Cisco Systems, Inc. (CSCO) Earnings Call Transcript & Summary

Welcome, everyone. Thank you for joining me to meet the new Cisco UCS X-Series modular system powered by Cisco Intersight. I'm James Moore, Cloud and Compute Sales Manager for Cisco UCS and Cisco HyperFlex Asia Pacific, Japan and Greater China. I welcome direct contact from you. Please feel free to reach out to me, tell me about your experience with your Cisco UCS and Cisco HyperFlex. Or if you're new to the Cisco world of computing, I look forward to understanding your requirements and answering any of your questions after this webinar. I'm based in Sydney, Australia. And I have the pleasure of connecting with customers anywhere, as far as Auckland, New Zealand to Mumbai in India. Now at Cisco, we understand that you want to increase your digital agility in a cloud smart world. You want to have your consumers and your business customers have an exceptional experience. An exceptional experience requires application assurance and performance. Now that application might be at the edge. It might be in the core. It might be in the public cloud. You might be building a private cloud so that you've got hybrid cloud capabilities. You want higher levels of automation. You want greater levels of observability, and you want it to be cloud native as applications move to a more container-based environment. And it's happening fast. IDC estimate that in the next 2 years, the number of applications being deployed within data centers and Edge locations will increase by 300%. But 61% of customers are telling us that their manual processes and siloed teams approach are real bottlenecks in terms of achieving this pace of application delivery. And 75% of Global 2000 IT organizations are clear. They want to adopt higher levels of automated operations and practices so that they can deliver at the pace of the business. So a question for you. Is your IT team an innovation enabler? Are you able to say, "Yes, we can," or is it, "I might get back to you on that?" Yes, we can, and yes, we will. And that's what the technologies are all about today that we're going to be discussing over the next 30 minutes. Now in order to move to yes, we believe that customers are crying out for a cloud operating model. The pillars on the left: high levels of automation; clear line of sight of observability; definitely cloud native; bridging between private cloud and public cloud; services such as single sign-on, multi-tenancy, RBAC, programmability through SDK; modules such as orchestration, optimization; programmability as infrastructure as code; the ability to deploy containers on-prem from a turnkey perspective; the ability to see all my virtualization assets on a single screen. They're independent of what brand they might be. The ability to deploy Kubernetes on-prem or in the cloud and manage from a single console. In terms of infrastructure management, show me my inventory. Is my ability to log a support case? Is my infrastructure compatible with the operating system that's running on it? And is my ability to maintain the firmware updates that are required to stay pace with compatibility or resolve problems? These are all aspects of a cloud operating model, a hybrid cloud operating model that Cisco has built, and it's called Cisco Intersight. So Cisco Intersight, available today, delivers a hybrid cloud platform. In the next few screens, I'm going to take you on a journey through what are these modular SaaS offerings that you can consume at a pace that meets your business requirements. So let's log in to the Cisco Intersight system via my Cisco ID that's linked to my Intersight account or via your organization's single sign-on. I log in, and I'm presented with my device's screen. My device's screen is a summary of all the hardware devices associated with my Cisco Intersight account. Now these devices are Cisco devices. So we're not looking at third-party servers and controllers within this environment. This is my Cisco device view, my server health summary, my server inventory, what models I've got, what's the support status, all of this information on this one screen. Let's say that I want to have a look at the next level of detail of my servers. Are they healthy? What's their management IP address? What if I wanted to log a case with Cisco Technical Assistance Center? You probably know of them. We call them TAC. Right mouse click, open TAC case, tech support files, and all of the details associated with this server get transferred to Cisco TAC to assist with a more timely resolution of the query from the customer. Let's say that I want to understand, well, what are the alarms and notifications? Why are they red? Why are they orange? Why are they blue? Click through on the alarms, and it will tell me which devices within my environments need attention. And what is the attention? Actually leading me to take the action that's required to move from red to green. What about important advisories? Here's one, Cisco fabric services denial of service