Led by Ken Surplice Category Manager, Mission Critical Solutions, Matthias covers the combination of ESX, Superdome Flex and where support will arrive in the coming months
Ken: “Matthias, thank you for talking with me today. In your role as Senior Systems Engineer with VMware’s Technical Alliance Manager EMEA, how do you engage with customers?”
Matthias: “Thank you for your introduction and your question, Ken. In my responsibility, I care for our OEM partners of VMware and HPE is our most important partner based on HPE’s contribution to our business.”
Ken: “I remember VMware’s Project Capstone. Your company said that with today’s scale-up x86 servers and high performance all-flash storage even the most demanding workloads could be virtualised with VMware. How is it going?”
Matthias: “With Capstone, we have demonstrated already in 2015 the capability of vSphere to deliver more performance, if the number of hosted VM’s is increased. Since mission-critical workloads do require very large systems, the VM’s in the test utilized 256 GB of memory as well as 30, 60 and even 120 virtual processors. We could increase the number of orders per minute by a factor 3.7 in adding three additional VM’s with 120 vCPU’s– an almost linear grow of performance.
It is amazing to observe, how HPE SuperDome X delivers such scalability, which required the use of expensive RISC processors in the past. We used HPE SuperDome X supporting 240 cores and 12 TB of RAM as the physical platform for our Capstone project. HPE SuperDome X heads up the x86 server market in scalability, performance and – another important criteria for critical workloads – availability features.”
Ken: “We met at HPE’s Discover symposium in November where HPE introduced its new modular large memory scale-up x86 server, Superdome Flex. You seemed very interested. What do you think the combination of VMware and Superdome Flex can do for customers?”
Matthias: “Customers using scale-up x86 server architectures are always looking for even larger server platform. They see positive economic results in reducing the number of servers to safe maintenance costs and downtime.
As you kindly presented me HPE SuperDome Flex in Madrid, I got impressed about the new modular architecture. If I recall correctly, this just launched new platform will deliver more than twice the performance compared with HPE SuperDome X.
HPE SuperDome Flex will convince new customers to migrate their critical and perhaps already virtualized workloads to this new flagship on x86 based computing platforms.”
Ken: “What advice would you have for customers when choosing between a scale-out and a scale-up environment for VMware?”
Matthias: “Even with recent approaches, developing programs for scale-out environments, the majority of critical workloads are still using legacy applications, which require a scale-up environment. Take SAP HANA as a prominent example – as more processors and memory you add to a single server or a single virtual machine as more performance you gain.
Let me list a number of advantages of a scale-up platform for virtualized workloads. In some cases, the application requires a (very) large virtual machine and this configuration defines the minimum number of physical cores and the amount of physical memory the platform needs to support.
Additionally, typical multi-tier applications can communicate between each other much faster, if they are hosted on the same physical platform. In this case, the network traffic between two virtual machines can stay inside the server, which reduces latency caused by components like physical network controllers or physical switches.
As you can see, for me a scale-up solution has clear benefits compared with scale-out scenarios. Besides improved performance, customers will also see a reduction of operational costs and downtime with such a system.”