VMware Infrastructure

VMware Infrastructure
Jeff Patton and Team Network Administrator Reinsel Kuntz Lesher LLP 1

VMware Infrastructure Adoption Trends
Agenda VMware Infrastructure Adoption Trends There are solid reasons why VMware is the most deployed virtualization solution on earth.

VMware virtualization
First Things First: Virtualization VMware decouples software from hardware. Enterprise App Operating System Operating System VMware virtualization Virtualization: fundamentally better IT environment a “truly transformational change in computing” (DG in IPO presos) Run VM on any system without modification Entire systems can be saved, moved or copied just like a file Run multiple VMs on a single system BTW: we introduced this back in 1999

Virtualization: Fundamentally Better
Run several operating systems on a single machine. Create shared pools of resources to optimize your infrastructure.

The Virtual Infrastructure Stack Today
Infrastructure Optimization Business Continuity Desktop Management Software Lifecycle  Management & Automation Resource Mgt Availability Mobility Security Virtual Infrastructure In Summary – a wide range of products and features from VMware enable further rapid adoption across the board. Extending out the virtualization capabitilie of VMware virtual infrastructure helps make it the right choice for virtualization in datacenters worldwide. ESX Server 3i 3.5 ESX Server 3.5 Virtualization Platform Œ 5

Cross-platform support
ESX Server Virtualizes servers, storage and network Runs multiple virtual machines on the same server Virtual Machines Performance Scalability Stability Cross-platform support ESX Server Only third generation hypervisor Production proven in more than 40,000 customer deployments Unparalleled reliability: more than 1,000 says of uptime and counting Physical Server VMware® ESX Server abstracts processor, memory, storage and networking resources into multiple virtual machines. Review: each VM looks, feels, and acts like a real server <CLICK> ESX Server delivers the highest levels of performance, scalability and robustness required for enterprise IT environments. Its bare metal architecture provides full control over the hardware resources, and near native performance Scalability: ESX Server allow to create powerful virtual machines that can support more than 95% of all x86 workloads. VMs can include up to 4 virtual CPUs and 16 GB of virtual memory. Stability: >3 years of uptime at a customer, 86% of customers use in production, Very importantly, ESX Server is rigorously tested and certified to work with a wide variety of hardware systems and storage. It also provides the widest guest OS support including windows (multi-generations), Linux, Solaris x86, Novell

Virtualization Platform Enhancements in ESX 3.5
Storage Network Virtual Machines ESX Server CPU Memory Performance Scale Compatibility Paravirtualization 64GB virtual RAM Ubuntu Windows Vista Large memory pages 256 GB of physical RAM TCP Segmentation Offload (TSO) Jumbo Frames 10 GigE Infiniband First we cover Scale & Compatibility – then a deep dive into performance. With ESX server 3.5 ; VMs as big as 64GB of memory can run, ESX server can run on hardware shipping with up to 256 GB of physical RAM On the compatibility side, a growing ecosystem around VMware have partnered with us to create expanded storage and networking choices for ESX Server by contributing through the community source program. High speed networking interconnects (e.g. 10 Gig E ) are now enabled by ESX server 3.5. (Note: iSCSI over 10 GigE is currently experimental.) SATA local storage is also supported as are low cost storage options such as SAS arrays. NPIV Support SATA devices

Application Performance Benefits – an example
Achieve 60,000 iops with ESX Server 3.5 Average across 15,000 Oracle databases is 1280 iops As an example of workloads that typically concern people, consider Oracle. Note the number of i/o operations per second achievable with ESX Server 3.5 and compare to the average number of iops collected across Oracle databases. What this comparison shows that most databases that do not have peak i/o requirements that exceed 60,000 iops are good virtualization candidates. These results with ESX Server 3.5 are the result of a number of optimizations which we cover in the following slides (From Richard McDougall’s blog) Database designers go to significant effort to avoid doing disk I/O when possible. This is because the latency of a disk I/O is substantially higher than the time a transaction will spend on the CPU. For example, a disk I/O takes on the order of 10 milliseconds, while the typical transaction takes just a few milliseconds of CPU time. If an I/O to disk can be cached in memory, it then could be serviced in a fraction of a millisecond. In addition, since disks are expensive, the cost of the storage systems for databases is often more affected by the I/Os per second that it can deliver, than by pure storage space. Lowering overall disk throughput can mean significantly lowering the cost of the system. Larger memory sizes help Oracle by caching more disk blocks in memory. Consider this simple example: if a database system is using memory to cache it’s disk I/O, and is yielding a 90% cache hit rate, this means that one in every ten accesses causes a physical I/O. For 10,000 accesses a second, we would see 1,000 I/O’s per second. If we increase memory to improve the cache hit rate to 99%, then we reduce the I/Os to one in one-hundred, reducing the physical I/O by 10x to only 100 I/Os per second. Often, over 80% of the memory used by the guest operating system is used by the Oracle disk block cache. A general rule of thumb is that the database cache be sized at 5-10% of the database size, and that doubling memory improves throughput by about 20%. This is obviously very workload dependent, but you can see that larger memory sizes help improve resource efficiency in other areas of the system, and that generally, more is better. For these reasons, large memory is very important for databases. VMware ESX 3.0 allowed 16GB of RAM per guest, and 3.5 increases the capability to 64GB per guest. Average number is data collected from Capacity Planner which collects iops info throughout the day at regular sampling intervals 60,000 iops is equivalent to 60,000 Exchange users

Virtualization Platform Enhancements Paravirtualization
Storage Network Virtual Machines ESX Server Memory Paravirtualization is making the guest OS virtualization aware and more performant running on a hypervisor VMware led successful industry collaboration effort to create a standard interface - paravirt ops - that allows standard Linux kernel (2.6.21) to be VMware virtualization aware Ubuntu 7.04 now uses this interface Enables Linux to run at higher perform-ance by making it virtualization-aware Performance benefits seen for workloads such as large databases, multi-process apps The slide pretty much covers it. Emphasis – no customization required to Linux to make it virtualization aware. Generally all applications that run on Linux distros that have the new paravirt interface will benefit from paravirtualization – the specific workloads pointed out are where we see the significant benefits Web servers such as Apache are sometime multi-process

