Building the Foundation..Server Virtualisation and Management

Julius DaviesDatacenter Technology SpecialistMicrosoft UKJulius.Davies@microsoft.com

Clive WatsonDatacenter Technology SpecialistMicrosoft UKClive.Watson@microsoft.com

Where are we today?

How can we optimise?

HYPER-V

Thin Provisioning

VHDVM

Guest OS needs to see 100GBbut may only consume % of that

With Fixed VHDs, a 100GB VHDwould consume 100GB on SAN

With Dynamic VHDs, the physicalspace consumed is only equal tothat consumed by Guest OS

VHD Performance Whitepaper: Link here

Dynamic StorageFlexible solution for adjusting available VM storage without downtimeUtilises SCSI Controller for Hot-Add and Hot-Remove of VHD/PTDEach VM can have up to 4 SCSI ControllersEach SCSI Controller can have up to 64 disks attached

Hyper-v Networking3 types of network:Private, Internal, External

Private = VM 2 VMInternal = VM 2 VM & VM 2 HostExternal = VM 2 VM, VM 2 HostVM 2 VM across Hosts

Each VM can have up to 12 vNICs8 Synthetic & 4 Legacy (PXE)Each with different VLAN ID

• Teaming Support provided by NIC Vendor• Intel = PROSet, Broadcom = BACS, HP = NCU• Best practice: install/enable Hyper-V, then install networking

utilities

Hyper-V Networking for Clusters

Great guide here: http://technet.microsoft.com/en-us/library/ff428137(WS.10).aspxBest Practice Suggests:

1 Network for Host Management1 Network for Cluster Heartbeat1 Network for Cluster Shared Volumes1 Network for Live Migration1 Network for Virtual Machine TrafficIf using ISCSI: 2 Networks for iSCSI Storage with MPIO

The above numbers represent networks, not ports. You may wish to team certain ports to provide resiliency.

High availability – Clustering

1. 2 Hyper-V R2 Nodes in a Failover Cluster. Each Node has 2 VMs running. VMs are stored on the SAN.

2. Node 1 Fails, and also brings down 2 VMs3. Failover Clustering in Hyper-V R2 ensures that VMs restart

on Node 2 of the Hyper-V Cluster

SAN

Cluster Shared VolumesEnabling multiple nodes to concurrently access a single ‘truly’ shared LUNProvides VM’s complete transparency with respect to which nodes actually own a LUNGuest VMs can be moved without requiring any drive ownership changesNo dismounting and remounting of volumes is required

Cluster Shared Volumes

1. We’ve set up a WS2008 R2 Cluster, and created 4 LUNs on the SAN.

2. We’ve made the LUNs available to the Cluster3. In Failover Clustering MMC, we mark the LUNs as CSV’s.4. Each Node in our Cluster then has a Consistent Namespace

for accessing the LUNs. We can now drop as many VMs on each CSV as we like

SAN

C:\ClusterStorage\Volume1C:\ClusterStorage\Volume2C:\ClusterStorage\Volume3C:\ClusterStorage\Volume4

C:\ClusterStorage\Volume1C:\ClusterStorage\Volume2C:\ClusterStorage\Volume3C:\ClusterStorage\Volume4

Live Migration

1. 2 Hyper-V R2 Nodes in a Failover Cluster. Each Node has 2 VMs running. VMs are stored on the SAN.

2. We decide we’d like to Live Migrate a VM from Node 1 to Node 2.

3. Live Migration in Hyper-V R2 ensures that VMs aremigrated with no downtime.

SAN

Dynamic MemoryAutomatic, dynamic balancing of memory between running VMsUnderstands the needs of the Guest OSAvailable as part of WS2008 R2 SP1 at no cost“on the hardware I was testing with, I saw an increase from 64 VMs

(Windows 7 on Hyper-V R2) to 133 VMs (Windows 7 on Hyper-V R2 SP1), we also ran performance testing against this so this wasn't a case of "let's see how many VMs we can fire up“ Matt Evans, Quest Software

RemoteFXNot a replacement for RDP!