vulnerability, that sounds serious. Click through on that one, it will show me the devices and the action that needs to be taken. Now unfortunately, these things do come. These advisories, these security briefing notices, they affect all vendor infrastructure. It's more about how do you deal with them and how do you get notified. How do you remedy them quickly. And through Cisco Intersight, you can do all of that from this one console. What if I wanted to see my virtual machine summary? So what have I got on my screen here? Okay. These are the operating systems, the guests OSs, what cluster are they associated with, CPU utilization, et cetera. I can start/stop virtual machines. In my next screen, I can optimize them and orchestrate them. Okay, these are orchestration workflows, orchestration workflows that discover the targets, whether they be servers, hypervisor managers, operating systems, storage assets, cloud assets, and I can orchestrate them with a workflow. So here is a view of a workflow. And I've created this workflow by dragging tasks from a task menu into a sequence that delivers the outcome, all of these orchestration workflows available within Intersight for an IT administrator to call upon as and when needed. Wouldn't that be great? Well, available today. What if I wanted to optimize resources to ensure application performance? Once again, Cisco Workload Optimizer discovers the targets that I have the access rights to. Then it understands the interdependencies of these targets. They might be containers, virtual machines, storage systems, hypervisor controllers. Once Workload Optimization understands the interdependency of these targets, it builds an interdependencies map and then comes up with recommendations that are known as actions. So you can see on the screen here that these recommendations are in an AWS environment because all of these modular SaaS offerings that I'm talking through now are hardware-agnostic. Only the support functions were linked to Cisco specifically. These are independent of the underlying infrastructure. When I'm happy with these actions, I can then automate them. So essentially, I move to a human-approved action to an automated action taking place as and when a policy determines it's needed to ensure application efficiency, application resource, assurance, all available through Cisco Intersight. So that's just a brief snapshot of some of the capabilities of the modular SaaS offerings within the Cisco Intersight hybrid cloud platform. Now Intersight is available today. If you've got Cisco UCS or HyperFlex devices in your environment, you can claim those devices today. You can have the sorts of capabilities of the Intersight infrastructure servers, and then you can start to explore the other modular SaaS offerings that I've just talked through. Many of them are available via a trial basis. Or if it's time to go a bit deeper and understand proof of value, your Cisco support specialist, your Cisco account manager will be very happy to talk to you about exploring such a proof-of-value experience. Let's now move on now to infrastructure, and I think you probably agree with me that today's infrastructure has been built to service unique and now complex architectures of a single workload perspective. So we've typically got blades for lots of virtualization. We've typically got rack servers, a lot for VDI. Some customers choose rack servers for virtualization themselves; specialized I/O or computational-intensive calculations, high cores, lots of I/Os, not much storage; specialized storage subsystems to serve compute, typically with tiers of storage performance, using all flash, all NVMe for hot data and then spinning disk for warm/cold data. And many of you, I'm sure, have purchased very expensive dual-socket or 4-socket dedicated AIML type devices in order to do that sort of processing. But are these resources, are these investments of yours highly utilized? Can you deploy a rack server as a blade server? Well, not really because blade servers don't traditionally have enough storage capacity. Can you use the GPUs that are in your dedicated AI box to do some VDI processing for a CAD VDI workstation? Well, no, you can't because that's a separate piece of infrastructure. Do you have separate management tools for all of these different silos of workloads? I'm sure you do because over the last decade, you've purpose built -- you've purpose bought platforms to serve single workloads because that's the way it's been done. So no shame on any of this. That's what's been available. But there is a better way, and we're about to discuss that with Cisco UCS X-Series. So let's take a look. What if we had infrastructure that had a single API, that was programmable through the Internet, that could call upon infrastructure as code deployments, that could run popular applications, traditional, all the way to scale-out and cloud native in between? Of course, it needs to run today's multiple OS and hypervisors and enabled us to collapse the fabric so that