1. Regional Seminar Series Kirkland, WA
Virtualization and HA PI Systems: Strategies to Keep Your PI System Available, Scalable, and Portable
Chris Coen
Product Manager
OSIsoft, LLC
November 10, 2009

2. Overview Virtualization PI High Availability Server
Storage (DAS, NAS, SAN)
Application (Clients)
PI High Availability
Highly Available (HA) PI
HA PI in a virtual environment

3. Why Virtualization and PI?
You can realize substantial benefits using the combined strategies of virtualization (storage, server and application) and PI Collectives (HA).
These strategies provide you with:
Increased reliability
Reduced hardware and maintenance costs
Improved scalability
Use them separately or together
All four strategies (3 virtualizations and HA) can be used both separately and together

4. Why Now? You need to do more with less
Your projects need to show immediate ROI
IT is challenged to increase service levels with less staff
Virtualization and HA are valuable separately, but better together
Our customers are being challenged to do more with less. In some instances, only projects that show immediate ROI within 6 to 12 months are being approved. As budgets are cut or remain flat and IT staff are being asked to increase service levels, virtualization has proven to reduce both capital and operating expenditures and provide IT staff the required flexibility to rapidly respond to evolving business challenges.

5. Who Needs This?
PI users who cannot afford disruption in service (even for planned maintenance)
IT organizations looking to consolidate management of computing resources (fewer servers to buy and maintain)
IT organizations looking to streamline deployment of new tools for the user community (less IT time and resources)
IT organizations investigating new ways to provide ever- increasing amounts of storage for mission critical systems
A PI system administrator tasked with scaling PI to more users and other information systems
Companies investigating virtualized test environments for validating new software purchases
Disruption includes planned maintenance
Ordered by highest payback first. Consolidation=fewer physical servers.
These three technologies offer benefits on their own, but can also build upon one another. Here’s a story:
An IT System Engineer (or whatever you want to call the guy in charge of infrastructure) is trying to implement a more available PI system because as PI becomes more visible across the organization and used by more people and even other information systems, the need to keep it online increases. So, he investigates the PI HA solution and learns that he can support multiple members of a collective and keep them all synchronized with the same data (to promote failover and redundancy). Unfortunately, his hardware budget has been slashed and he can’t purchase additional hardware at this time.
Luckily, the IT department has started a virtualization project to help lower their TCO for various data and application servers throughout the enterprise. So the IT SE wants to provide HA PI using virtualized servers, which he can afford to implement right now.
Looking at his data storage and backup needs, he realizes that he needs a network storage solution to ensure that the PI Collective is not only available, but capable of expanding to fill the ever growing demand for its data. In addition, he realizes that he can centralize the storage of PI Client application content (ProcessBook displays and DataLink reports).
Finally, he can realize significant client software deployment savings by implementing a Terminal Services Gateway to virtually serve applications across the enterprise instead of installing them on every desktop.

6. A Simple System

7. A Simple HA PI System
Benefits discussed on next slide. Here discuss some scenarios and functionality .
Failover
Load balancing
Interface writing to all members at same time
Collective members don't need to be identical hardware.
Geographical separation

8. Built-in Benefits of HA PI
PI is there all the time – users trust it
No late night heroics to restore a backup or perform routine maintenance
Removes fear of a bad backup
Simple design is robust, low bandwidth and supported by WANs
Geographical independence (replace PI to PI)
Support more or specialized users
Facilitates capacity planning
Complements virtualization strategies:
PI is perfect for monitoring a virtualized environment (HyperV performance counters; VMWare SNMP interface)
See also this white paper on the OSIsoft web site:
This slide content revised from Jon P’s presentation:
Build and maintain trust with PI System users—PI is there all the time.
HA is insurance against ruined weekends or late night IT heroics required to restore a backup.
HA removes the fear of a bad backup, which could lose days of data.
HA PI Systems complement virtualization strategies.
Simple TCP/IP- and application-based design is robust, low bandwidth and supported by WANs.
HA is a great way to support more users or specialized users, and facilitates capacity planning.
See Selling HA presentation materials by Jon Peterson (5/7/2009)
From David Soll:
PI IT Monitor tools and virtual system performance counters (especially those in Hyper V) are a perfect match for monitoring your Virtual environment.