Paravirtualization Benefits
Paravirtualization benefits workloads when virtual machines are Switching between several different execution threads Executing network and disk i/o Large databases, multi-process applications, file servers, web servers typically benefit from these improvements The graphs on the left indicate the benefits achieved through paravirtualization The top one compares kernel microbenchmarks across a VM that’s using the regular BT (binary translation) and another that is using the VMI paravirtualization interface. What is a kernel microbenchmark? Real life workloads spend most of their time in user code (less privileged state) and little time in kernel code. We use Kernel micro-benchmarks so that the virtual machine spends almost all of its time in privileged state. The graphs in the charts show that significant benefits are achieved for a virtual machines when in privileged state. The bottom chart shows a comparison between a BT VM and a paravirtualized VM while executing network & disk I/o. Higher Mb/sec throughput is achieved with VMI VMs. More detail on benchmarks: The following kernel micro-benchmarks were used Js – javaSwitch: A Java program that switches between multiple java threads. dz – divZero: C program that generates divide by zero fault and handles it gp – getppid: C program that issues getppid system call ns – nativethread Switch: C program that creates threads and switches between them nf – nforkwait: C program that forks processes and waits for them to complete sv – segv: C program that generates segmentation fault and handles it As expected kernel micro-benchmarks run significantly faster in VMI VM. This is because VMI VM completely avoids the binary translation overhead involved in handling privileged instructions. While this result shows signficant performance gap with BT, this performance gap is highly dependent on the workload. Dbench & Tbench: Dbench and tbench simulate the disk and network I/O component of SMB I/O traffic respectively. The benchmark has a server and client component. For this experiment both the server and the client was run on the same machine. Dbench uses asynchronous I/O which does very little disk access due to filesystem buffering. Since there is no real network and disk I/O this benchmarks behave like a kernel micro-benchmark that exercises I/O calls without stressing the I/O devices. We run the benchmark with 4 and 8 client threads. VMI gains significantly over binary translation. It is important to note that the performance gap between BT and VMI decreases when there is more disk I/O as I/O latencies hides the CPU overhead. This is not shown in this graph, but is evident with workloads that are I/O intensive. Guest: SLES 10 SP1, VMI patched kernel VM: 1 vcpu, 1G

Virtualization Platform Enhancements Memory Management
Storage Network Virtual Machines ESX Server Memory Large memory pages for more efficient memory access Benefits memory intensive workloads such as databases, Java apps ESX Server translates the guest OS physical memory to actual hardware physical memory. This translation when done in the CPU hardware is faster than when done in software Larger memory pages for the guest OS and the hypervisor enable more efficient memory access The net effect of this is that memory intensive workloads such as databases, Citrix etc see benefits in performance. Here are a couple of references that explain the benefits of large memory pages… “4. 2MB PTE's for Oracle memory - We saw an 8% gain when we switched from using 4k PTE's to 2MB PTE's for the shared memory Oracle was allocating. 14GB of the 16GB of memory on the system was set aside for this "bigpage" usage. “

Virtualization Platform Enhancements Network
Storage Network Virtual Machines ESX Server CPU Memory TCP Segmentation Offload* reduces CPU overhead by streamlining TCP processing Jumbo frames* allow reduced TCP/IP processing by enabling larger frame sizes Support for 10 GigE for network intensive workloads** Networking performance ESX Server 3.5 introduces these enhancements that reduce the overhead associated with processing network i/o. o TCP Segment Offload (TSO) dramatically improves performance by reducing the CPU overhead involved with sending large amounts of TCP traffic. It is supported both in the guest operating system as well as in the ESX Server kernel TCP/IP stack. Thus it improves performance for TCP data coming from a virtual machine and for traffic, such as VMotion, that is sent out of the server. TSO is enabled for most virtual machines automatically. With NICs that support segmentation offload there is some additional benefit. To enable TSO in a virtual machine, you must select either the Enhanced vmxnet or e1000 virtual device. iSCSI with TSO is supported. TSO is only enabled with select guests. Guest list will be available with GA documentation. Jumbo Frames allow ESX Server 3.5 to send larger frames out onto the physical network. The network must support jumbo frames (end-to-end) for jumbo frames to be effective. Jumbo frames up to 9kB (9000 Bytes) are supported. NICs must support jumbo frames- there will be a separate HCL for this list. iSCSI with Jumbo Frames is not supported. This feature is available only for select guest operating systems and needs to be enabled via the command line. Benefit network intensive workloads (e.g. backups over LAN, webservers, Citrix, Win Terminal Services, iSCSI) in virtual machines * Supported for ESX Server hosts and select guests **10 GigE not supported for the ESX iSCSI software initiator

Networking Performance Benefits
TSO Enabled Throughput for Windows 2003 TSO & Jumbo frames enable faster throughput with lower CPU overhead I/O intensive workloads achieve faster throughput times Enables consolidation of greater number of workloads per server Throughput Mb/s Jumbo Frame Throughput These charts that show benefits obtained from networking performance optimizations in ESX Server 3.5. The top chart compares throughput of a Win2003 VM without TSO enabled to one with TSO – we see significant throughput improvements. The bottom chart shows that as Frame size increases, throughput goes up. Note that in the above experiments CPU overhead showed an actual reduction while throughput rates increased. This means that you can now have more better i/o throughput at higher consolidation ratios..

ESX Server 3i Compact, 32MB footprint
Only architecture with no reliance on a general purpose OS Integration in hardware eliminates installation Intuitive wizard driven start up experience dramatically reduces deployment time Standards-based management of the underlying hardware ESX Server 3i is the next generation, thin hypervisor integrated in server hardware Hypervisor: on its own, ESX Server 3i offers basic partitioning of server resources. Howerver, it also acts as the foundation for virtual infrastructure software, enabling VMotion, DRS, etc, the keys to the dynamic, automated data center Thin architecture: Small footprint (32 mb) for security, reliability and simplified management Server integration: Default feature makes deployment easy and fast Two target markets: Existing customers: Get all the performance and sophistication of ESX Server and fully supports VI3 features and benefits in a very small footprint architecture New customers: Easiest way to get started with virtualization Additional functionality on top of the hypervisor just requires the right licenses, not any changes to the code itself. No reinstalls and no VMFS changes to go from running a standalone instance of the hypervisor to a full VI3 Enterprise deployment Management is simplfiied because: No Linux command line skills required No user accounts or passwords to create and maintain No OS security hardening, antivirus, or backup effort required Security and reliability is increased because: Fewer interfaces minimize attack profile Locked-down, BIOS-like interface prevents users from running arbitrary code No dependence on failure-prone hard-drive enables disk-less servers More than 50% fewer patches* ON STANDALONE AVAILABILITY: A standalone , hard disk installable version will be available after the initial ESX Server 3i starts shipping embedded in servers. The date for this is still TBD. This standalone version may have a smaller compatibility list as compared to ESX Server 3 Note: HA support for ESX Server 3i is experimental initially. We expect we will shortly make it production level support. Server boot to running virtual machines in minutes Simplified management Increased security and reliability

