 אצד פלוימנס, קונטיינרס, ובאולטנסי. future of networking. רגע, רגע, תודה רבה כתוביות. אני אלמוג אלפאסי. אני קלוד אכיטקטי טרדאט. אני רשמל בבקשה ובאמצע ובאמצע, קלוד פלוימנסים של אופסור סטכנולוגיקס. אני עושה בכל סקטור, FSI, TELCO, וכתוביות. וכי היום אני אעשה את תלכו פרקט. Our project was for one of the top TELCO companies in our region, having several diseases and providing services across the state. Like a lot of our customers, multiple teams work on different platforms. For example, the network team on physical appliance sales, network appliances, their DevOps in development or virtualization stage, platform story, and so on. So our goals was where deploy one centrally managed platform for all the teams. Network functionalization at the edge and provide a container platform for CNF workload cloud native network function. Let's take a look at the NFV network functionalization to CNF cloud native network function journey. We can see that from the network appliance stage, physical stage, organization move to virtualization stage and after that to container platform. But what can we do if we have virtualization capabilities without a container platform? Actually, we stuck in the middle of transition between VMs to containers. NFV and CNF cloud native network function give us some advantage. For example, a general hardware instead of a dedicated hardware, better scale, multiple function on a single server and edge compute that I will talk about later. So NFV and CNF platform options was, where sorry, one virtualization platform as an infrastructure and on top of it container platform. Next, two separate platform, one for NFV and one for CNF. Or our choice, option three, remember one of the goals was one centrally managed platform. So option three, a container platform is an infrastructure that can run CNF workloads and on top of it, Qvirt for network VMs. What is Qvirt, right? Our virtualization is an operator that allows us to run VM inside a pod. Actually VMs and containers can manage on the same platform. For example, CNF workloads can run with network VMs like router, firewall, SBC, load balancer, et cetera, side by side. Let's go to the next step and the neighbor solution at the edge. As close as possible to the end users and ensure critical operation continue with limited connectivity. There are several options and levels for edge tier. No one size for all the organization and actually we need to choose the relevant tier for the situation. There are several deployment options for edge clusters. We want options that can run without a connection to the central DC. Because of that we chose three node cluster, cluster that provide a chain and each server will be master and worker and can work without a connection to the central DC. The second option, the second choice was single load, SNO OpenShift. One single server that will be both master and worker roles and also can work without a connection to the central DC. We use both deployment depending on the edge tiers. So we realize that we will need a lot of edge clusters and one central managed cluster can be critical for us. Our answer was ACM Advanced Cluster Management. As an operator that we installed in the central DC, this cluster called hub cluster and from it we deployed, upgraded and managed edge clusters. The edge cluster is called managed clusters. One more advantage we received by ACM was the deployment process. We use assisted installer tool or service for deploy edge clusters, bare metal clusters. Without assisted installer the deployment process can be a complex process. So actually we have a hub cluster in the central DC with ACM and from it in no time we deployed edge clusters as needed. Three node or SNO is closed as possible to the end users. So after we solved the options and technologies we chose, let's see how everything works together in our project. We have two central data centers, one for the outside disaster recovery. In each central data center we have hub cluster with ACM and from it we deployed edge clusters that will run the workloads with special requirements that I will explain soon. So the special requirements or the buzzword, right? Let's take a look at the severed edge. Basically bare metal node with ODF, OpenShift Data Foundation and local disk for storage solution, CPU pinning, huge pages and with multi-CNI we add additional VLAN and SRIV plugin directly to the workload. Our workload was VM inside a pod by OpenShift authorization and provides network function like router, firewall, SBC, etc. So we start here and our achievements, one platform for all the teams and the workloads. VMs capability with cuVirt OpenShift authorization for network function, firewall, SBC, etc. and containers capability ready for 5G and CNF, the next generation workload. Of course, one centrally managed platform for all the edge clusters. Our time was limited but feel free to contact me here or by social media. Thank you very much.