Enhancement to the graphical capabilities of RDP 7.1vGPU (WDDM)

Single GPU for multiple Hyper-V Guests

Host Side RenderingApps run at full-speed on host

Intelligent Screen Capture & Hardware-Based EncodeScreen deltas sent to client based on network/client availability

Bitmap Remoting & Hardware-Based DecodeFull range of client devices – HW and SW manifestations by design

Hyper-V R2 SP1 – SummaryBusiness Continuity - High Availability & Live MigrationHost Scalability - 64 Cores & 1TB RAM VM Scalability - 64 GB RAM & 4vCPUs Per VMDensity – Dynamic Memory included with SP1Power Efficiency - Core Parking & Many Power ImprovementsDynamic Storage - Add/Remove disks with no downtimeThin Provisioned VHDs - Use Less StorageNetworking Improvements - NIC Teaming via NIC Vendor, Jumbo Frames, TCP Offload, VMq, vLANs etc.Familiarity - Based on Windows, managed through Windows and System CenterHardware Optimised – Takes advantage of latest h/w innovations (e.g. SLAT)Huge HCL – http://www.windowsservercatalog.comOS Support - In-lifecycle Windows Server/Clients & Linux (SUSE/RHEL/CENTOS)

How can we better manage?

VMware

Self Service Portal 1.0SQL DATABASELIBRARY SERVER

VMM SERVER

SCVMM 2008 R2 SP1ADMIN CONSOLE

SCVMM 2008 R2 SP1Multi-HypervisorP2V & V2VLive Migration SupportQuick Storage MigrationOpsMgr Integration:

Unlocks PRO CapabilitiesRapid ProvisioningIntelligent PlacementLibrary & Web PortalAD IntegrationGranular ManagementPowerShellMaintenance Mode

Self Service Portal 1.0

DEMO:SCVMM 2008R2 SP1

SCVMM 2012 - key pillars

HA VMM SERVER

UPGRADE

CUSTOM PROPERTIES

POWERSHELL

DEPLOYMENT

HYPER-V BARE METAL

HYPER-V MANAGEMENT

VMWARE MANAGEMENT

XENSERVER MANAGEMENT

CLUSTER CREATION

DYNAMIC OPTIMIZATION

POWER MANAGEMENT

NETWORKMANAGEMENT

MONITORINGINTEGRATION

STORAGEMANAGEMENT

FABRIC

SERVICELIFECYCLE

APP DEPLOYMENT

IMAGE BASED SERVICING

CLOUD

CAPACITY & CAPABILITY

DELEGATION & QUOTA

SELF-SERVICE

APP OWNER USAGE

SERVICES

SCVMM 2012 - interface

SCVMM 2012 in action:System Center Virtual Machine Manager 2012:Fabric Management for the Private Cloud

http://www.msteched.com/2010/Europe/MGT306System Center Virtual Machine Manager 2012:Service Lifecycle Management for the Private Cloud

http://www.msteched.com/2010/Europe/MGT206

How can we provideSELF SERVICE?

VMware

Self Service Portal V2

Datacenter and Line-of-Business Administrators

Datacenter Administra

tor

FocusedSolutions

Flexible Management

LOB Administrat

or

End Users/Consumers

SLA Driven

24x7 InfrastructureManagement & Monitoring

ProcurementSecurity

VMM Self-Service Portal 2.0Step 1- Configuration and Extensibility

Pool Infrastructure Assets in toolkit

Extend Virtual Machine actions through Extensibility UI

Step 2- Onboarding and Infrastructure Request

Onboard Business Unit

Create Infrastructure Request (i.e. request a sandbox)

Step 3- Approval /Provisioning

Verify Asset Availability and Capacity

Assign Assets

Approve Infrastructure Request and Provision

Step 4- Self Service VM Provisioning

Manage Environment

Manage VMs

Access Reports

LegalSales FinanceHR

Infrastructure Service AProduction Environment

Corporate NetworkResources: Network access, storage allocation & quotas, access control

Infrastructure Service BTest/Dev Environment

Devevelopment NetworkResources: Network access, storage allocation & quotas, access control

Shared Resource Pool ofNetworkStorageCompute

Web Front Ends

Service Role

Reporting Servers

Service Role

VMM Self-Service Portal 2.0

Infrastructure

Web Front Ends

Service Role

Reporting Servers

Service Role

DEMO:SELF SERVICE PORTAL V2

Learn More

SYSTEM CENTER: http://www.microsoft.com/systemcenter

HYPER-V: http://www.microsoft.com/hyperv

PRIVATE CLOUD: http://microsoft.com/privatecloud

APPLICATION VIRTUALISATION: http://microsoft.com/appv

© 2008 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.