From Server Boot to Running VMs in Minutes
 Power on server and boot into hypervisor Configure Admin Password (optional) Modify network configuration Connect VI Client to IP Address Or manage with VirtualCenter    We’ve also simplified the user experience for this new form factor – it becomes as easy to set up as a home router appliance. For e.g. It picks up IP addresses automatically using DHCP – or it can be manually set The whole setup time takes a few minutes, customers are ready to run VMs within minutes of booting into ESX Server 3i

Enabling the ‘Plug-and-Play’ Datacenter
Plug: Power on a new server with ESX Server 3i. The new server joins a DRS cluster. Play: All VMs in the cluster are automatically rebalanced taking in consideration the newly available resources On-demand capacity Easy scalability How does this enable a plug and play datacenter? Plug in and run ESX Server 3i. DRS takes over and automatically manages this new compute capacity, migrating Virtual Machines to optimally load balance. Adding capacity and scaling your datacenter becomes a simple, non-disruptive process. ESX Server 3i nodes can be viewed as standardized compute nodes: No installation required Small footprint Reliability and manageability of hardware

Traditional ESX Server
98% 2% Agent … Agent RPM RHEL3-based Service Console Helpers VMM VMM VMM VMkernel Storage Networking Traditional ESX Server has a RHEL service console attached to it that is used primarily for booting up from a variety of different hardware, to run scripts and hardware monitoring agents etc. While ESX Server is not dependent on this service console ( the Console can be non-responsive but have no effect on VMs) – there is still extra effort administrators have to spend in securing this console, patching it and maintaining it Resource Management HAL and Device Drivers Disk Footprint: 2 GB Disk Footprint: 32 MB Percent of Patches 50%

ESX Server 3i: Thin Virtualization!
98% 2% Agent … Agent RPM RHEL3-based Service Console Helpers VMM VMM VMM VMkernel Storage Networking ESX Server 3i now comes pre installed in the hardware eliminating the need for a service console. Scripts can be run in a remote CLI. Hardware monitoring agents can talk to the underlying hardware through CIM. The resulting compact 32MB footprint hypervisor is more easily maintained, updated and secure. Resource Management HAL and Device Drivers Disk Footprint: 2 GB Disk Footprint: 32 MB Percent of Patches 50%

Infrastructure Optimization Business Continuity Desktop Management Software Lifecycle  Management & Automation Resource Mgt Availability Mobility Security Virtual Infrastructure DRS HA VCB Storage VMotion VMotion Update Manager In Summary – a wide range of products and features from VMware enable further rapid adoption across the board. Extending out the virtualization capabitilie of VMware virtual infrastructure helps make it the right choice for virtualization in datacenters worldwide. ESX Server 3i 3.5 ESX Server 3.5 Virtualization Platform Œ 19

62% of VMware customers have implemented VMotion
VMware VMotion 62% of VMware customers have implemented VMotion Live migration of virtual machines Zero downtime In the are of mobility, VMotion lets you move live, running virtual machines from one host to another while maintaining continuous service availability. Neither users nor the application / VM know it’s been moved. Applications don’t have to be taken off line. VMotion is a huge advantage when doing hardware maintenance Supported on Fibre Channel and iSCSI SAN and NAS

Storage VMotion Storage independent migration of virtual machine disks
Zero downtime to virtual machines LUN independent Supported for Fibre Channel SANs Minimizes planned downtime due to storage Complete planned downtime management solution across servers and storage with VMotion and Storage VMotion Storage VMotion completes the non-disruptive management of planned downtime Migrating single VMs and their disks from one array to another is accomplished today by moving entire LUNs with datamovers/SAN tools – almost always involving downtime Now performance optimization of VMs with the right type of storage becomes a trivial problem to solve HOW DOES IT WORK: VMware Storage VMotion allows virtual machines’ disks to be relocated to different data store locations completely transparently while the virtual machine is running, with zero downtime. Storage VMotion takes advantage of core technologies that VMware has developed, such as disk snapshots, REDO logs, parent/child disk relations, and snapshot consolidation. Before moving a virtual machines disk file, Storage VMotion moves the “home directory” of the virtual machine to the new location. The “home directory” contains meta information about the virtual machine i.e. configuration, swap, log files. It next “self VMotions” to the new VM home location. The disk movements follow after the home directory migration. First, Storage VMotion creates a “child disk” for each virtual machine disk to be migrated. Once the migration operation has started, all disk writes are directed to this “child disk”. Second, the “parent”, or original, virtual disk is then copied from the old storage device to the new storage device. In step three, the child disk that is capturing the write operations are then “re-parented” to the newly copied parent disks. And, in the final step the child disk is consolidated into the new parent disk and the ESX host is now re-directed to the new parent disk location. The switchover process of home directory and disk migration, creation of child disks and parent disks, re- parenting and consolidation of child disks happens within sub-2 seconds, fast enough to be unnoticeable to the application user.

