Microsoft Hyper-V 2.0 Features

Below are my rough session notes from a session of Microsoft Hyper-V features.  These are pretty much bullet point only though they cover the key areas.  The only thing to keep in mind when reading this is that Hyper-V 2.0 will only be available with the release of W2K8 R2 which is slated to be released in Q1 2010.  So a lot of those features which will bring Hyper-V more in line to what ESX is currently offering won’t unfortunately be available until then.  Hyper-V 1.0 is still worth a look however and will no doubt continue to have a more expansive install-base.

Microsoft Hyper-V 2.0:

  • Included with x64 W2K8 editions only.
  • 64GB per VM
  • 1TB per W2K8 server.
  • Hardware Requirements:
  • AMD-V or Intel VT
  • DEP
  • >64 Hard disks per controllers.
  • >12 network adapters per VM (each with own Mac address)

3 types of virtual switches:

  • Private – no packets go out onto the wire.
  • Internal – VM’s talk to themselves and the parent partition.
  • External – goes out to the usual network.

Now with support for systems with 24 LPs and up to 192 running VM’s!

Upon installing the Hyper-V role the Hypervisor is installed and insert to run directly on the hardware. The Hypervisor is run upon the reboot after installing the Hyper-V role.

Traditionally – to access a network packet the VM needs to go via User mode to the Kernel mode to the Hypervisor and back.

Now: In kernel mode. VSC makes request to VMBus that talks to VSP in parent partition that then in turn talks to the VM Drivers (also in the Parent Partition) in the Kernel Mode.

Xen Enabled Linux (eg: SuSe, etc) is not emulated.

Can run x86 and x64 VM’s

Supports up to 4 cores (SMP) per VM

HA provided via clustering.

BitLocker: Seamless, secure data encryption.

Live Backup: Volume Shadow Service Integration. Ensures that applications put them self into a state for backups.

Pass-through disk access for VM’s as opposed to using a VHD: Good for working with large datasets, eg: >300GB

Virtual Machine snapshots

New hardware sharing architecture (VSP/VSC.VMBus)

  • Disk, networking, input, video

Robust networking: VLANs and NLB

Security – Isolation:

  • No sharing of virtualized devices
  • Seperate VMBus instance per VM to the parent
  • No sharing of memory
    • Each has its own address space – so no man-in-the-middle attack risk
  • VMs cannot communicate with each other, except through traditional networking.
  • Guests can’t perform DMA attacks because they’re never mapped to physical devices
  • Guests cannot write to the hypervisor
  • Parent partition cannot write to the hypervisor.

Hyper-V Networking:

Two physical adapters at a minimum

  • 1 x Management
  • 1 x (or more) for VM traffic
  • Dedicated NIC(s) for iSCSI
  • Connect parent to back-end management network

If you want to PXE boot a VM you need to use the ‘legacy’ NIC.

Always ensure that the ‘Synthetic’ device drivers are installed and not the ‘Emulated’ device drivers as these are much slower.

Tip: When creating W2K3 template VM’s ensure that it is created with dual processor HAL as even if running on a single CPU VM it is compatible – but not the other way around.

 

Technorati Tags: ,,,
About Simon Seagrave 706 Articles
Simon is a UK based Virtualization, Cloud & IT Technology Evangelist working as a Senior Technology Consultant and vSpecialist for EMC. He loves working in the ever changing IT industry & spends most of his time working with Virtualization, Cloud & other Enterprise IT based technologies, in particular VMware, EMC and HP products. As well as on this site, you can find him on Twitter and Google+

Be the first to comment

Leave a Reply

Your email address will not be published.


*


This site uses Akismet to reduce spam. Learn how your comment data is processed.