9. A Simple Virtualized System
Emphasis on the fact that different server-based applications can be installed on virtual machines and operated on the same (physical) virtual host. Each virtual machine has it’s own Operating System, so this design can also support server-based applications that have competing OS requirements or that cannot be installed on the same server for some reason.

10. Virtualization Servers Storage Applications
What it is, the benefits, how it applies to PI

11. Server Virtualization
Instead of having physical machines, virtual servers run on a physical host
Case Study: AtlantiCare
Eliminated need to expand or relocate data center
Microsoft® Virtual Server 2005 used to consolidate infrastructure and legacy application servers
Consolidation ratio achieved of 33:2
Server consolidation: By consolidating multiple workloads onto a single hardware platform, you can maintain “one application / one server” while reducing server sprawl. This way, you can fully provision the business with less hardware, resulting in lower equipment costs, less electrical consumption for server power and cooling, and less physical space required for the server farm.
Data came from Microsoft.

12. Example: Server Consolidation
Typically server workloads only consume a small fraction of total physical server capacity, wasting hardware, space, and electricity OS
APP 9%
utilization 6%
14%
30%
Through virtualization, these workloads can be consolidated onto fewer physical servers, saving resources and increasing flexibility
The optimization of server applications that do not require full CPU or RAM capacity on a physical server may be achieved through server virtualization. The reduced power consumption of a single physical box (in this case) is also an advantage.

13. Application Virtualization Presentation Virtualization
Reduce System Costs
Challenges:
Terminal server silos
Server underutilization
Solution: Enable any applications to run side by side on any server at the same time
Application Virtualization
Presentation Virtualization
Case Study: Russell Investment Group
Consolidated servers by over 40 percent
Support legacy and LOB Applications: Terminal Services and application virtualization can enable applications written on older operating platforms to be supported in a current operating system without software code revisions.

14. Increase Availability
Challenge:
Providing disaster recovery for business continuity for operating systems and applications
Backup
Solution: Virtualize operating systems and applications—to support easy backup, replication, and movement to available servers
Robust disaster recovery: A virtualization strategy helps you to maintain an instant failover plan that provides business continuity throughout disruptive events. With the right tools, you can support automated backup, replication, and rapid movement of servers, desktops, and applications.
Maximize uptime: By compartmentalizing workloads, you can prevent one application from affecting the performance of another or causing a system crash. Even unstable legacy applications can be operated in a secure, isolated environment.
Server A
Server B

15. Increase Availability
Challenge:
Providing disaster recovery for business continuity for operating systems and applications
Backup
Microsoft
System Center
Virtual Application
Server A
Solution: Virtualize operating systems and applications—supporting easy backup, replication, and movement to available servers
You can also back up your applications and desktop images, for rapid recovery following a catastrophic event. Because the applications are residing in one centralized location and are streamed on demand, it isn’t necessary to rebuild individual images across the organization.
Microsoft
System Center Virtual Application
Server B

16. Benefits of Virtualization*
Increase user density
Improve availability
Scalability is a click away
Reduce costs
*Gillen, A., Grieser, T., Perry, R Business Value of Virtualization: Realizing the Benefits of Integrated solutions. IDC.
Support more users with fewer physical servers