Resource Optimization with VMware DRS
Dynamic and intelligent allocation of hardware resources to ensure optimal alignment between business and IT What is it? Dynamic balancing of computing resources across resource pools Intelligent resource allocation based on pre-defined rules Customer Impact Align IT resources with business priorities Operational simplicity; dramatically increase system administrator productivity Add hardware dynamically to avoid over-provisioning to peak load Automate hardware maintenance Business Demand VMware® Distributed Resource Scheduler (DRS) dynamically allocates and balances computing capacity across a collection of hardware resources aggregated into logical resource pools. VMware DRS continuously monitors utilization across resource pools and intelligently allocates available resources among the virtual machines based on pre-defined rules that reflect business needs and changing priorities. When a virtual machine experiences an increased load, VMware DRS automatically allocates additional resources by redistributing virtual machines among the physical servers within the network. VMware DRS allows IT organizations to: Prioritize resources to the highest value applications in order to align resources with business goals. Optimize hardware utilization automatically and continuously to respond to changing conditions. Provide dedicated (virtual) infrastructure to business units while giving central IT complete control over hardware. Conduct zero-downtime server maintenance How does VMware DRS work? VMware DRS dynamically allocates and balances computing capacity across collections of hardware resources aggregated into logical resource pools. VMware DRS continuously monitors utilization across the resource pools and intelligently allocates available resources among virtual machines. VMware DRS allows users to define the rules and policies how virtual machines share resources and how these resources are prioritized among multiple virtual machines. When a virtual machine experiences increased load, VMware DRS first evaluates its priority against the established resource allocation rules and policies, and if justified, allocates additional resource. .Additional resources are allocated to the virtual machine by either migrating it to another server with more available resources or by making more “space” for it on the same server by migrating other virtual machines to different servers.. VMware DRS can be configured to operate in automatic or manual mode. In automatic mode, it migrates virtual machines to the most appropriate physical servers. The live migration of the virtual machines to the different physical server is executed completely transparent to end-users though VMware VMotion. In manual mode, VMware DRS provides a recommendation for optimal placement of virtual machines, and leaves it to the system administrator to decide whether to make the change. Flexible hierarchical organization of resource pools allows administrators to match available IT resources to the needs of the business organization. Individual business units can receive dedicated IT resources while still benefiting from the efficiency of resource pooling. Robust access privileges make it possible to delegate routine infrastructure tasks for a business unit resource pool to a business unit administrator. VMware DRS optimizes IT environments to align resources with business goals while ensuring flexibility and efficient utilization of hardware resources. How Is VMware DRS Used in the Enterprise? Align IT resources with business priorities. Define rules and policies how resources are prioritized among virtual machines. VMware DRS dynamically and intelligently allocates IT resources to the highest priority virtual machines to ensure optimal alignment between business and IT. Guarantee IT autonomy and service levels to business organizations. Provide dedicated IT infrastructure to business units while still profiting from higher hardware utilization through resource pooling. Dramatically increase system administrator productivity. Enable a single system administrator to monitor and effectively manage a large pool of infrastructure resources. Automate hardware maintenance. Place a physical server in maintenance mode and VMware DRS will automatically migrate all virtual machines to other physical servers, allowing server maintenance with zero downtime. Resource Pool

Distributed Power Management (experimental)
Consolidates workloads onto fewer servers when the cluster needs fewer resources Places unneeded servers in standby mode Brings servers back online as workload needs increase Resource Pool Minimizes power consumption while guaranteeing service levels No disruption or downtime to virtual machines Most servers consume 50% of their peak power requirement even when idle. Distributed Power Management helps you really manage your power bill without compromising on resource availability to virtual machines Put host in stand-by mode if: total demand + reserve <= total capacity minus host capacity MORE DETAIL FOR INTERESTED CUSTOMERS: Users can define: Reserve capacity to always be available Time for which load history can be monitored before the power off decision is made. Time for which load history can be monitored before the power on decision is made. Power on will also be triggered when there aren’t enough resources available to power-on a VM or when more spare capacity needed for HA. Stand-by mode means the host is powered off (S5). There are sub-modes to standby mode: When a recommendation to go into standby is accepted, the host immediately enters standby/entering mode. In this mode the machine is still powered on and may still have some VMs running on it. They must be migrated off, and no new VMs can be started/migrated on the host. When the host is evacuated, it can be powered off (standby/off) At some later point, if a recommendation to leave standby mode is accepted, the machine is powered up and begins to boot (standby/leaving). No VMs can be migrated on it until it's fully up. Automation options: Automation level Automatic – DPM makes power on/off recommendations and executes them Manual – DPM makes power on/off recommendations and admin chooses whether or not to execute them Off – DPM is disabled Automation level defined on per-cluster & per-host basis Advanced options for tuning load parameters Manual power off/on of hosts Requirements and considerations: Running Virtual Center with DRS VMotion compatible cluster Shared storage Compatible processor architectures VMkernel/VMotion network Hosts with Wake on LAN NICs with Wake on LAN Configure these as the VMotion NICs Test Wake on LAN prior to using DPM in production Manually power off and then power on each host using VC UI Physical Servers

Zero-downtime maintenance using VMware
Title Month Year Zero-downtime maintenance using VMware Use VMotion to evacuate hosts Move running applications to other servers without disruption Perform maintenance at any time of day Automate with DRS maintenance mode Automates moving virtual machines to other hosts Automates re-balancing after maintenance complete VMotion VMotion DRS migrates running virtual machines to other hosts Activate Maintenance Mode for physical host Shut down idle host and perform maintenance [Note: slide has manual animation to step through DRS Maintenance Mode sequence of events] Key Points: Just with VMotion, you can move running virtual machines off of a server for planned maintenance downtime For customers using DRS, Maintenance Mode further automates process of evacuating (i.e. moving virtual machines off of) a physical server Some scenarios this applies to: firmware updates, replacing failed components (that didn’t bring down the server), adding or replacing memory Restart host; DRS automatically rebalances workloads 24

Ensure High availability with VMware HA
VMware HA enables cost-effective high availability for all applications What is it? Automatic restart of virtual machines in case of server failure Customer Impact Cost effective high availability for all applications No need for dedicated stand-by hardware None of the cost and complexity of clustering X What is VMware HA? VMware® High Availability (HA) provides easy to use, cost effective high availability for applications running in virtual machines. In the event of server failure, affected virtual machines are automatically restarted on other production servers with spare capacity. VMware HA allows IT organizations to: Minimize downtime and IT service disruption while eliminating the need for dedicated stand-by hardware and installation of additional software. Provide uniform high availability across the entire virtualized IT environment without the cost and complexity of failover solutions tied to either operating systems or specific applications. How Is VMware HA Used in the Enterprise? VMware HA allows companies to provide high availability to any application running in a virtual machine. With VMware HA IT organizations can: Protect applications with no other failover option. Provide cost-effective high availability for any application running in a virtual machine. High availability solutions are often relatively complex and expensive, and typically reserved for mission critical applications. VMware HA provides a cost-effective high availability solution that makes high availability possible for software applications that were formerly left unprotected. Establish consistent “first line of defense” for an entire IT environment. Unlike other high availability solutions that are operating system or software application specific, VMware HA represents a consistent, easy to manage high availability solution for the entire IT environment. VMware HA provides basic failover for any application with minimum cost and management overhead. How Does VMware HA Work? VMware HA continuously monitors all servers in a resource pool and detects server failures. An agent placed on each server maintains a “heartbeat” with the other servers in the resource pool and a loss of “heartbeat” initiates the restart process of all affected virtual machines on other servers. VMware HA ensures that sufficient resources are available in the resource pool at all times to be able to restart virtual machines on different physical servers in the event of server failure. Restart of virtual machines is made possible by the Virtual Machine File System (VMFS) clustered file system which gives multiple ESX Server instances read-write access to the same virtual machine files, concurrently. VMware HA is easily configured for a resource pool through VirtualCenter. Resource Pool

