so as I said before, we start with definition, then we create some deployments and then some instances. So I will create some deployments.49:23So I don't want to create only one, I want to create multiple ones, so I will click on multiple deployments and click on start. So it's asking me what will be the name of the deployment, what will be the name of the instance that will be generated out of it. So let's say the.49:34Deployment name will be quite straightforward and I want one deployment, I want one deployment per region, right? So let's put the region that I've said before, so east, west, and central, and then I just need to provide also name for my instance.49:53What will be the prefix of all of my instance? So yes East and West and central. Let's start with that and click on create. So now if I go to deployment.50:10I will be able to see the free deployment that were created based on the definition that I've created before, right? So those are just definitions in in some ways, right? What I want to do now is assign that deployment to.50:28Hardwares. So to do that, I will just click on one of them, I will modify it. Actually there is even though we have a blueprint, there is some specific thing that you can change inside those specific workloads because.50:44E.g., in my, in my example, we have multiple regions, right? So, let's say you don't have the same schedule maintenance on all of those regions because they don't have the same time zone, the same time zone. So let's say you want to have a schedule maintenance that's different between the regions that the set that is a setting that is.51:01That need to be provided inside the workload itself and not on the definition. So what we will do is actually put a resource qualification on that workload. So basically this is a way to to set a rule that will match current and future hardware resources. So believe.51:21I will just put a rule that will select for me automatically what are the the the unified edge chassis that I want to use to for that I want to use for that workload. So I wanted to match my path tag. So basically the tag that I created before.51:41So that's why it's useful. So yes, I want all of this, all of the unified edge with that specific tag to be associated with that specific workload. So clicking on next and then clicking on next again.51:59And save on and validate. So yeah, so the validation is ongoing, and now you can just click on prepare to deploy, and yes, so you can see now the status of.52:16Of the of the deployment has changed. We can see that we have a staging at five and we can see that we have some, we have some instance conformance at five. So basically has found five resources to allocate that specific deployment alright?52:33So, yes, so basically if I go to instance, now I can see five instance. Before that, it was zero. And each of those instance is actually linked to a specific location. So my store in filter, my store in DC, my store in New York will get those specific.52:51Deployment are linked to that specific deployment with those instances. So I can just click on that and actually go to the setting and click on deploy. So now I will deploy my configuration. Inside that deployment definition, you can see two rollout strategy.53:13So the canary one is, e.g., if you want to only update a specific set of, of, of systems. So e.g., if you only want to update the New York City store 1st, you want to only, you want to update this one 1st to check if everything is going fine.53:31And then we can share that specific update or that specific configuration with the rest of the stores, ok? That's an example. But in our example we in our, in our demo, we'll use the batch that will actually spread it to all of them.53:50So now clicking on deploy. I will do the same for the central and for the west. It will only take me a minute. So same as before scrolling down, going to the resource, putting some qualification can be qualification for the location for the same number or for the path tag, but as I want some regions.54:08And not only.54:28The workload per per city, e.g., I want region based configurations so I will use the central one clicking on next and validate. So in 1 s I will be able to deploy it exactly.54:30So same as before, clicking on the setting and then deploy as a batch. And I will do the same for the west, clicking on on edit.54:49Click on next, next, and then add my qualification rule this time on the west tag. Oops, I was a bit too quick.55:04And now clicking on prepare to deploy yes. So going to the setting and deploy. So you can you can see already on the top right that I have a lot of actions that are ongoing, a lot of creation of of creation of the instances creation of the of the profile of the policies and.55:24Such, so actually what I have done is.55:44I have deployed the configuration on my, on my, all of my servers, so I have eleven of them so I have deployed all of that all of the configuration in eleven on eleven stores in only a few minutes. So of course it's not there yet. It will be it will.55:48Will be a few minutes to get everything in the right in the right process. But yeah, basically the way we want to to to show it to you is that only only a few minutes you create your definition, you create your deployment.56:08And you can spread the whole thing quite easily. So I guess so if e.g. if now if I go to, if I go now, if I go now sorry if I go now to to profile, ok, so if I go to now profile.56:21I can see that the unified edge profile have been automatically created, ok? So they are linked to the workload that you have created, of course, and we can see that they are still being configured and spread and spread into the various locations, right? So basically the edge the unified edge profile is a configuration.56:40Before the chassis itself.57:00And we have also the server profile that are not there yet because it's a bit too too soon we need this a bit more time to to see them appearing just right there. But in just a few minutes we will start to see also the server profile just right there that will be actually the