we don't have separate fiber channel with all of the optics in between and expensive cabling routes and Ethernet or separate Ethernet and a separate management collapsing to a unified fabric. What if the actual infrastructure that housed the servers themselves enabled for a larger server footprint? So that server could take on a persona as a blade server or a rack server. Well, introducing Cisco UCS X-Series modular system powered by Cisco Intersight. It is cloud operated, deployed, operated and managed by Cisco Intersight. It's got incredible flexibility, which I'm going to take you through in the next few slides, enabling it to take on the persona of an I/O-intensive high core or a compute node suitable for a VDI workload or a compute node suitable for an AIML workload. And if you were to invest in something like blade infrastructure, would you like that investment to last 10 years so that you're not having to take on another API set to rip and replace infrastructure out of your data center that you've sunk very good dollars into in order to try and achieve the sort of longevity of 5 years? What about 10 years? What if it was built for 10 years? UCS X-Series is built for 10 years. Let's have a look slightly under the hood. So if you can bring your eyes now over to our modular nodes on the left-hand side of the screen, we're not calling these blade servers anymore. These are modular nodes, and they come today in the form of a compute node. And in the near future, they will come as storage nodes, accelerator nodes, persistent memory nodes. The I/O is up to 100 gigabits per second to each server, have 2 network interface cards on the compute node and double it to 200. I/O is not constrained within X-Series. UCS X-Fabric. That sounds new. It is. It's brand new. X-Fabric is going to allow us to share resources between modules. I'll talk through that on the next slide. Power and cooling. How do we cater for the CPUs and GPU requirements of the future? Well, it's built into the design of this infrastructure. The ability to do liquid cooling, CXL I/O, the ability to power the CPUs and GPUs of today and of the next 10 years have been engineered into this infrastructure. So let's now better understand how does UCS X-Fabric technology really work. Well, we want the ability to run modern applications and assembling that using modules within the system. We want to deploy them via software policy that is driven out of Intersight. So let's just say that I add in 4 compute nodes into this single UCS X-Series enclosure. By the way, we could have 20 enclosures into a single UCS domain. This example, just to keep it simple, is one enclosure. And then I add in GPU accelerators. I add in a complete storage module full of up to 30 terabytes of NVMe. I add in external persistent memory that I want to share as resources to those compute nodes, and this is where I add X-Fabric. So each node has got an interface into the X-Fabric, which will allow the sharing of these resources on the compute node to the accelerator storage and memory node. Here's a use case. So I've got a single Cisco UCS X-Series system, and I want to run AIML. We've talked about the earlier example of AIML-dedicated system, purpose-built, not shared and not utilized. Maybe there's some downtime for that AIML system. Maybe you do most AIML calculations in the after hours, from 5 until 9 rather than 9 until 5. Well, this particular AIML shared modular approach, we've got compute node 1 sharing GPUs and calling upon storage from the storage node. In the same enclosure, we have compute node 3 with CPU, memory and I/O calling upon storage resources within the enclosure. And we've got big data drawing lots more storage, and the appropriate memory and I/O as compute node 4. Now the mind boggles as to what the combinations thereof can be here. Maybe during business hours, all 4 compute codes are VDI. Maybe from 5 until midnight, they are big data processing. Maybe from midnight until 8 a.m., they're AIML because these compute nodes are programmable. They're policy-based. So we can draw upon the resources within the X-Series modular system to see your business need based on the time and day to increase the efficiency and utilization of this investment. Here is an overview of the compute node. It's larger, and it can cater for 2 Intel third-generation Xeon CPUs that will deliver 1.5 to 3x the performance of yesterday's technologies. So talking about greater density, absolutely, up to 40 cores per socket, up to 32 DIMMs per server, connectivity for X-Fabric when it comes later this year, huge amounts of I/O through the Cisco virtual interface network card and up to 30 terabytes of NVMe in the front-facing drives. So this is the X210c M6 Compute Node available today. Now there's no traces within this enclosure. This enclosure is completely sheet metal. All of the systems in the front being nodes and fabrics in the back are interchangeable. You may not require X-Fabric day 1, but you may choose to add it day 2. Superior technology will