IT Infrastructure From All Angles
Steve Sibley, Eric Herzog and Bruce Anthony discuss how IBM infrastructure, storage and hybrid cloud solutions offer cohesive solutions to meet your business needs.
By Neil Tardy11/30/2020
As a year with unprecedented challenges and opportunities draws to a close, IT leaders are searching for solutions that will move their organizations forward. Together, IBM Power Systems™ servers, IBM Storage and hybrid cloud solutions—on the IBM Cloud® or elsewhere—help clients meet those challenges with cohesive solutions.
With this in mind, we talked to three IBMers who are prominently involved in IBM Power Systems management and development, to get their thoughts on what we’ll see and what we can accomplish in the year ahead. In three separate interviews, you’ll get thoughts from Steve Sibley, vice president, IBM Power Systems Offering Management; Eric Herzog, CMO and vice president of Global Storage Channels for IBM Storage; and Bruce Anthony, Distinguished Engineer and CTO with Cognitive Cloud Systems.
In these interviews, you’ll read broadly about how IBM technological advancements are bringing a cohesiveness to IT environments. More than ever, servers, storage and cloud must function seamlessly so that businesses can innovate quickly and efficiently. You’ll find plenty about the ongoing evolution of cloud computing and how clients benefit from the IBM/Red Hat® partnership. You’ll also get some insight on IBM POWER10, which is expected to arrive late next year.
We hope you’ll find these conversations enlightening and informative. Here’s to meeting your 2021 business goals from a solid IT infrastructure foundation.
The Road to IBM POWER10
Steve Sibley explains the benefits of the Power Systems software stack and discusses the anticipation for IBM POWER10
“We’re building more automation and capability so our clients, as they continue their shift to cloud, can manage their remote systems more easily than ever.”
Q: Servers, storage and cloud are all pieces of a larger infrastructure puzzle. Can you talk more about that?
Steve Sibley (SS): Clearly, IBM’s ability to bring together storage and compute delivers unique value and performance to our clients. Our storage team has done an amazing job around flash technology, which when combined with IBM POWER®, enables greater throughput in capacity. Leveraging our storage team’s innovation, we’ve been the leader on I/O throughput and the capability to access and drive transactions for the past couple of generations. In terms of performance, providing greater resiliency in storage gives our clients the best overall solution on prem. The IBM Cloud brings IBM Power Systems servers and IBM Storage together into a similar architecture that provides consistency for our clients.
Q: At this pivotal moment in business, what makes infrastructure in particular so important?
SS: There was already a clear demand that clients innovate more quickly. If they don’t, their competitors will. Now clients must be agile even as they enable work from home and other rapid business adjustments.
Certainly, some of our clients are back in their data centers, but most continue to enable remote work. It’s been gratifying to hear many clients tell us that their Power® infrastructures have responded to increased demand on the supply chain or shifting use in workloads. Clients cannot compromise on their need for business continuity and scale, so the capability for these infrastructures to respond quickly and seamlessly with business continuity is more important than ever.
Q: How does the partnership with Red Hat impact clients’ on-premise infrastructure?
SS: We’re building more automation and capability so as our clients continue their shift to cloud, they can manage their remote systems more easily than ever. Clients still need the most secure environment possible, so we’re addressing security not only at the chip level, but all the way up the stack. Our relationship with Red Hat and the capabilities they bring is a key part of our increased focus on automation.
Red Hat OpenShift® gets a lot of publicity, but the Red Hat Ansible® project delivers automation capabilities for clients both on premises and in the cloud and help simplify their operations. It enables a consistency of skills across server architectures, allowing our clients to automate most of the routine tasks that confront them. This lowers administration costs as well as reduces errors.
Ansible has become an important vehicle for us to drive automation. So is OpenShift. As we talk about innovating quickly, the new focus from an application development standpoint is clearly around containers and microservices with Kubernetes. OpenShift providing that management infrastructure allows our clients to bring that new application development capability for new services right in step with their core business applications and data that are running on Power today.
Q: There’s been considerable discussion of public clouds and their evolution, but IBM recently came out with a significant private cloud offering. Let’s talk more about that.