17. Benefits of Server Virtualization*
Less hardware required (HP went from 85 data centers to 6)
up to 35% reduction of annual server costs per user ($100-$200K per year per server)
Better utilization of hardware (HP decreased servers by 40%)
Reduce power consumption (HP reduced energy by 40%)
Provide higher availability by supporting redundancy
Rapidly deliver adaptive and reliable IT services
Tie diverse components together into a single managed entity
Storage efficiency can lead to higher storage utilization
*Gillen, A., Grieser, T., Perry, R Business Value of Virtualization: Realizing the Benefits of Integrated solutions. IDC.
Bristol Myers Squibb is completely virtualized now; they are looking to virtualize RtReports and PI WebParts right now.
The NOC is completely virtualized
All EA customers are asking about this now
Shell is moving in this direction (they are moving their testing environment there now)
IDC white paper:
Hewlett Packard (HP) is using its own virtualization platform to transform its data centers (see
The transformation is expected to enable HP (1) to:
Reduce spending on internal IT from approximately 4 percent of revenue in 2005 to less than 2 percent in 2009;
Consolidate more than 85 internal IT legacy data centers globally to six next-generation data centers in three geographic locations equipped with new, standardized and automated technology. These data centers have 342,000 square feet of computing “white space” – expandable to more than double that amount – to accommodate growth, including acquisitions such as EDS;
Consolidate more than 6,000 applications running the business to approximately 1,500 standardized applications;
Reduce annual energy consumption in its data centers by 60 percent;
Decrease the number of servers by 40 percent while increasing processing power by 250 percent, by utilizing HP virtualization and energy-efficiency technologies;
Reduce networking costs by 50 percent while tripling bandwidth;
Eliminate more than 700 data marts and create one enterprise data warehouse where employees are accessing consistent data to make business decisions; and
Through portfolio management, deliver hundreds of high-priority business innovation projects while transforming the company’s IT infrastructure and operations.

18. PI and Server Virtualization
Validated environments need a test bed (any pharmaceutical company; BMS; Shell)
Environments that require portability of IT assets (Cargill Deicing Technology – Salt mining)
Deploying new sites (Rio Tinto)
Known customers using VMs: BMS, NOC, Shell (Calpine, but don’t mention as they are having financial issues)
Need to be portable: Cargill Deicing Technology UC Presentation:
Validated environments: Amgen UC Presentation:
Casual users: Inco Limited (ActiveView issue in particular) Clientele call #241805) – metals and mining
New sites: acquisitions and mergers (paper industry--IP, mining industry—Rio Tinto)

19. Storage Virtualization
Challenge: Grow available storage space without disrupting applications and servers
Solution: Storage Area Networks (SAN) allow dynamic sizing of available storage
Data Store
Dynamically size the storage for each virtual server.
Server A
Server B

20. Storage Types and Virtualization
DAS: Direct Attached Storage (local hard drives)
NAS: Network Attached Storage (e.g., mapped drives)
SAN: Storage Area Network, essentially virtual storage
SAN—arrays of SCSI disks; hosting system with sophisticated RAM based cache. Fibre Channel access from computer. Logical units the can be sized on the fly. Effectively virtualization of storage.
A storage area network (SAN) is an architecture to attach remote computer storage devices (such as disk arrays, tape libraries, and optical jukeboxes) to servers in such a way that the devices appear as locally attached to the operating system. The cost and complexity of SANs are dropping.
Network attached storage (NAS), in contrast to SAN, uses file-based protocols where it is clear that the storage is remote.
Both are used to provide virtual storage
For more information:

21. Benefits of SAN Technology
Additional storage appears to be local to the host so users don’t have to know where the files are stored
Improve the ties between centralized storage and virtual infrastructure
Provide virtual-machine consistent backups for data stores and the ability to restore virtual machines in a few clicks
Provide relief from disk subsystem access in virtualized environments (biggest performance hit on virtual host)
Consolidate disk resources
We have customers who provide desktop or local shortcuts pointing to client application files (e.g., PI ProcessBook displays) so that they don’t need to know where the files are actually stored. This solution removes the need for that workaround.
Vendors are integrating with virtualization vendors to provide network-based storage that works with the file format and operation of virtual machines to enable seamless storage growth and backup capabilities.

22. PI and Storage Virtualization
Keep more and higher fidelity data online; add or expand PI archive files
Support aggregated PI Systems; VSS support enables PI backups
Store PI Client files centrally
Backup virtualized application and data servers
Backup virtualized Terminal Server hosts
Complete system backup storage
No specific customer examples for this, but it is easy to see how the need would be present.
Consolidate file storage and backups.

