1. Virtualisation Assessment & Roadmap
Giovanni Marchetti

2. Agenda Virtualisation Flavours
Infrastructure Maturity & Virtualisation
DLPARs on AMD & Intel
Microsoft Virtualisation
Q&A

3. Definition
Virtualisation is a technique to pool computing resources in a way that masks their physical boundaries to the resource users

4. Virtualisation Flavours
Workload Management on a single host, single o/s
WSRM, ArmTech, Unix WLM, etc …
Multiple virtual machines on same host o/s
Virtual Server, VMWare, Xen, Unix VM, zVM, etc …
Dynamic Logical Partitions (DLPARs)
Windows Virtualisation (200X), several Unix/MF
All of the above + model-based management = dynamic infrastructure
DSI, Autonomic Computing, etc …

5. Dynamic LPARs LPAR = logical partition
Not bound to HW structure (e.g. can use % of CPU)
O/S runs on top, may (or may not) be LPAR-aware
DLPAR can change physical resource allocation dynamically
Requires O/S support (e.g. dynamic add/remove RAM or CPU)
Applications may (or may not) be able to take advantage
Most efficient virtualisation technique

6. Infrastructure Maturity Model

7. Tool Map
WSRM
DSI VS
DLPAR +
WSRM

8. … on commodity hardware
The Road to DLPAR
… on commodity hardware

9. Type-1 VMM (Hypervisor)
Virtualisation Types
Type-2 VMM
Hybrid VMM
Type-1 VMM (Hypervisor)
Guest 1
Guest 2
VMM
Guest 1
Guest 2
Guest 1
Guest 2
Host OS
Host OS
VMM
VMM
Hardware
Hardware
Hardware
Examples: JVM CLR
Examples: Virtual PC & Virtual Server
Examples: Windows Virtualisation
What we have today
What we’re building

10. Tasks of a Virtual Machine Monitor
OS and Apps in a VM don't know that the VMM exists or that they share CPU resources with other VMs
VM0
VM1
VMM isolates Guest SW stacks from one another
...
...
App
App
App
App
App
App
Guest OS0
...
Guest OS1
VMM runs protected from all
Guest software
VM Monitor
VMM presents a virtual platform interface to Guest SW
Platform Hardware

11. Run VMM in Ring-0 as a collection of fault handlers
SW Solution: Guest Ring De-privileging
Run Guest OS above Ring-0 and have privileged instructions generate faults...
Run VMM in Ring-0 as a collection of fault handlers
Top IA Virtualisation Issues:
Ring Aliasing
Non-trapping instructions
Excessive Faulting
Interrupt Virtualisation Issues
CPU state context switching
Addr Space Compression
Complex Software Techniques:
Source guest OS Modifications
Binary guest OS Modifications
VM0
VM1
...
...
App
App
App
App
App
App
Guest OS0
...
Guest OS1
VM Monitor
Platform Hardware

12. Guest SW runs deprivileged in a new operating mode:
Intel® Virtualisation Technology
Guest SW runs deprivileged in a new operating mode:
Apps run deprivileged in ring 3
OS runs deprivileged in ring 0
VMM runs in new mode with full privilege
VM0
VM1
...
...
App
App
App
App
App
App
Guest OS0
...
Guest OS1
VMM preempts execution of Guest SW via new HW-based transition mechanism
VM Monitor
Platform Hardware

13. Operating Modes VMX root operation: VMX non-root operation:
Fully privileged, intended for VM monitor
VMX non-root operation:
Not fully privileged, intended for guest software
Reduces Guest SW privilege w/o relying on rings

14. VM Entry and VM Exit ... VM Entry VM Exit Transition from VMM to Guest
Enters VMX non-root operation Loads Guest state and Exit criteria from VMCS
VMLAUNCH instruction used on initial entry VMRESUME instruction used on subsequent entries
VM Exit
VMEXIT instruction used on transition from Guest to VMM
Enters VMX root operation
Saves Guest state in VMCS
Loads VMM state from VMCS
VM0
VM1
...
...
App
App
App
...
App
App
App
Guest OS0
Guest OS1
VM Monitor
VM Exit
VM Entry
Physical Host Hardware

15. VT-x Operations VMX Non-root . . . Operation IA-32 VMX Root Operation
Ring 0
Ring 3
VM 1
Ring 0
Ring 3
VM 2
Ring 0
Ring 3
VM n
VMX
Non-root
Operation
. . .
VMCS 1
VMCS 2
VMCS n
VM Exit
Ring 0
Ring 3
IA-32
Operation
VMX Root
Operation
VMRESUME
VMLAUNCH
VMXON

16. AMD Virtualisation Technology
Hosts executes VMRUN to start guest stream
Guest runs until it exits back to host
Host resumes after VMRUN
VMCB holds guest state