VMware Consolidated Backup
Agentless file level backup Move backup out of the virtual machine A single agent running on the proxy server rather than an agent on every virtual machine. Eliminate backup traffic on the local area network Utilize Fibre Channel tape devices for virtual machine backups Pre-integrated with major 3rd-party backup products VMware® Consolidated Backup provides an easy to use, centralized backup facility for virtual machines. It enables virtual machine contents to be backed up from a centralized Microsoft® Windows 2003 proxy server rather than directly from ESX Server. Reduce the load on ESX Server by allowing it to run more efficiently and run more virtual machines. Improve manageability of IT resources by using a single agent running on the proxy server rather than an agent on every virtual machine. Eliminate backup traffic on the local area network by utilizing Fibre Channel tape devices for virtual machine backups Copyright © 2005 VMware, Inc. All rights reserved. 26

* Note: RHEL guests can only be scanned, not remediated
VMware Update Manager Automates patch management for ESX Server hosts and select Microsoft and RHEL virtual machines Scans and remedies online as well as offline virtual machines* and online ESX Server hosts Snapshots virtual machines prior to patching and allows rollback to snapshot OFFLINE Eliminates manual tracking of patch levels of ESX Server hosts and virtual machines Automates enforcement of patch standards Reduces risk through snapshots and offline virtual machine patching What is VMware Update Manager? It is an automated patch management solution for VMware ESX hosts as well as Microsoft and Linux virtual machines Two main benefits compared to traditional patching solutions : 1. patching of offline /suspended machines is done securely. Non compliant machines are patched in a quarantined state so that the rest of the network is not exposed to them 2. Can patch and update ESX server and ESX Server 3i MORE DETAIL FOR INTERESTED CUSTOMERS: VMware Update Manager is used to enforce compliance to patch standards in four steps: 1.Getting information on the latest patches: V Mware Update Manager automatically gathers the latest patch data from VMware as well as application vendors such as Microsoft, Adobe and Mozilla via the Internet. To enable secure off-network patching, VMware Update Manager has a disconnected patch mode. A separate download service assists Update Manager in gathering patch data used for off-network patches and updates. 2. Set baselines : The information collected by Update Manager is used to set baselines . Baselines contain one or more service packs, patches and /or updates The baselines data that Update Manager gathers gives IT administrators granular control in defining patch levels. These baselines updates can be static baselines defined manually or dynamic baselines that are set automatically depending on the significance of the patch data from the system vendor. 3. Compare physical hosts and virtual machines against the baselines: VMware Update Manager scans the state of the physical VMware ESX Server hosts as well as select Microsoft and Linux guest operating systems and compares it with baselines set by the administrator. Scans can be initiated on entire data centers, clusters, resource pools, templates, folders or individual hosts and virtual machines. They can be run immediately or scheduled as necessary. After a scan is complete, non-compliant machines are flagged for patch updates. Remediate the selected set of physical or virtual machines. Virtual machine patching: VMware Update Manager supports either manual or scheduled patching of the non compliant machines. If a reboot is required on a manual patch or update, the administrator has the option to reboot immediately or delay system restart by up to 60 minutes. To reduce the risk of virtual machine patching failures, the VMware Update Manager automates snapshotting the virtual machine state prior to applying a patch. Snapshots are stored for a user-defined period so administrators can rollback a patched virtual machine to a known working state if there are any problems VMware Update Manager also patches offline or suspended virtual machines. When remediating offline or suspended virtual machines, VMware Update Manager disables their NICs during the patching so the network is not exposed to non-compliant virtual machines. Update Manager Host Server * Note: RHEL guests can only be scanned, not remediated

Non-disruptive ESX Server Patching with Update Manager and DRS
Update Manager patches entire DRS clusters Each host in the cluster enters DRS maintenance mode, one at a time VMs are migrated off, host is patched & rebooted if required VMs are migrated back on Next host is selected Update Manager VMotion VMotion Resource Pool DRS enabled patching brings new levels of automation and resource management to the datacenter With DRS enabled patching, system admins can be almost completely handsoff in applying patches to ESX Server hosts. DRS manages the entire process for entire clusters of physical hosts with NO disruption or downtime to any of the virtual machines Automates patching of large number of hosts with zero downtime to virtual machines

Update Manager Benefits
Guest Patching Administrative time – 3064 hrs, $153,200 saved annually Calculated for 100 virtual machines, assuming 75 patches per machine Scan machines Assess patch requirements Remediate systems Troubleshoot Rollback Per virtual machine Per patch Manual 15 min Automated 6 min Manual 18 min Automated 6 min Manual 33 min Automated 12 min Annual Savings for 100 VMs 1125 hrs, $56,250 Annual Savings for 100 VMs 1939 hrs, $96, 950 Annual Savings for 100 VMs 3064 hrs, $153, 200 In a typical 100 VM environment, you save on 1 FTE dedicated to patching and maintaining VMs. This value is delivered in all three VI 3 Editions at no extra price! Offline Machine Patching Reduces exposure from non-compliant offline/suspended virtual machines Systems have NICs disabled during patching to reduce risk

Infrastructure Optimization Business Continuity Desktop Management Software Lifecycle  Virtual Center Site Recovery Manager VDI ACE Lab Manager Management & Automation Resource Mgt Availability Mobility Security Virtual Infrastructure DRS HA VCB Storage VMotion VMotion Update Manager In Summary – a wide range of products and features from VMware enable further rapid adoption across the board. Extending out the virtualization capabitilie of VMware virtual infrastructure helps make it the right choice for virtualization in datacenters worldwide. ESX Server 3i 3.5 ESX Server 3.5 Virtualization Platform Œ 30