23. Application Virtualization

24. Application Virtualization
Customers currently use Citrix or Terminal Server to reduce deployment costs and maintenance for client apps
Windows 2008 Server offers a service that provides applications over an SSL connection (HTTPS) without client- side deployment (a thin deployment) – Terminal Services Gateway
Terminal Services Gateway provides URL access to a host (like Remote Desktop connections, without the VPN requirement) or to specific applications on a host (even more secure for those outside the firewall)
Make sure to explain implications on licensing—concurrent is straightforward, but seat (individual) basically means only one person can use that license.

25. Benefits of Application Virtualization
One point of installation makes deployment simpler
Access to applications secured
All users have the same version of the software; no version or compatibility issues
Casual users do not need to install anything to get started
Save money on hardware upgrade investments by deploying client software in one place
Whenever PI ProcessBook is released, for example, customers find issues with editing the same file in two different versions of the application. This solution removes that problem. Furthermore, using a Storage Area Network (SAN) makes the (PI ProcessBook) files look local to the user, even if they are stored somewhere out on a disk array. Therefore, there is no need to create workarounds, like Windows shortcuts, to the files.

26. PI and Application Virtualization (ProcessBook)OR
Launch from Desktop icon
Launch from web page
There are two ways to provide access to virtualized applications through Windows Terminal Services 2008.

27. PI and Application Virtualization (ProcessBook)
This technique makes it easy to get ProcessBook and DataLink to many users. Of course, any time we make it easier for our customers to use our products, the better off we and our customers are.

28. PI and Application Virtualization
Environments with casual client users who need low barrier to entry for system access (Inco Limited)
Terminal Server users (a partial list)
Georgia Pacific, Kellogg, SASO, SAPPI Fine Paper, Wacker Chemie, Alcoa, Eli Lilly, ExxonMobil Upstream, Iberdrola, Progress Energy Services
Citrix users (a partial list)
SDG&E , Water Corporation, Amgen, Bayer Material Science, Genmab, PPG, Vaxgen, Katahdin Paper, Celanese Chemicals, Novo Nordisk, Queensland Alumina, Total
Windows 2008 Terminal Services Gateway
OSIsoft
Biggest question is licensing:
Client tools have concurrent licensing, which allows a number of sessions on a Terminal Server/Citrix box without needing a license for each individual user.
Lots of customers have tried both Citrix and Terminal Server

29. Five Principles for Virtualization Success*
Treat virtual machines as if they were physical machines
Invest in Enterprise-level hardware and software
Do not mix virtual and physical on the same host
Use qualified Virtualization support personnel
Test on the target platform
*OSIsoft Center of Excellence
Provide VMs with the same RAM/CPU/disk allocation you would provide for a real machine
Use Enterprise class software and hardware to get enterprise class results (e.g., you wouldn’t put a production PI Server on a laptop): This means, for virtualization, Microsoft HyperV or VMWare ESX Server (GSX is freeware and not recommended). For disk access (running multiple virtual machines on a host), a SAN is recommended to relieve the stress of local disk access.
You can mix physical and virtual implementations, just not on the same host, since the physical implementation will compete for resources and win at the expense of the virtual implementation
Use staff knowledgeable about virtualization to manage and monitor your virtual environment, just as you would for any other IT resource
Like anything else, if you plan to deploy there, you must test there as well.

30. Dynamic system management
Challenge:
Inability to respond quickly to changing needs at desktops and the data center
Solution: Live migration and dynamic provisioning of servers and desktops are based on real-time needs using a single management platform
Streamlined provisioning: Adding workload resources can be accelerated and decoupled from a hardware acquisition process. If a particular business process requires an additional capability to meet business needs (for example, a Web commerce engine), adding this capability is streamlined and immediate. In an advanced virtualized environment, workload requirements can become self-provisioning, resulting in dynamic resource allocation.