SS: In May, we introduced IBM Power Private Cloud, which is a new way to purchase capacity for Power Systems environments. This provides tremendous flexibility because clients can dynamically add new capacity when they need it, on essentially a pay as you go basis.
Power Private Cloud is unique in its ability to optimize cost. We monitor entire infrastructures across multiple data centers and only charge clients for what they use in total aggregate across their entire pool or landscape of resources. By moving to Power Private Cloud, clients can reduce their costs 30-40% without any challenges or impact to their performance, which translates to huge interest and excitement in the offering.
Q: The IBM POWER10 chip was announced at Hot Chips earlier this year. What can clients look forward to?
SS: We always introduce next-generation technology at Hot Chips. In this case it happens to be about a year ahead of when our systems will likely launch, but with the interest and excitement around the capabilities that our development teams have built into the IBM POWER10, we’ve obviously shifted our foundry.
We’re using Samsung 7 nanometer technology, a great capability that drives more energy into chips and gives us a greater chance to deliver more performance and more scale to our clients. We’ll pack up to three times more performance in a socket than we have with IBM POWER9™. IBM POWER9 is already leading the industry in many different performance metrics, so when you talk about an IBM POWER10 chip that delivers that much more overall capacity and performance, people are getting really excited. IBM POWER10 also has a key focus on security: the capabilities to do end to end encryption in memory and encrypt containers. Memory inception will transform how our clients put their systems together. So there’s a lot in IBM POWER10 that’s exciting.
Q: Security and reliability have come up a bit in your answers. Can you talk about how they’re baked into the IBM Power Systems platform?
SS: Most clients run most of their critical data or most critical transaction processing applications on Power Systems technology. That says two things about their priorities. One priority is availability. Their systems must be available all the time within their environments: 24-7, 365 days a year. That’s why, from an engineering standpoint, IBM is so focused on our systems’ availability. Of course, along with the hardware, we provide the system software on top of that, so if there ever is an availability issue, solutions like IBM PowerHA® can provide very fast failover.
The other priority is security. Security is designed and built into the processor itself, and our software, starting with IBM PowerSC®, automates security and compliance in a way that simplifies things for our clients. We are focused on the full stack. A few years ago, when Spectre and Meltdown emerged as key security issues, this focus enabled us to respond to these urgent circumstances. From the virtualization layer up through the firmware, the security closures were there. Because IBM owns the chip, the virtualization and the software on top, we have the ability to move more quickly and provide complete solutions.
Storage for Future-Forward Workloads
Eric Herzog explains how IBM Storage offerings enable clients to operate in a cloud-connected, multivendor world
“According to IDC, large enterprises use five to seven different clouds. So tiering data from our Spectrum Protect to the world’s four largest clouds gives our clients several strong choices for their hybrid cloud configurations.”
Q: Let’s start broadly: How do storage, Power, and hybrid cloud all work together to help businesses meet today’s challenges?
Eric Herzog (EH): All companies—big, medium or small—are doing hybrid clouds. Certain data sets stay on-premise and some go off-premise, so what’s critical is that there’s a way to move data seamlessly back and forth. Powerful servers and storage are needed on the on-premise side. You also need the right kind of software. For example, IBM Storage has a solution called Spectrum® Virtualize for public cloud that sits in IBM Cloud or Amazon Web Services (AWS). Next year we’ll also be in Azure. That’s what’s needed in a hybrid cloud environment: The right infrastructure, on and off prem, and, then, the capability to seamlessly move data automatically.
Q: In addition to Azure, what else can clients look forward to in the coming year?
EH: We see that data access varying across multiple server infrastructure configurations: containerized, virtualized, and “bare metal” (applications deployed directly on servers). The larger the company, the more likely you need all three models: some applications sitting in a virtualization engine, some applications in a container environment and some applications in bare metal. IBM Storage supports all of those variants of your infrastructures.
Q: Can you talk about the Red Hat partnership and what it means for clients, particularly from a storage perspective?
EH: IBM Storage has worked with Red Hat for more than 10 years. From a storage perspective, we’ve always had a tight integration with Red Hat. We support everything in the Red Hat OpenShift environment. Likewise, Red Hat works with our environments: Spectrum Scale, IBM Cloud storage and Spectrum Virtualize software, as well as the associated storage arrays that leverage that software.