Guided Consolidation Automatically discovers physical servers
Analyzes utilization and usage patterns Converts physical servers to VMs placed intelligently based on user response Discover Analyze Lowers training requirements for new virtualization users Steers users through the entire consolidation process For first time virtualization users, a new feature in VirtualCenter 2.5 guides users through the process of server consolidation. Recommended for smaller, simpler environments – this feature steers users through discovering physical servers, collecting performance data from these servers and converting these servers to virtual machines placed intelligently on the most appropriate hosts. Allows new users to quickly realize benefits from server consolidation, reduces training requirements for first time “virtualizers” Current constraints in Guided Consolidation – FOR CUSTOMERS WHO ARE INTERESTED: Guided Consolidation currently requires default username and password (needed for discovery) to only contain ASCII strings. As such the discovery feature may not work on non-English OSes. This is essentially a bug in collector which currently does not support Unicode. (we hope to fix this by GA) Users cannot limit discovery to a certain IP range or filter search results based on wildcards or other search criteria Guided consolidation does not discover across multiple AD sources as the collector engine can only discover and report on domains managed by a single AD source Guided Consolidation relies on AD and/or LANMAN to discover machines on the network and gather data on the same. Systems that have LANMAN service disabled (typical in larger enterprise environments) will not show up in discovery in the absence of AD. In such a scenario, VC needs to be installed on a system that is part of an AD domain and the user needs to provide credentials that has read access on the AD domain to begin discovery and data gather. Needless to say in the absence of both LANMAN and AD, guided consolidation will not be able to discover any physical systems Convert

Virtual Desktop Infrastructure
Virtual desktops hosted on VMware Virtual Infrastructure Virtual Infrastructure optimizes and manages servers to deploy desktops 3 1 3 Clients can be a desktop, a laptop or a specialized thin client Displays desktop of remote virtual machine 2 Secure Data Center VirtualCenter Virtual Desktops 1 Virtual Desktop Infrastructure is built on VMware virtual infrastructure hardware. Desktop virtual machines are hosted on server hardware running VMware Virtual Infrastructure software. These desktop environments are then accessed remotely from a thick or thin client via a remote desktop protocol such as RDP. ESX Server 2 2 3 Connection using remote display software (e.g. RDP, VNC, etc.) 2 3 Copyright © 2006 VMware, Inc. All rights reserved. 32

VDI – Virtual Desktop Manager (VDM)
VMware VDM Centralized Virtual Desktops Enterprise-class, scalable connection broker Central administration and policy enforcement Automatic desktop provisioning with optional “smart pooling” Desktop persistence and secure tunneling options Microsoft AD integration and optional 2-factor authentication via RSA SecurID® Clients End-to-end enterprise-class desktop control and manageability Familiar end user experience Tightly integrated with VMware’s proven virtualization platform (VI3) Scalability, security and availability suitable for organizations of all sizes Note –VDM is part of our VDI offering. This is not in ESX Server 3.5/Virtual Center 2.5 Control and manageability in an end-to-end product: Central administration of desktops from any location including policy enforcement options such as session time-outs Enterprise-class connection brokering, managing connections between remote clients and their centralized virtual desktops Automatic provisioning of desktops using optional “smart pooling” capabilities Desktop persistence options: maintains desktop state on log-out or reverts to “known” state Secure tunneling via a proxy server so that all connections are completely encrypted Two-factor authentication via RSA SecurID® Full Microsoft Active Directory (AD) -based integration Policy enforcement: Familiar end-user experience: Give users similar desktop experience, just running over a remote connection. No changes to the apps. No retraining. Individuals can customize their desktop. Desktops are isolated, not shared. Usability improvements to RDP for local printing, PDA, smart cards, etc. Secure Single Sign On to desktops on same domain as client Enterprise-class deployment, scalability & high availability: Brings powerful VI3 and backend data center server capabilities to the desktop Deploy desktops in minutes instead of weeks through automated desktop provisioning Simplify backups and DR with centralized desktops that can leverage shared storage No single point of failure: fully synchronized connection mgmt servers and support for industry standard server load balancing Tighter integration with VI3 – the leading virtualization platform: Familiar VI3 look and feel for consistent administration across desktops and servers in the data center Third party extensibility Built by the leader in virtualization software (VMware)

ACE Solution Overview Create and Package Manage ACE Package
ACE Management Server ACE Workstation ACE Package

ACE on Portable Flash Media
Deploy — Pocket ACE ACE on Portable Flash Media A secure policy managed virtual machine stored on portable media Virtual machine is completely protected and isolated from the host PC Optimizations for running off portable media A better media than DVD for many use cases Reduces client-PC hard disk space constraints as a deployment concern Pocket ACE creates secure portable computer environments without a PC.

Introducing VMware Site Recovery Manager
Site Recovery Manager leverages VMware Infrastructure to transform disaster recovery Simplifies and automates disaster recovery workflows: Setup, testing, failover, failback Provides central management of recovery plans from VirtualCenter Turns manual recovery processes into automated recovery plans Simplifies integration with 3rd-party storage replication Note: Site Recovery Manager is an add-on product – not part of ESX Server 3.5/VirtualCenter 2.5 Provides a central point of management for disaster recovery plans Plug-in to VMware VirtualCenter Manage recovery plans for virtual machines managed by VirtualCenter Enables pre-programming of disaster recovery plans Map resources between production and recovery sites Program steps of recovery process Automates key disaster recovery workflows Simplified setup of recovery plans. Create, configure, and manage recovery plans from VMware VirtualCenter, making it possible to manage disaster recovery plans as an integral part of virtual infrastructure. Turn the complex, manual runbooks that are commonly required in traditional disaster recovery solutions into automated recovery workflows maintained and executed automatically at a push of a button from within Site Recovery Manager. Automated failover. Once recovery plans have been created, automate execution of those recovery plans in order to make recovery dramatically faster by eliminating the slow and unreliable manual processes otherwise required. Easier testing of recovery plans. Perform non-disruptive, automated tests of the same recovery plans that will be used in a failover scenario. Expected to be available in H will be a separately licensed component Makes disaster recovery rapid, reliable, manageable, affordable

Site Recovery Manager Key Components
Production Disaster Recovery Storage Servers VMware Infrastructure VirtualCenter Site Recovery Manager Service VirtualCenter Protected virtual machines VMware Infrastructure Servers This slide illustrates the key components of a site recovery manager deployment The requirement is for protected VMs to be running off of storage that is being replicated to a secondary site. Virtual Center is required to manage both sites Site Recovery Manager manages the mapping of components ( VMs, resource pools, networks etc) between the two sites and provides workflow automation for setup, failover, failback and test of this DR environment. Storage 3rd-party Replication

Summary of Benefits Site Recovery Manager Leverages VMware Infrastructure to Make Disaster Recovery Rapid Automate disaster recovery setup, failover, failback, and testing Eliminate complexities of traditional recovery Reliable Ensure proper execution of recovery plan Enable easier, more frequent tests Manageable Centrally manage recovery plans Make plans dynamic to match environment Affordable Utilize recovery site infrastructure Reduce management costs These features are representative of feature areas under development. Feature commitments must not be included in contracts, purchase orders, or sales agreements of any kind. Technical feasibility and market demand will affect final delivery.