configuration for the server itself, but we're created.57:02With the workload clear with the workload definition, deployment instances.57:22But you can see just right there. So yes, I want to show you also another demo that will be on the. So let's say that you have created all of your all of your configuration and you want to update it by.57:27Example adding a new VLAN or something like that. It's also quite easy with the fleet management so you just go to the definition itself. You will go to manage version and actually you can see we can create multiple versions of the same definitions so let's say I want to create a new one with the source being.57:47The 1st one, same as before, I will go to next and I said that I wanted to add my new VLAN, so clicking on add VLAN and now I'm reaching the VLAN 40 e.g., clicking on add, clicking now on next, and going to also change the configuration for the VLAN two.58:07Putting the 50 VLAN, and now I'd also need to just modify the VLAN configuration of the chassis to add the new specific VLAN that were set up before. So clicking on next.58:27Scrolling down to the last part and click create and validate, you can see it's the version two of the definition, right? So I'm setting up as the new main I'm setting it up as the new preferred version and I can also select the new.58:44The deployment, the strategy, the rollout strategy that they want.59:03For that specific version. So I want to spread it as a batch and click here on save. So now I can just close that specific window I can see also the list of all the versions that are there, ok? So you can see the list. I can see which one is active and which one is not active.59:08Receive anymore?59:28So you have also a way to roll out actually by doing that, and we can see that now all of my deployments are in changes, meaning that in a few minutes it will actually spread the the new configuration into all of the instances, and yes, now we can see everything is in progress. If we go to the request tab.59:30We can see all the deployment that are ongoing, starting to be ongoing actually.59:50And if I go back to instances now, I can see that all of my servers are starting to get new the updates of the contents that have set in the definition. So other set up as a batch, of course it's all under all of them at the same time, I could have done it in a, in a, in a canary way.59:57And yes, so now we can see all of them are starting to be updated and after a few minutes everything is is in status, ok? And if you go to unified edge menu you can see also all the new workload definition and the unified edge profile assigned to those specific.01:00:17Specific unified edge chassis, you can see the profile and we can see also all the servers are configured with the server profile that and that's it for the demo parts.01:00:35Yes. So basically, we were just we wanted to showcase the fleet management. Yeah, this is what we just showcase, so it's a way to easily provide a configuration for the whole whole fleet of system.01:00:52And also we have some easy way to provide help. E.g., as I mentioned before, if you choose the blueprints on that, we provide a template to easily create also the full configuration of the system. So in that way you have the whole thing configured in in a really easy way.01:01:12Of course, for people that are used to the usual way to configure IMM servers, let's say, if you are used to also already configure a server with the server profile with your own policies and search, it's also available for the unified edge, of course. So you are not it's not mandatory to.01:01:31User fleet management is just the new recommended way for those unified edge.01:01:51Servers. And before we finish, we wanted to show you a comparison of the legacy edge and the new unified edge with the specific system. So basically in the legacy edge, as we showed before.01:01:57We add a lot of requirements so basically of course you add your router, but you also had maybe multiple data plane switches, maybe also one management switches. For all of those switches we you need it, you need one power cable. And of course you also need all the network cable between those and the router, then.01:02:13Between those switches to the servers, if you have multiple servers, those servers need also to be connected to the power with the power cable. So by just doing that, you have a lot of cables, you have a lot of power power utilization. So this is the legacy edge.01:02:33And that's what, why we also we wanted to have a new way to do that with only one chassis.01:02:53That is providing you all of that. So in the unified edge chassis, you will have NO NO cable outside of just set up your to just connecting your chassis to the rest of the network and of course to put the chassis with the, with a power cable, and that's only the.01:03:00Only thing that you need because everything else is already inside the chassis, you already have the the whole management part you have all the switching part, the compute part, everything is already inside that specific chassis. So before we leave, we have one more question for you.01:03:20So how confident are you in adopting unified edge solutions in your organization? So still on Slido so if you can answer that question and also for your information before we leave.01:03:41Once you will quit that specific webinar, you will get some survey to to to to complete. So please do do do do complete it because from those from your feedback, we can also.01:03:58Get better and we can we can be better for next webinars yeah ok thank you for the somewhat confidence. I I would say the best way, the best way with those kind of solutions is also to.01:04:15You know to put your hands inside of it because yeah it's all the the only true way to learn, to learn exactly how to use it. So I think we can finish on that thank you thank you everyone to.01:04:31For being there, don't hesitate to, to fill out the the survey. It will only help, it will help us and help you for the future. So yeah, please do so. And yeah, thank you very much and have a good good day.01:04:47