31. How does HA PI play into virtualization?
PI collectives (HA) and interfaces
Virtualized HA PI

32. A Simple Virtual HA PI System
Two servers, maybe focus on virtualizing this environment; also add a slide for the client virtualization
Start with 1 PI Server and interface nodes, build animation to a more complex system that features redundancy, virtualization, SAN, etc.
Virtual Machines (e.g., PI Collective, AF Server)
Hosts repeated multiple times for scalability as well as reliability
Advantages:
Dedicate PI Servers to specific tasks:
Data mining
Testing
Deployment
High Availability is designed for geographic separation (disaster avoidance)

33. Virtual HA PI with SAN
Application = PI Server or PI System. This approach is ideal to help the IT shop deliver QoS; i.e., one VM hosting PI for “real time” users and the other for “data miners”. Give a reason for customers to buy HA. Of course, as hosts are repeated, the interface node needs to fan more. There will be ramifications on the network and thus diminishing returns.
You would want your PI Server collective members on separate hosts to protect against host failure. This illustration shows what a particular host might look like.
Challenges include:
Sizing the host;
determining the ideal number of supportable virtual machines per host;
determining the number of hosts;
sizing each virtual machine
Backing up the virtual machines
Backing up the hosts
Expanding the storage available to the virtual machines and hosts
Replicating completed configurations to deploy new sites, if needed

34. Virtual System including Clients
LAN, WAN, Extranet connectivity available. Not cloud computing, SaaS or S+S.
Connect to applications even from the Internet, past firewalls, with Terminal Services Gateway. The Gateway can be virtualized as well.
PI Collective members can be on separate virtual hosts, across WAN links, etc.
The SAN is not a single physical entity. The best are integrated with the virtual technology so that they appear to be locally attached.

35. Recommendation: Virtualized PI System
Multiple hosts (cluster)
Collective can be split across hosts
PI Server components can run as separate virtual machines for scalability and performance
SAN can offload storage
Pros:
PI Server may be isolated from any virtualization guests; performance of PI Server is not directly impacted
Reduced hardware costs to two servers
Increased fault tolerance through multiple virtualization hosts (additional software or licensing may be required for automatic fault tolerance)
Cons:
Additional hardware required for additional virtualization hosts
Scalability:
The design allows for the addition of PI System components to the virtualization host without the possible need for additional hardware. If there are resource constraints for memory or CPU, these components may be added or upgraded on the server to allow for additional scalability on servers. Utilizing two or more virtual hosts, an exponential quantity of PI System components may be installed on the hosts.
Reliability:
Using individual virtual guests for each PI System component, process isolation is achieved. The process isolation will prevent the failure of one virtual guest from effecting the other virtualization guests. In addition to isolation, the multiple enterprise class hosts allow for the migration of virtual machines from one host to another. This process may be done manually or automatically in the event of a problem; installed products and licensing will vary the capabilities.
Efficiency:
Although a minimum of two servers is required for this design, the overall hardware requirements may be reduced. Reducing any hardware requirements may drastically reduce the hardware expenditures needed for the PI System implementation.
The reduced cost and increased fault tolerance provide a beneficial implementation

36. Benefits of PI in a Virtualization Project
Value of HA PI—
Availability, Quality of Service (QoS)
No data loss
Scaling
Improved IT management
Value of virtual machines and SAN
Consolidation
Scaling and hardware utilization
Centralized IT management
All adds up to higher quality of service for less cost

37. Benefits: PI in a Virtualization Project
PI works as well in a virtual environment as it does on physical hardware
PI is perfect for monitoring a virtualized environment
If you are thinking about virtualization, it’s a good time to consider the value of HA PI
If you are thinking about network storage, it’s a good time to consider the value of virtualization and PI with SAN support
If you are thinking about problems with client software deployment, it’s a good time to consider the value of Terminal Services Gateway, virtualization and PI

38. More Information
Whitepapers and Tech Support bulletins on OSIsoft web site
Vendor web sites
OSIsoft internal expertise
Microsoft representatives for Hyper V and Terminal Server Gateway solutions