Over the past two years, we’ve integrated across our portfolio: Modern data protection to keep the data safe, primary storage for high transactional workloads such as our flash systems or DS8900, applications that are focused on big data, AI and analytics workloads with Spectrum Scale, and IBM Cloud object storage. In addition to working in your data center in virtualized or bare metal environments, all of IBM Storage’s solutions are very tightly tied to Red Hat container configurations.
Q: How do storage needs change with innovative environments like AI and big data?
EH: First, you need massive amounts of storage. Some of our Spectrum Scale and IBM Cloud object storage clients have more than an exabyte of storage in production (which is giant). Those environments also need to heavily scale. They require the capability to seamlessly move data back and forth across a cloud environment, which we provide. From a performance perspective, those applications—AI, analytics and big data—must be heavily tied to the bandwidth of the storage. So the bigger the bandwidth, the better the performance for your AI, big data or analytics workloads.
Q: What are some other ways that IBM Storage helps clients cope with large quantities of data?
EH: Data scientists and AI analysts, they’re not storage experts. Searching exabytes of data is a big chore and with our Spectrum Discover solution, we can make that exceedingly easy. We’ve created APIs so AI, analytics and big data software can seamlessly talk to our storage. It’s about having high availability, high performance and high integration across clouds, while integrating with third-party software that is important for our end users.
With our flash system products, we can seamlessly replicate data back to IBM or AWS cloud environments, helping you avoid malware and ransomware. And with IBM Spectrum Virtualize, you can move data seamlessly between on-prem and cloud storage.
You want to automate; you want to tier data. For example, Spectrum Protect—our data protection cyber resiliency software—keeps data safe. It facilitates backups to IBM Cloud, AWS, Azure and, as of this October, Google Cloud. Large global enterprises don’t use one cloud provider, they use many. According to IDC, large enterprises use five to seven different clouds. So tiering data from our Spectrum Protect to the world’s four largest clouds gives our clients several strong choices for their hybrid cloud configurations. That’s very important. It’s even more important given the events of 2020.
Q: A final question: Earlier this year SAP announced an extension of support for Business Suite 7, essentially adding two years to mandate to migrate to SAP HANA. Can you talk about how storage fits into the equation as clients try to decide if, when, and how they’re going to migrate to HANA?
EH: We’re certified for traditional SAP and also for SAP HANA. In fact, we have more certifications for SAP HANA than any other storage vendor. IBM also has substantial documentation, including Redbooks, that address SAP HANA configuration. Of course, we lead with Power and SAP HANA, but whatever clients choose from a server perspective, IBM Storage fully supports.
Your Toolbox for the Cloud
Bruce Anthony discusses the inner workings of cloud environments and explains how IBM is aligning the pieces to provide clients with powerful systems and flexible tools
“I expect to see rapid growth in the development and deployment of container-based workloads.”
Q: Steve Sibley and Eric Herzog are business executives, but you have a different role. Could you discuss what it is you do from day to day?
Bruce Anthony (BA): My role is to be the thought leader, strategist and architect for delivering on IBM’s hybrid multicloud strategy with Red Hat OpenShift on Power. I partner closely with a development executive to lead the team. I tend to focus on the technology big picture and how everything fits together and my partner focuses on engineering execution and product delivery. You could say I’m the visionary laying the tracks while my partner runs the engineering trains to deliver the products.
Q: That’s a great analogy. Before diving deeper, could you discuss the benefits of running cloud environments on IBM Power Systems infrastructure as opposed to other platforms?
BA: Running our cloud software stack, Red Hat OpenShift, leverages capabilities in the platform that have traditionally delivered great value for clients. Our performance gives us significantly more compute capacity per core. We have a more secure and robust infrastructure that’s less vulnerable to attacks. We manage workload consumption on our system so that clients receive more effective utilization; there’s more bang for your buck on the hardware.
Being able to deliver Power capabilities in an OpenShift cloud setting provides great flexibility. Clients can rapidly create and deploy new software and easily connect the worlds of on-premise and cloud-based computing. We have a great degree of flexibility in how you can assemble the pieces of your solution and where you can operate them. I like to say you develop it once, deploy it anywhere and operate it everywhere.
