Green IT, Green IT, everywhere you turn there is something about Green IT.  With increased power and cooling costs a lot of people are using virtualization to reduce these costs but Windows Server 2008 R2 brings a new feature called Core Parking which helps as well.  It does this by moving all processing to a core (or cores) and putting the other cores to “sleep”.

Image courtesy of Microsoft Corporation

As processing power increases more cores “wake up” and take on the additional processing.  Considering Windows Server 2008 R2 supports up to 256 cores this feature can really provide some savings while not impacting performance.

Further to Core Parking, Windows Server 2008 R2 can also adjust the “P-State” to reduce the power consumption of a core.  In a nutshell this is like Intel SpeedStep which can “slow” down the CPU on your notebook to conserve power when you run on battery.

Image courtesy of Microsoft Corporation

All of this can be controlled via Group Policy in an AD domain or via the Local Policy on a standalone server.

I can’t think of any feature we’ve been asked more in regards to Hyper-V has been around Live Migration, known as VMotion in the VMWare world.  Hyper-V R2 now supports Live Migration, partly due to Cluster Shared Volumes.

Live Migration allows a virtual machine to be moved from one node in a cluster, to another node without any downtime.

Image courtesy of Microsoft Corporation

In a nutshell Live Migration accomplishes this by…

1. Using SCVMM 2008 (an update will be required) you initiate a Live Migration from the source host to another node in the cluster.
2. Once initiated, a virtual machine is created on the target host.  Since the VHD(s) is stored on a CSV there is no copy of a VHD over the network.
3. Once the VM is created on the target host the data stored in the source hosts RAM is copied to the RAM in the target host.  In essence Hyper-V is mirroring the contents of memory on both hosts.
4. Up until the memory “mirror” is complete, clients continue to be connected to the original VM on the source host.
5. Once the memory “mirror” is complete, clients are redirected to the VM on the target host.

Now one thing to remember is that both nodes of the cluster need to be online.  This is not a DR solution, although the VMs can be made highly available.  In a DR situation where the source node fails, the highly available VMs will be restarted on the other node(s) in the cluster.

One of the challenges faced with clustered Hyper-V servers today is that for each VHD on a highly available virtual machine, you need to carve out a LUN on your SAN to place that VHD.  While it might not be a monumental task to do so, it is a management nightmare.  With less than 24 drive letters free you often end up using GUIDs.

Well that changes in Server 2008 R2 with the introduction of Cluster Shared Volumes (CSV).  CSV allows you to place multiple VHDs on a single LUN while still looking like it is on its own LUN. 

Image courtesy of Microsoft Corporation

There are a number of benefits to CSV including…

• provides consistent namespace to all nodes in the cluster
• they still use NTFS meaning deployment will note require you to reformat your SAN
• also using NTFS means you can use a file system you are familiar with rather than s proprietary file system
• no special management tools are required to create and manage the CSV

With all CSV volumes stored in the ClusterStorage root directory, it becomes easy to navigate the file system as well.  As an example if the Cluster Storage folder is stored on the S:\ drive then the path to Volume1 would be S:\ClusterStorage\Volume1 and the path to Volume63 would be S:\ClusterStorage\Volume63.  No running out of drive letters and no GUIDs!

VPNs are a thorn in the side of all admins.  While the setup and management has improved they still lead to a number of calls from users on the road who have difficulties connecting.  There have been a number of technologies along that way that have reduced the need for VPNs, like OWA, RPC over HTTPs, and secure SharePoint publishing but we still have VPNs because we need them.  DirectAccess is a new feature in Windows Server 2008 R2 that aims to solve that problem.

Windows Server 2008 R2, used in conjunction with Windows 7, can eliminate the need for a VPN as it provides a transparent connection to internal resources.

Image courtesy Microsoft Corporation

DirectAcces can provide this secure connection thanks to a few built in features.

• Network Policy Server – Controls what clients have access to the internal network
• Network Access Protection – Ensures clients are up to date before creating the connection
• Access Control – Using AD clients can be given different levels of access to the internal network
• Authentication – DirectAccess handles the authentication allowing the client to connect before the user logs in
• Authentication Support – Provides support for smart cards and other forms of multi-factor authentication
• Encryption – All communications are encrypted to protect the privacy of the data being transmitted
• Routing – DirectAccess leverages “split tunnel” routing allowing only traffic destined for the internal network to pass while directing all other traffic to the WAN connection.

The best part about this is that as an admin, you have full management control of a client connected via DirectAccess.  You can push updates, apply policies and manage the client without the need for a user to initiate a VPN connection or even log on.

Do more with less?  Ever hear that term?  Probably around the same time your budget was cut!  Bandwidth between the home office and the branches can be a rather large cost from the actual money spent on the connection to the user downtime when the connection is slow.  Having worked in an engineering company that was always sending large CAD drawings back and forth, Branch Cache looks to help.

In a nutshell Branch Cache is very much like a DNS cache.  For example, when you do a lookup of thelazyadmin.com via DNS the resulting IP address for the server is stored in your cache until the TTL expires.  Branch Cache does the same thing only with data that you have transferred the remote site.  There are two types of configurations, peer based and server based.

With peer based caching, User 1 copies Document1 from the file server at the head office to their PC at the remote office.  When User2, located at the same remote office attempts to download Document1 from the file server at the head office they are redirected to User1’s computer and will download Document1 from there, saving bandwidth and speeding up the time it takes for the download to complete.

Image courtesy of Microsoft Corporation

A server based Branch Cache configuration works in a similar manner however a server is installed at the remote site and a centralized cache is created there.  This is ideal for larger organizations.

Branch Cache works with both SMB and HTTP content and can be configured manually or via group policy.

There is a lot of VHD integration in Windows 7 and Server 2008 R2.  We already showed how to boot from VHD but you can also mount a VHD from within a Windows 7 / Server 2008 R2 environment.

Open up Disk Management, you can Right-Click on Computer and choose Manage and then select Disk Management to do this.

Right-Click on Disk Management and choose Attach VHD.

Enter the path to the VHD.  You can choose to mark it Read-Only as well. Click OK.

The drive will mount and display in Disk Management.  You’ll notice the disk icon is also blue to help you distinguish it as a mounted VHD.

To dismount, Right-Click the disk and select Detatch VHD.

You will be given the option to delete the VHD as well, but it is not the default.  Click OK to detatch.

For more information be sure to check out the Windows 7 resource page on Technet.com