17. AMD Virtualisation Technology

18. Principal Causes of EXIT (applies to both)
Paging state exits
CPU State-based exits allow function virtualisation
Selective exception and I/O exiting
Controls provided for asynchronous events (e.g. interrupts)
Guest inactivity

19. The Road to DLPAR
… on Windows

20. Windows Virtualisation
Windows virtualisation for servers
Hypervisor-based
Separate, small management partition (parent)
Targeting availability in the Longhorn wave
Definitions:
Parent partition: a partition that manages its children
Child partition: any number of partitions that are started, managed, and shut down by their parent
Virtualisation Stack: The collection of components that runs in the parent partition for VM management

21. Windows Virtualisation for Servers
Some proposed features
32-bit and 64-bit guests, x64-only hosts
Guest multiprocessing
Virtualised devices
WMI management and control API
Save & restore
Snapshotting
CPU and I/O resource controls
Tuning for NUMA
Dynamic resource addition & removal
Live migration

22. The Hypervisor Very thin layer of software No built-in driver model
Highly reliable
Much smaller Trusted Computing Base (TCB)
No built-in driver model
Leverage the large base of Windows drivers
Drivers run in a partition
Will have a well-defined, published interface
Allow others to create support for their OS’s as guests
Hardware virtualisation required

23. Monolithic vs. Microkernelised
Monolithic hypervisor
Simpler than a modern kernel, but still complex
Contains its own drivers model
Microkernelised hypervisor
Simple partitioning
Increase reliability and minimize TCB
No third-party code
Drivers run within guests
VM 1
('Admin')
VM 2
VM 3
VM 2
('Child')
VM 3
('Child')
VM 1 ('Parent')
Virtual-
ization
Stack
Hypervisor
Drivers
Drivers
Drivers
Drivers
Hypervisor
Hardware
Hardware

24. The Hypervisor Thin layer of software running on the hardware
Supports creation of partitions
Each partition is a virtual machine
Each partition has one or more virtual processors
Partitions can own or share hardware resources
Software running in partition is called a guest
Enforces memory access rules
Enforces policy for CPU usage
Virtual processors are scheduled on real processors
Enforces ownership of other devices
Provides simple inter-partition messaging
Messages appear as interrupts
Exposes simple programmatic interface called 'hypercalls'
Parent Partition
(Minimum Footprint Windows)
Hypervisor
Hard Drive
Ethernet NIC
CPU
RAM

25. Virtualisation Infrastructure Driver Hypervisor API & Message Library
Virtualisation Stack
Collection of user-mode & kernel-mode components
Runs within a partition on top of a (minimal) OS
Contains all VM support not in the hypervisor
Interacts with hypervisor
Calls the hypervisor to perform certain actions
Responds to messages from the hypervisor or from other partitions
Creates and manages a group of 'child partitions'
Manages memory for child partitions
Virtualises devices for child partitions
Exposes a management interface
VM Worker Process
VM Service
VM Worker Process
WMI Provider
VM Worker Process
Virtualisation Infrastructure Driver
VMBus
Bus Driver
Hypervisor API & Message Library
Parent Partition
Child Partition 1
Child Partition 2
Hypervisor

26. Device Virtualisation
Provides a method for sharing hardware efficiently
Physical devices are still managed by their device drivers
Definitions
Virtualisation Service Providers (VSPs) & Clients (VSCs)
VSP = provider, VSC = consumer
VSP typically run in a partition that 'owns' a hardware resource
VSP/VSC pair per device type (storage, network, etc.)
May expose bandwidth resource controls
Protocol is specific to device type, but is generally OS-agnostic

27. Device Virtualisation
Physical devices
Managed by traditional driver stacks
Virtualisation service providers (VSPs)
Virtualise a specific class of device (e.g. networking, storage, etc.)
Expose an abstract device interface
Run within the partition that owns the corresponding physical device
Virtualisation service clients (VSCs)
Consume virtualised hardware service
VMBus
Software 'bus' (enumeration, hot plug, etc.)
Enables VSPs and VSCs to communicate efficiently
Uses memory sharing and hypervisor IPC messages
Storage VSP
Storage Stack
Storage Stack
Storage VSC
Port Driver
Parent Partition
VMBus
VMBus
Hypervisor
Disk

28. Windows Enlightenments
Modifications to an OS to make it aware that it’s running within a VM
Windows codenamed 'Longhorn' enlightenments
Optimisations in memory manager (MM)
Win32 and kernel API: Am I running on a virtual machine?
Looking at additional enlightenments in the future

29. Some System Requirements
Must support hardware virtualisation
Intel’s Virtualisation Technology
AMD’s 'Pacifica'
We are not planning on supporting any other implementations (e.g. VT-i)
Must support x64 extensions

30. Community Resources Windows Hardware & Driver Central (WHDC)
Technical Communities
Non-Microsoft Community Sites
Microsoft Public Newsgroups
Technical Chats and Webcasts
Microsoft Blogs

31. Additional Resources Email: msvirtex @ microsoft.com Web Resources