Q: What about the benefits of cloud native for Power Systems clients?
BA: Many of our clients are entering an increasingly competitive world where they must improve their software and solutions to better serve their clients, partners and employees. They’re looking to gain new insights into their data through analytics or AI. Red Hat OpenShift is our cloud native computing stack for Power and IBM has a large investment in software applications that run on OpenShift called Cloud Pak® solutions. A popular one is the IBM Cloud Pak for Data, which runs on Power providing an end-to-end data, analytics and AI solution. For example, a very large bank in Europe is using this installation to derive new business insights using analytics tools in the IBM Cloud Pak for data running on Power. This gives their data scientists an end to end workflow to connect the data to learning systems, and then implement the insight that comes from it.
In addition, with Red Hat OpenShift on Power Systems, clients can develop software more rapidly. Whether they have AIX®, IBM i or enterprise Linux, clients can create new applications in Red Hat OpenShift that run colocated on the same Power system. This allows them to securely connect their existing data with new applications, creating new value and functionality. OpenShift on Power provides secure, robust price performance for clients to run their core business. We’ve seen examples of clients getting up to 50% more throughput running containers on an IBM Power Systems server as they would on comparable competitor’s system. Clients gain a significant amount of capacity while having fewer physical boxes on site that consume less power and occupy less physical space.
Q: What are some specific ways that the Red Hat partnership benefits clients?
BA: Our entire team works day to day with Red Hat, creating differentiating capabilities that we can deliver to IBM Power Systems clients. We’re also partnering with Red Hat in specific industries. For example, a number of banking clients have purchased OpenShift on Power. A bank in the mid-Pacific region wanted to get greater insight into their data while combining it with insights from social media and other public cloud services. The current process they were using was time-consuming and had concerns about potential security exposures. They constructed a more secure solution on Power that also improved response times by 50%. We’re creating and delivering compelling new application environments for banks by marrying the agility and broad software ecosystem from Red Hat OpenShift with the strength that Power brings in security, utilization and performance.
Q: What are some other ways that OpenShift and Cloud Pak solutions help complete the IT infrastructure puzzle?
BA: When you think about the overall solution component, there’s obviously a server—a compute part. There’s storage to keep the data, there’s the OS and container management layer. Then there’s middleware. Across IBM, we’ve taken the middleware we’ve been working on for years and converted it into a container-based infrastructure, grouped around various use cases. For instance, there’s a Cloud Pak for application developers to help them create new container applications; there’s a Cloud Pak for data, which is all about managing and gaining insight from your data; there’s one for integration, and that’s about connecting your data and business processes. All of these give you the building blocks to assemble solutions for your particular business. Some clients take those and create their own custom applications, while others go to system integrators and ISVs to create a specific, tailored solution.
Q: Do you anticipate any other ways that containers will change the IT landscape?
BA: I expect to see rapid growth in the development and deployment of container-based workloads. These workloads will need to scale across a range from hundreds to millions of client interactions at the same time. Having systems that dynamically scale to those needs will be critical. Our IBM POWER10 systems will be well-aligned for containers. They’ll be capable of handling those fluctuating workloads and they’ll support more containers on a system than nearly anything other system in the industry.
Also, security is particularly important for containers. We’re working to make each individual container more secure and less vulnerable to attacks not just from the outside but internally from bad actors in your own business, so that you can provide a secure environment for executing containers for your most privileged data and information.
Q: Let’s talk more about IBM POWER10. What will it mean for cloud workloads in particular?
BA: Processor threads are the basic computing unit for Containers. A large IBM POWER10 system will have over 1,900 threads, far more than competing systems, allowing us to support solutions with many more containers on a single system. With everything on a single system, containers can communicate more quickly and securely with far greater performance and reliability. Looking forward—and this is a couple of years down the road—IBM POWER10 and its memory inception capabilities offer some really exciting possibilities for containers. We could see whole new ways of architecting solutions built around memory inception architecture with very large pools of memory that are always active with tens of thousands of processing cores operating on the data in that memory. I’m excited to see what will flow from that in terms of fundamentally new styles of applications providing the scale, resiliency and the incredible response time users need.
Neil Tardy is a contributing writer to TechChannel.