Lab Manager VMware Lab Manager Image Storage Library
VMware Lab Manager determines the best host servers, then deploys the machines. User selects a multi-machine configuration, clicks deploy. VMware Lab Manager Image Storage Library LAN/SAN Application Developer Waldorf Dev Facility QA Automation Engineer BIG IDEA: This slide illustrates how Lab Manager Works (animated slide). First, the user selects a multi-machine configuration from the library and clicks deploy. Note that the “atomic unit” for Lab Manager is a multi-machine configuration, that is all of the machines of a multi-tier systems, networked together and ready for use. Next, VMware Lab Manager determines the best host servers in the server pool, and deploys the machines in that configuration onto that server pool. Once the configuration is deployed, the user directly interacts with each machine as if sitting in from of the actual machine consoles – all from within the Lab Manager user interface. Bangalore Outsource Partner VMware Infrastructure Application Developer Once deployed, user directly interacts with the machines, as if sitting at each console Virtualized Server Pool New Jersey Dev Facility Automated Virtual AD Lab

Infrastructure Wide Solutions Drive Adoption
Infrastructure Optimization Consolidate servers Reduce operating costs: real estate, power, cooling Business Continuity Reduce planned and unplanned downtime Simplify disaster recovery Desktop Management For enterprises and end users Improve security and mobility Software Lifecycle Automation Rapidly provision machines Improve software quality Let’s recap what are the most popular use cases for virtualization: Infrastructure & Capacity Optimization (aka Server Consolidation) was the original use case for virtualization, and remains one of the biggest drivers for adoption. Deploying multiple workloads on the same server means that companies can do more with less – less servers, less storage, less networking cables, less space in the datacenter, less power and cooling. This use case is applicable to the overwhelming majority of servers in the datacenter. But increasingly companies find that virtualization represents a new way of thinking about business continuity in all its forms – ensuring high availability of critical applications as well as planning for a disaster or recovering from a disaster. A new, and very exciting extension of virtualization beyond the datacenter is what we call Virtual Desktop Infrastructure – the capability to host end user desktops in the datacenter, and expose them to end users through a thin terminal. This allows companies to drastically decrease the cost of end-user desktops, and apply the same rigorous security and data protection available in the datacenter to end-user desktops. Virtualization offers a lot of flexibility to manage software lifecycle – from development, testing, QA, staging to production. These processes typically entail creating and maintaining multiple environment which now can be co-located on the same system. VMware offers capabilities to create libraries of these environments, and instantiate them as needed. Let’s illustrate these use cases with some customer examples

Consolidation: Just the Beginning
Leading North American Utility AFTER BEFORE 10 server racks 20 power whips 200 server racks 400 power whips Facilities 300 cables/ports 3,000 cables/ports Network 140 TB SAN & NAS 270 TB DAS Storage 80 1,000 Servers Impact of VMware Hard cost savings 70-80% less datacenter space, power and cooling $8 mil savings in 2 years Operational efficiency Server rebuild and app load from hrs to mins 10,000 man hours saved per year

Reduce Costs and Power Consumption
For every server virtualized, save… ~$700 and ~7,000 kWh / year $300 in carbon offsets over lifetime 4 tons of CO2 emissions / year Plus Power down underutilized physical servers, saving 40% Desktop management reduces PC power usage 35% / year 10:1 consolidation = 80-90% less energy

Consolidate, Manage, Optimize
Deploy and move VMs – easily, safely, transparently Consolidate servers, reduce costs Manage resources for high availability Automate processes; accelerate tasks Allocate resources dynamically Johnson Controls Saved $760,000 in hardware renewals in first 2 years Reclaimed 32 tons of cooling capacity Provisions new servers 80% faster QUALCOMM Drove server utilization from 5% to nearly 100%· Consolidated 12:1 Optimized server management and instant server provisioning Easily upgrade software with zero downtime

From Consolidation to Strategic Architecture
Software Lifecycle Provisioning Business Continuity Desktop Management Resource Management Service Level Management Change Management Application Availability VMware Infrastructure Consolidation & Containment

BEFORE virtualization
Instant Provisioning Week 1 Week 2 Week 3 Week 4 Week 5 Week 6 BEFORE virtualization Procure hardware Configure hardware Install OS Configure OS & Tools Assign IP Addr Configure Network WITH virtualization 20-40 hrs of work 4-6 week lead time <1 hour of work 1-2 days lead time Before virtualization the provisioning process took hrs to build a server and re-load application… Build and configure hardware Load operating system Load configuration tools (Backup, Resource Kit, Monitoring, etc…) Assign 2 IP addresses Build 3 network connections, copper or fiber Turn over to applications team to re-load and re-configure software Test applications Coordinate outage / data migration With VI min to copy a virtual machine and restart Redirect virtual disk to new VMware virtual machine instance [Tools already loaded] [Application already loaded, configured] Done Deploy from Template Power on VM

Number of Workloads per Infrastructure Admin
A measure of IT infrastructure management efficiency is how many workloads a single infrastructure admin can support 177 VMs with 2 admins 2,000 VMs with 10 admins 1,500 VMs with 5 admins Probably the best summary proof point for how VI affects IT management is one key efficiency ratio – how many servers/applications/workloads can a single admin support. Before VI, the typical ratios we see are less than 50 workloads per sys admin. With VI we see VMs per sys admin, in some extreme cases – up to 600 VMs per admin. . 700 VMs with 2 admins 1200 VMs with 2 admins