come over the next decade, just like we've seen with the evolution of all of the systems of today. And X-Series will be capable of catering for those technologies like silicon photonomics (sic) [ photonics ], like liquid cooling, like next-generation Intel or AMD road map as they come to bear. You will not need to replace this enclosure or the upstream switching in terms of being able to adopt these sorts of technologies. That's pretty cool. That's protecting your investment. Now in terms of use cases, we feel that X-Series is ideal for virtualization, particularly because it's got dramatically increased bandwidth. And with the core densities available in CPUs today, that leads to a dramatically higher VM density not constrained by bandwidth. Bringing local NVMe drives directly via PCIe bus to the CPU will deliver much higher application performance. And if you're a customer that likes density, well, 8 1RU rack servers is 8 RUs, whereas 8 compute nodes in the X-Series is 7 RU. So talk about density, more VMs in lower footprint. These are the sorts of reasons why we believe X-Series is ideal for virtualization, but I'm sure you've got database deployments in your environment. So we're testing and qualifying SAP. Because of SAP's high amount of memory consumption, you're not constrained with any CPU with up to 128 gig, memory DIMMs of 32 thereof. Also, the ability to have very large CPUs in terms of core density or core frequency, usually frequency driving much more power consumption, you're not constrained with power consumption in UCS X-Series. Oracle databases. Typically, you're looking for lower-core, ultrahigh frequency. Available, cooled, powered within X-Series. And of course, Microsoft SQL server, similar requirements, lots of local storage, lots of flexibility in terms of CPUs and local disk configurations that are available to you with UCS X-Series. Now in the event that local storage consumption is not right for your use case, then Cisco has been leading converged infrastructure solutions since 2010 with the introduction of VxBlock and shortly thereafter, the introduction of FlexPod and more recently, the introduction of FlashStack. These 3 offerings are confirmed to be available via validated designs with UCS X-Series. So FlexPod from NetApp, FlashStack from Pure, VxBlock from Dell EMC will all be tested and validated and released via validated design, Cisco-validated design with UCS X-Series. In fact, FlexPod and FlashStack are available today. And VxBlock, well, will be available later this year. It's been a very, very successful venture for our storage alliance partners and Cisco with over 15,000 customers still running converged infrastructure solutions using UCS B-Series technologies today, multiple different workloads, multiple different use cases. And if you're not familiar, a Cisco-validated design is an engineering blueprint on how you can deploy your workload on this tried-and-tested single-vendor support infrastructure. So in summary, UCS-X Series modular system is ready now and very much designed for the future. It's deployed through the cloud through Cisco Intersight, managed and operated with the ability to add modular SaaS offerings to improve the efficiency of your cloud operating model environment. It is the most flexible infrastructure available in the market today, being capable of being policy programmed and adapted from being what we know as traditional blade infrastructure today to more flexible modular rack-like infrastructure also today. And it's future-ready. Cisco UCS B-Series was released in 2009. It's the same chassis that's still available today and powering the UCS B200 M6 using third-generation Intel Xeon CPUs, the same CPUs that are now available in UCS X-Series. So Cisco has a proven track record in delivering over a decade of longevity of the UCS B-Series, and we know we're delivering the same with UCS X-Series. UCS is proven to deliver fewer components. We know that we can power with X-Series double what's coming in the future. And now with Cisco Intersight, whether it's via a SaaS connection or a virtual appliance on-prem within your data center, we have a single control point in order for you to deploy, manage and operate UCS X-Series. So this brings us to our last slide. Let the UCS X-Series journey begin. Please access the assets that are available on the screen for you to discover more about the suitability of UCS X-Series for your business. Go to cisco.com/goucsx. Demo Intersight for yourself, cisco.com/go/intersight. We've got lots of great webinars. They go into more detail on the Intersight modular SaaS offerings like workload optimization, like HashiCorp Terraform for business cloud, like our Kubernetes servers, or please e-mail me directly. I'd love to connect with you, hear about your experience with your existing UCS or HyperFlex or, of course, start a new discussion in regards to UCS X-Series. Thanks for joining me. Wishing you a great rest of your day and look forward to connecting with you soon. Bye for now.