[MUSIC] Good day. Hi, my name is Fulcan Fong I'm the Cloud Solution Architect in the CSP TSS. Today, I will be presenting to you the 100 level overview of cloud orchestration. The agenda today is that I'm going to talk about the evolution to cloud orchestration. What are some of the key orchestrators out there and the components or the features of clouds orchestrators? Then we'll talk about the intel impact in the cloud orchestrator community and then the benefits of cloud orchestration. Thank you. And here we go. So on the first slide here, we've got to look at the evolution to cloud Orchestration. When you look on the left side in the pre cloud or pre software defined error. Orchestration was basically a centralized function, right? That involved a lot of people to coordinate the silo of resources that is compute storage and networking in order to support a or any application that runs in the data center. So automation was really limited at that time and it requires a substantial amount of manual labor to manage the different systems, right? That is the networking equipment, the storage equipment and also the, service themselves. And on top of that during the pre cloud era, there's a lack of a unified view of the different resources that are in the data center. So that actually turns into a challenge for reporting and measurement because whatever reporting that is being provided to coordinate all the resources, they are not real time. So it is a very difficult job to manage these resources. And so the additional challenge to is in this pre cloud orchestration era is that or the not only the computing storage and networking resources operating in silos the workload themselves, they are basically island of information so they don't have their own resources. So this all opposes as a big challenge. That leads to cloud orchestration. And this is where you see on the right hand side of this slide right? Where cloud orchestration can provide all the functions to manage the different types of resources that is to unit to provide and unified view of the different resources, compute network and storage. And then scheduling them and provision them and launching them for running any type of workloads in the data center. So in this next slide let's talk about what is cloud orchestration? So it is the management and coordination of cloud resources. As I briefly mentioned in the previous life. This is also a self service pay as you go model where the user or the subscribe or the class subscriber can request cloud services to run workloads without the involvement of IT. Make it very easy to consume the cloud resources. Cloud orchestration is a it's basically using a lot of the policies and procedures to govern the automatic provisioning of the different resources for the different workloads that is allowing that these resources to either scale up or scale out depending on the workload requirements. And finally there are mechanisms such as the schedule that are in place to ensure a timely and efficient delivery of these requested resources to run the workload from end to end. This next slide, we can look at the different types of orchestrators that's up there in the marketplace. So there are those that are open source which are Kubernetes, OpenStack and OpenShift. Then there are those orchestrators that are developed by the SVS or ISVs from like or Red Hat and then of course the Hyper Scaler such as Microsoft measure, AWS, Tencent Alibaba, IBM cloud, they all have their own version of cloud orchestrators. Okay, so depending on the type of cloud your customer has, they may choose one or more of these orchestrators to manage the cloud problem. So let's get really want this definition of some of these clouds orchestrators I think opens back, it will be a very good example because it is probably one of the most popular orchestrators out there being used by enterprise customers. Com Service Providers and also some CSPs too. So the mission of like, let's say, OpenStack right, is to produce a common open source cloud computing platform that it's easy to use, simple to implement and provide interoperable between the deployment and the work well at all scales that meets the needs of both the customers and also the cloud service providers for both public and private club. Now I mentioned earlier open stack has been adopted by many, many customers today, Kubernetes, it's also gaining a lot of traction because containers is being also adopted by many of these users. And do keep in mind Kubernetes and OpenStack, they can coexist they can work together to read the benefits of all the cloud resources that's out there. So on this next slide, let's look at what are some of the key common components shared by most of these orchestrators? Actually all of these clouds orchestrators. So looking at the table here, right? The first three rows, of course compute storage and networking, those are the resources that needs to be managed and they are available on all of these orchestrators. Then the other important part will be the schedule. The orchestration piece, right? That will be used to manage, the cluster of resources, right? The workflow messaging, reservation of these resources setting up the arms, right? Then, there's also the component of shared services, right? Share services, including like key management, identity management, application data protection, okay. Doing indexing and search and also within these orchestrator. There's the the hardware lifecycle management component, right? That is the provisioning of bare metal or the life cycle management of the different types of resources such as accelerators. Right? Then there's also the application lifecycle management, right? The availability of instances to run in the cloud. All right. In order to meet the specific service level agreements. There's also the application lifecycle management of the application catalog, the software lifecycle management of the orchestrator itself. Right then there's also the lifecycle management of the backup resource and disaster recovery. And then another key component of the cloud orchestrator will be the dashboard. And I'll show you some examples in the next couple of slides. And then finally three key component here is on security. Let me just give you some examples of what kind of the cloud orchestration. That's what looks like. And this is from open stack. Right. So the next few screenshots shows you that any users operators, they can actually see the different resources that's available to them. Or they can request specific resources, such as virtual CPU storage or memory, that they can request those resources and run workloads right? In the cloud. And then there's also this dashboard on open stack that shows any of the users the network topology. Then in this next slide, this is the snapshot of the AWS management console rifle during cloud orchestration. So, definitely a different look and for you,but the functionalities are very similar to open stack or any other orchestrator. So the users, they can configure and compose the resources to run different cloud workloads, but more importantly, as a hyperscaler or CSP right. AWS is able to give their customers