39. Next Steps Knowledge Base article #3062OSI8
Learn whether there are plans for (or an existing) virtualization environment in your organization
Estimate the hardware reduction to be gained by virtualizing your existing server applications
Estimate the hardware reduction for server applications both with and without a SAN available (more hosted servers per host if data storage is offloaded, for example).
Estimate the hardware, software and support reduction to be gained by moving your client applications to a hosted environment (e.g., Terminal Server)
Consider the value of monitoring the virtualized environment with PI
3062OSI8 - Does OSIsoft support a PI Server running on a virtual machine (such as VMWare or Microsoft Virtual Server)?
Deployment recommendations are the 5 principles and multiple hosts on previous slides
Provision like real machines
Invest in enterprise level technology
Do not mix virtual and real applications/services on same machine
Use virtualization professionals to deploy and manage
Test on the virtualized environment

40. Thank you for your time
your questions to:
Contact your salesperson for more detailed information

42. Virtualization and HA Details

43. Getting Started No 3rd party software or special hardware for HA PI
Licensing models
One virtual PI server = one real PI server
One virtual PI client connection = one concurrent PI client connection
Virtualization hardware/software
Maintenance
Additional storage, memory, etc., as needed over time
Virtualization Per user (from IDC)
Basic virtualization: $24.1 (over 3 years); benefit = $144.9
Advanced virtualization: $23.3 (over 3 years); benefit = $212.4
From IDC white paper:
Basic Virtualization
Total benefits $144.9
Total investment $24.1
Advanced Virtualization
Total benefits $212.4
Total investment $23.3
From Dell white paper ( :
Costs
The key cost categories associated with the Dell Solution total $545,684 and are as follows:
• IT labor costs to plan and test the Dell Solution: $3,000. The Organization’s IT staff
(Windows administrators, storage engineers, and network engineers) spent 60 hours
(in aggregate) planning and testing the Dell Solution.
• Eight Dell PowerEdge (virtual hosts) servers for testing and development:
$56,000. These servers are being used for consolidating old servers in the testing and
development environment.
• VMware VIN (Virtual Infrastructure Node) licenses for testing and development:
$180,000. This software includes the following for the 30 physical servers: ESX,
VMotion, Virtual Center in the licensing package, and three years (2006, 2007, and
2008) of Gold Support (all paid in 2006).
• VMware VIN Gold Support for fourth year (2009): $31,500. (see bullet above for
software details).

44. When is Virtualization NOT a good solution?
No one in the organization is familiar with managing virtual environments
The project is geographically dispersed to the point where there is no benefit of having multiple virtual machines on a single host
All the equipment being used is identical and it is easier to just clone machines than to manage a new virtual environment
Cost to start may be prohibitive
Performance notes:
Settings and parameters may be important for performance
Sizing the virtual machine appropriately makes a difference
Sizing the virtual host appropriately makes a difference
For example, a project in the middle east for an oil refinery found it easier to just clone existing equipment and replace it if needed. The scale of the implementation made virtualization more complex than simple cloning and replacing.
One known issue with HyperV and EMC SANs:
OSIsoft had to resolve an issue with slow performance for HyperV virtual machines (VMs) hosted on our EMC SAN. 
With Microsoft’s support, we were able to find and implement a solution that returned our SAN based VMs to the level of performance we expect.  We were not experiencing any performance  issues with VMs running from server based disks, the issues were only with configurations running from the SAN. 
The Microsoft KB referenced by Microsoft is
Once I disabled Receive-Side Scaling State (RSS), the responsiveness of the SAN based VMs was immediately improved.  Microsoft suggested making sure TCP Offload State was also disabled. 
On our Dell servers, with Server 2008 and the Hyper-V role installed, the RSS setting was enabled by default.  On our installations, the TCP offload was already disabled.
This information may be of value to others using Hyper-V …… changing the setting has made our installations substantially more responsive.
C:\Users\bbarnett>netsh int tcp show global
Querying active state...
TCP Global Parameters
Receive-Side Scaling State          : enabled
Chimney Offload State               : disabled
Receive Window Auto-Tuning Level    : normal
Add-On Congestion Control Provider  : ctcp
ECN Capability                      : disabled
RFC 1323 Timestamps                 : disabled
C:\Users\bbarnett>netsh int tcp set global rss=disabled
Ok.
Receive-Side Scaling State          : disabled

