VMWARE Professionals - Intro and Scale

Editor's Notes

• #9: Before Hyper-V Server 2012Hyper-V was first introduced as part of Windows Server 2008, but later that year, in October 2008, a standalone, free edition of Hyper-V, known as Microsoft Hyper-V Server 2008 was released. This standalone version contained only the core virtualization components and Windows driver model, and as a result, offered a reduced disk footprint, and a reduced number of services running on the host OS.In October 2009, just 1 year later, both Windows Server 2008 and Microsoft Hyper-V Server 2008 were updated with the R2 release, which introduced a number of compelling capabilities to help organizations reduce costs, whilst increasing agility and flexibility. Key features introduced included:Live Migration – Enabling the movement of virtual machines (VMs) with no interruption or downtimeCluster Shared Volumes – Highly scalable and flexible use of shared storage (SAN) for VMsProcessor Compatibility – Increase the Flexibility for Live Migration across hosts with differing CPU architecturesHot Add Storage – Flexibly add or remove storage to and from VMsImproved Virtual Networking Performance – Support for Jumbo Frames and Virtual Machine Queue (VMq)Finally, in February 2011, With the addition of Service Pack 1 (SP1) for Hyper-V, Microsoft introduced 2 new, key capabilities to help organizations realize even greater value from the platform:Dynamic Memory – More efficient use of memory while maintaining consistent workload performance and scalability.RemoteFX – Provides the richest virtualized Windows 7 experience for Virtual Desktop Infrastructure (VDI) deployments.And that brings us to today…
• #11: Virtualization technologies help customers lower costs and deliver greater agility and economies of scale. Either as a stand-alone product or an integrated part of Windows Server, Hyper‑V is the leading virtualization platform for today and the transformational opportunity with cloud computing.With Hyper‑V, it is now easier than ever for organizations to take advantage of the cost savings of virtualization, and make the optimum use of server hardware investments by consolidating multiple server roles as separate virtual machines that are running on a single physical machine. Customers can use Hyper‑V to efficiently run multiple operating systems, Windows, Linux, and others, in parallel, on a single server. Windows Server 2012 and Hyper-V Server 2012 extend this with more features, greater scalability and further inbuilt reliability mechanisms.In the data center, on the desktop, and now in the cloud, the Microsoft virtualization platform, which is led by Hyper‑V and management tools, simply makes more sense and offers better value for money when compared to the competition.This presentation will focus on comparing Windows Server 2012 Hyper-V &Hyper-V Server 2012, with the VMware vSphere Hypervisor and vSphere 5.1 Enterprise Plus, across 4 key areas:Scalability, Performance & DensitySecurity & MultitenancyFlexible InfrastructureHigh Availability & Resiliency
• #12: In the previous releases of both Windows Server 2008 R2 Hyper-V, and the Microsoft Hyper-V Server 2008 R2, we supported configuring virtual machines with a maximum of four virtual processors and up to 64 GB of memory. However, IT organizations increasingly want to use virtualization when they deploy mission‑critical, tier-1 business applications. Large, demanding workloads such as online transaction processing (OLTP) databases and online transaction analysis (OLTA) solutions typically run on systems with 16 or more processors and demand large amounts of memory. For this class of workloads, more virtual processors and larger amounts of virtual machine memory are a core requirement.Hyper‑V in Windows Server 2012, and Hyper-V Server 2012 greatly expands support for host processors and memory. New features include support for up to 64 virtual processors and 1TB of memory for Hyper‑V guests, a new VHDX virtual hard disk format with larger disk capacity of up to 64 TB, and additional resiliency. These features help ensure that the virtualization infrastructure can support the configuration of large, high-performance virtual machines to support workloads that might need to scale up significantly.Significant improvements have been made across the board, with Hyper-V now supporting increased cluster sizes, up from 16 nodes, to 64 nodes in size, supporting up to 8,000 VMs per cluster. Hyper-V also now supports a significantly higher number of active virtual machines per host, and additionally, more advanced performance features such as in-guest Non-Uniform Memory Access (NUMA). This ensures customers can achieve the highest levels of scalability, performance and density for their mission-critical workloads.
• #13: In terms of comparison, the table shows that Hyper-V offers significantly greater scale across Host, VM and Cluster when compared with VMware vSphere Hypervisor. VMware positions vSphere Hypervisor as a simple, entry-level solution designed to allow users to experience the benefits of VMware’s virtualization platform at no cost, however on closer examination, certain restrictions are imposed which prevent customers utilizing the solution at scale, meaning customers have to purchase, at significant cost, one of the more advanced vSphere editions. Examples of this include the capping of physical memory on the vSphere Hypervisor to 32GB, limiting scalability, and subsequently maximum virtual machine memory size as a result.Since the launch of vSphere 5.0, back in 2011, VMware has regularly discussed the inclusion of 32 virtual processors within a virtual machine, yet this was exclusive, at the time, to the Enterprise Plus edition of vSphere, and not vSphere Hypervisor, vSphere 5.0 Essentials, Essentials Plus, Standard, and Enterprise editions, which were all capped at 8 virtual processors per virtual machine. With vSphere 5.1 however, the Enterprise edition can deliver VMs with up to 32 vCPUs, and the Enterprise Plus edition 64 vCPUs. Compare this with Hyper-V in both Windows Server 2012 and Hyper-V Server 2012, and customers not only receive up to 64 virtual processors per virtual machine, but this comes with no SKU-specific restrictions. Customers are free to run the most demanding of their workloads on Hyper-V, without additional costs or expensive edition upgrades.The table also shows that both Windows Server 2012 Hyper-V, along with Hyper-V Server 2012, and vSphere 5.1 Enterprise Plus deliver up to 1TB of memory to an individual virtual machineWhilst virtualization itself is an incredibly important aspect within the datacenter, resiliency and high availability of workloads is of equal importance. The inclusion of Failover Clustering with Windows Server 2012 and Hyper-V Server 2012 enables customers to achieve massive scale with an unparalleled number of nodes within a cluster, and virtual machines per cluster. Unfortunately, the free vSphere Hypervisor alone doesn’t provide any high availability, or resiliency features, and customers must purchase vSphere 5.1 to unlock these features, and even then, cluster sizes are restricted to only 32 nodes, and 4,000 virtual machines per cluster, which is considerably smaller than the 64 nodes, and 8000 VMs supported by Windows Server 2012 Hyper-V and Hyper-V Server 2012.
• #14: Windows Server 2012 Hyper‑V supports NUMA in a virtual machine. What is NUMA?NUMA, or Non-Uniform Memory Access, refers to a computer architecture in multiprocessor systems in which the time required for a processor to access memory depends on the memory’s location relative to the processor. With NUMA, a processor can access local memory (memory attached directly to the processor) faster than it can access remote memory (memory that is local to another processor in the system). Modern operating systems and high-performance applications such as SQL Server have developed optimizations to recognize the system’s NUMA topology and consider NUMA when they schedule threads or allocate memory to increase performance.Guest NUMAProjecting a virtual NUMA topology onto a virtual machine provides optimal performance and workload scalability in large virtual machine configurations. It does this by allowing the guest operating system and applications such as SQL Server to take advantage of their inherent NUMA performance optimizations (for example, making intelligent NUMA decisions about thread and memory allocation). The default virtual NUMA topology projected into a virtual machine running Hyper‑V is optimized to match the host’s NUMA topology, as shown in the figure.