the additional ability to subscribe to AWS cloud services. Right. So as we've just seen the examples from OpenStack dashboard and also the AWS dashboard on cloud orchestration. So let's look at why it matters to Intel. Right. So from inter perspective, we have actually contributed to many of these open source orchestration projects like OpenStack and Kubernetes. And it can also help our ISVs and also cloud service providers to optimize their thought orchestrators in order for these different orchestrators to expose the underlying intel features behind the curtain of the cloud. So then the customer workloads can get equivalent performance right? When running in private, public hybrid cloud environment. So here are some really good examples of the intel technologies that Intel contributed to these plow orchestrators. Okay, so the first one is the note feature of discovery. So what it does is it actually detects the hardware features available on every node, right, and advertise those features. So then the capabilities of the platform can in turn be used to facilitate some, I would say intelligent scheduling of the workload to be run in the cloud. Some of these features are like, you know, CPU ID, SRIV or the networking side, resource director technology, right, RDT, turbo boost input and output memory management. SSD storage and network topology and also maybe the features of PCI availability. So, in addition to the node discovery feature that is talked about, there's also the accelerated pooling, which is basically a management framework for the pooling and sharing of accelerators right inside the cloud infrastructure of the service provider. Right. So accelerators that include FPGAs, GPUs, NVME, SSDs, DPDK so those cannot be shared right by the users. And then in addition to that, there's also the enhanced platform awareness. So what it is is that because there's a growing demand for the cloud OS right to have a greater awareness of the capabilities of the platform that it controls. So, enhanced platform awareness facilitates kind of better informed decisions for the operators, right to do a better VM placement or matching to allow workloads to land on the platform that best run those workloads. Right. And actually provides better intelligence scheduling of, running these types of workloads. So, for example, in the enhanced platform awareness feature, it can automatically launch a let's say, a crypto workload on a platform that has the hardware based crypto accelerator. Right. So making things a little bit easier for both the operator and also the users, but also for the cloud service provider. Right? They can actually create value added services based on this type of enhanced platform awareness features there on the software guard extensions, Intel. SGX that offers the hardware based memory encryption. Right? That isolate a specific application code of data in memory and I will give you more information in a later slide. And then another big contribution that intel provides is on the counter container. So basically is it provides a significant improvement on the security and the isolation on container workloads. So in the next few slide I'm going to show you the software stack for the different technologies that we contributed to communities and also to the open stack orchestrators. Right? So this first one is again I mentioned earlier, is the Intel accelerator polling in the cloud orchestration. So again, the accelerated pooling right basically allows the service provider to offer these shared accelerated resources for the for the customers instead of offering dedicated accelerator to each customer's. So this way the from a service provider standpoint, they get the economy of scale right, better efficiency and cost saving that can be passed on to the customers. Okay. And this is a more viable solution from a financial point of view to the service provider. Another technology that I did not mention is the Intel Quick Assist Technology QAT. So it actually, this technology accelerates and compress cryptographic workloads by offloading the data, the hardware that are capable of optimizing these functions. So in other words, again, the service provider can create differentiating services for the customers using this technology. This can be exposed here. The cloud orchestrator is a new service back to software guard extensions. So with SGX right for open stack and also kubernetes you know, these orchestrators, they actually provide the all the ingredients necessary to deploy privacy and integrity solutions, but it is up to the service providers to deploy them security. So a couple of good examples here is both IBM cloud and Microsoft Azure. They've actually created new services based on or using Intel SGX. So in the case for Microsoft Azure, the services, what they call it, the confidential computing. And finally, in addition to our process with technologies that I just mentioned, our memory technology such as the OptaneApp direct mode, this is also available to be used and exposed. Being different types of clouds orchestrators. To summarize this session on cloud orchestration. The benefits of cloud orchestration is actually quite obvious for both the users and also operators. Because cloud orchestration enables self service in the cloud that is to pay as you go mobile, right? Because the user, they can actually request the specific amount of compute network or storage resource to run the workload in the south. And then the cloud orchestration of the dashboard provides a unified view of all these resources that's available to the user's cloud services that they can subscribe to. And on top of that is the capability of reporting the usage. So again ties back to the pay as you go model. And then another key benefits is because of the scheduling and then the rules that are being set up via the cloud orchestrator. It allows a lot of the automation on a lot of the simple and small operations within the infrastructure to minimize any type of human intervention. Okay, so automation is another great benefits of cloud orchestration. The other benefit is better service level agreements are being met or improved uptime. Right? Again, because of automation. Because of the reporting alerts that the users of the operator can see from the dashboard. Right. Because they are all event driven or policy based. So, can the remediation and the downtime can be minimized. Right. And then another point here is as a part of the summary is that the clouds orchestrators actually exposes a lot of the Intel features or the Intel goodness. That allows the cloud service providers to create mute and differentiated cloud services. And because of cloud orchestration software developers are able to develop cloud native applications more quickly because they can request those resources for developing there cloud into software. So, thank you again for your time to listen to this cloud orchestration session. Hopefully you learned something. I've also provided a reference page that you can actually look at more information regarding Kubernetes, OpenStack or any of the Intel technologies that impact cloud orchestration. Thank you. And you have a good day.