Maximize Continuity and Uptime
Systems up. Costs down. Reduce planned / unplanned downtime High availability with less complexity Disaster recovery made simpler Bowdoin College Replaced tape backup system requiring weeks for recovery Avoided $500,000 in hardware costs Saved $15,000 in annual server maintenance Bowdoin College Bowdoin: top college in Maine; 1,800 students and 800 faculty/staff. No viable DR plan: servers were backed up to tape stored in a different building, requiring weeks or months to recover in the event of a disaster. Now, Bowdoin and Loyola Marymount in LA are partners in an enterprise-class cross-country DR and business continuity solution with VI3. Results Continued to grow IT capabilities without increasing datacenter footprint Achieved higher availability, better load balancing, and enhanced fault tolerance Co-located emergency Web, DNS, and Active Directory servers: critical systems now DR-ready Subaru of Indiana Reduced downtime 40% Saved $200K in productivity due to greater uptime Avoided $20K+ in hardware expenditures

Manage All Types of Downtime
Prevent planned outages Quick recovery from unplanned outages Component NIC Teaming Multipathing Server DRS with Maintenance Mode VMotion HA Storage Storage VMotion VCB Data N/A Site Site Recovery Manager A variety of solutions available with VMware virtual infrastructure support management of all types of downtime

Manage your Desktops Manage, lock down, and control PCs wherever they are Centralized administration and control End users keep physical-PC experience Fast and easy desktop deployment Bell Canada 3000 hosted desktops deployed A new desktop in 20 minutes Can hire more telecommuters -- without security or management worries “Here VMware offers a message of bringing data center availability and security to a client operating environment - something we believe will resonate stronger then simply cost reduction."- John Humphreys, IDC, "The New Killer Server App...the Desktop?," (1/18/2007) Cardinal Health 1650 hosted desktops deployed 100 new virtual desktops per hour One desktop technician per 300 call center workers

Desktop Momentum

Optimize and Automate your Lab Infrastructure
Reduce IT costs and time required to manage lab environments Consolidate servers, networking and storage resources Rapidly provision multi-tier systems on demand Princeton Softech Improved server-to-administration ratio by 25x Reduced provisioning time -- from weeks to seconds Deferred $200,000+ in hardware costs "VMware Lab Manager is an excellent example of VMware's innovation that brings together existing features and functionality to deliver complex test and development environments with minimal effort and time.... The core value of the solution comes from three key capabilities: self-service provisioning, the ability to manage multiple VMs as an atomic unit and an image library management system." -Brian Babineau and Mark Bowker, Enterprise Strategy Group, "Identifying Successful Virtualization Solutions with VMware," (Feb 2007) AGFA One blade server enclosure supports 100-person dev team Entire virtual lab managed by one person IT no longer overwhelmed with new-machine and provisioning requests

VMware Infrastructure Adoption Trends
Agenda VMware Infrastructure Adoption Trends There are solid reasons why VMware is the most deployed virtualization solution on earth.

Customers are Increasingly Standardizing
Adoption path of VMware customers: Escalating value. >40,000 virtual Infrastructure customers 86% in production 65% running different OS on the same server 63% plan to virtualize more than 50% of their x86 infrastructure in 3 years 43% standardizing on virtual infrastructure VMware customers want more than a hypervisor – management, dynamic resource allocation etc. Customers upgrading to VI3 quickly. 66% of customers using VirtualCenter; 56% using VMotion “Virtualization will save us $60M in OpEx and CapEx through 2010.” A large VMware enterprise customer Explore Expand Standardize Source: Market Research Survey, 2006 and 2007

Customers Move Rapidly Along the Adoption Curve
Customer Example: Large Wireless Technology Company Standardization 300 1000 Expanded Rollout 250 800 200 Active Virtual Machines Departmental Rollout 600 ESX Server Instances 150 400 100 Qualcomm’s experience illustrates well the typical customer adoption phases Qualcomm started a POC in the first half of Today, Qualcomm has implemented a VMware first policy, i.e. they have standardized on VMware for x86 worklods. Today 60% of Qualcomm’s x86 environment is virtualized (1900 total servers/1150 are virtualized). The number of physical servers has grown from 950 to 1900 over the past 2.5 years, and because of the much simplified provisioning with virtualization, they have been able to maintain the same number of server admin's today. They provision 68 new VM's/month. This would be impossible in the physical world without dramatic staffing increases. Which means that the number of physical servers a single sys admin can manage has more than doubled. This translates into substantial operational savings for the company. Provisioning has gone from 6-8 weeks to between an hour and 2 days depending on the group's workload. They're is a joke within IT at Qualcomm that: "Projects kickoff in the morning and the servers are ready in the afternoon." They've built a web based, front ends and business users simply request a server via the web and it gets routed to the server admin's who determine where to place the VM etc. In aggregate, they've saved $4.5M over 3 yrs with VMware. This calculation doesn't include the additional cost for storage in the virtual world (all VM's are SAN connected now), but it also doesn't include the cost savings from power, cooling etc. Proof of Concept 200 50 2003 2004 2005 2006

Production Usage is Increasing
GP-15 N=307 B50 Please indicate the status of the following x86 virtualization software products in your company today Source: survey of customers in North America, N=307

VMware Infrastructure 3 at Work Nationwide – Increasing ROI
Physical Virtual Savings HW and VMware licensing $5,000 (HP DL38x 2-CPU server with 2 GB RAM) $2,000 $3,000 SAN Storage $1,800 ($1,800) Total $5,000 $3,800 $1,200 Physical Virtual Savings HW and VMware licensing $6,400 (HP DL38x 2-CPU server with 2 GB RAM) $1,700 $4,700 SAN Storage $250 ($250) Total Upfront $6,400 $1,950 $4,450 Monthly support $917 $675 $242 56

Strategic Business Benefits: agile, flexible, always-on
Your Adoption Path Start with immediate needs. Scale for enterprise-wide benefits. Strategic Business Benefits: agile, flexible, always-on Automate IT processes Create resource pools Capacity on-demand Infrastructure-wide management OpEx Savings Virtualize production servers High Availability Disaster Recovery Balance workloads CapEx Savings Partition Consolidate servers Explore Expand Standardize Source: IDC, Virtualization 2.0: The Next Phase in Customer Adoption, Nov 2006

The End!

VMware Infrastructure

Similar presentations

Presentation on theme: "VMware Infrastructure"— Presentation transcript:

Similar presentations

About project

Feedback

Log in

Auth with social network:

VMware Infrastructure

Similar presentations

Presentation on theme: "VMware Infrastructure"— Presentation transcript:

Similar presentations

About project

Feedback