45. Details of Server Virtualization
Available virtual technologies (partial list)
Microsoft (Hyper V, in particular)
VMware (ESX server, in particular)
For more information on virtual machine technologies:
Microsoft Hyper-V:
VMware:
On booting to a virtual machine and VM versus dual boot:

46. Virtual Vendors, Compared
VMware ESX Server
Microsoft Hyper-V
Support for 32/64 bit hosts
Yes
Only 64 bit hosts
Support for 32/64 bit guests
Device Driver Support
Hypervisor
Guest OS
Maximum RAM
64GB
Maximum Virtual CPUs 4
Shared VM Memory No
Boot VMs from SAN
Live Migration
Maximum Active VMs
128
Unlimited
Additional information about the features of each platform can be found at:
VMware:
Microsoft Hyper-V:
Microsoft Hyper-V Planning and Deployment Guide
VMWare Infrastructure 3 Primer
Windows Virtualization Team Blog (information about Import/Export/Copy/Backup features in Windows 2008 R2)
Live Migration and High Availability of Hyper-V Server 2008 R2 (
Microsoft Hyper-V Server 2008 R2 includes Live Migration and High Availability. Period. No Strings Attached.
Live Migration is a great solution for planned downtime such as servicing the underlying hardware like adding more memory, storage or applying a BIOS update. Simply Live Migrate the virtual machines to another server (without downtime) shutdown the physical server and perform the maintenance. When the maintenance is complete, Live Migrate the virtual machines back and your done.
High Availability is a great solution for unplanned downtime. For example, suppose someone accidentally unplugs the wrong power cable on a server. The virtual machine on the server that just unexpectedly went down will automatically restart on another node without any user intervention.
Hyper-V Server 2008 R2 includes both these capabilities as well as our new Cluster Shared Volumes (CSV) capabilities to simplify storage management and run multiple virtual machines from a single LUN.
Hyper-V Server 2008 R2 Live Migration and High Availability can be managed in a few different ways:
Failover Cluster Manager/Hyper-V Manager from a Windows Server 2008 R2 Server OR,
System Center Virtual Machine Manager 2008 R2 OR,
.using the FREELY (there's that word again) available Failover Cluster Manager/Hyper-V Manager for Windows 7. So, as you can see, there are a few different options depending on your needs and option three give you Live Migration and High Availability at zero cost.
BTW: If you decide to go with option #2 System Center Virtual Machine Manager 2008 R2, you certainly can do a lot more such as:
Heterogeneous Virtualization Management
Rich PowerShell Support for Datacenter Automation
Maintenance mode
Virtual Machine Library Support
Templates, Clones, Sysprep Integration
Performance Resource Optimization (PRO)

47. Microsoft’s Hyper-V R2
Live Migration--increased reliability/availability
Clustered Shared Volumes
Cluster Node Connectivity Fault Tolerance. Dynamic I/O redirection.
Improved management of VMs. Including multiple servers.
Improve Performance
VM Chimney/TCP Offload
Note--Intel now builds their chips with features to improve VM performance. (VM Chimney is an example).
Green
Core Parking
Simplified deployment.
.vhd files now the de facto standard.
Server 2008 R2 can boot from a vhd on local hard drive

48. Microsoft Virtualization Options and Licensing
From
* Each additional Windows guest VM requires a license.
If you need to acquire and host new server licenses, Windows Server 2008 Standard, Enterprise, and Datacenter provide the best value.

49. Details of Client Virtualization
Available Client virtualization examples
Terminal Server Gateway (Windows Server 2008)
Terminal Server 2003
Citrix
For more information on client virtual services:
Terminal Server Gateway: and (Terminal Services Team Blog) and (the App-V Team Blog)
Citrix:

50. Details of SAN
Sample SAN vendors (compatible with virtual technologies)
NetApp HP
EMC2
For more information on SAN vendors and products:
NetApp:
EMC: