Skip to main content

23. Virtualisation

09/12/22

What is Virtualisation​

Basic Idea​

  • Abstract hardware of a single computer into several different execution environments.
  • Similar to layered approach, but layer creates virtual system (virtual machine) on which operation systems or applications can run several components
  • A failure in a particular VM does not result in bringing down any others.
  • Virtual machine visualises an entire physical machine
    • Providing the illusion that software has full control over the hardware
    • As implication, you can run multiple instances of the as OS on the same machine.

Components​

  • Host - underlying hardware system
  • Virtual machine manager (VMM) or hypervisor - creates and runs virtual machines by providing an interface that is identical to the host
  • Guest - process provided with virtual copy of the host

Main properties​

  • Isolation - each VM is independent, so failures do not affect the host
  • Encapsulation - state can be captured into a file
    • Check pointing, migration. It is easier than migrating processes. Just move the memory image that contains OS tables
  • Interposition - All guest actions go through the monitor which can inspect, modify, deny operations
  • Fewer physical machines save money on hardware and electricity

Requirements for Virtualisation​

  • Safety - the hypervisor should have full control of the virtualised resources
  • Fidelity - the behaviour of a program on a VM should be identical to that of the same program running on bare hardware.
  • Efficiency - much of the code in the VM should run without intervention by the hypervisor (overheads). For instance with VMware:
    • CPU intensive apps: 2-10% overhead
    • I/O intensive apps: 25-60% overhead

Approaches to Virtualisation​

  • Java VM is very different from the physical machine that runs it

  • Like a simulator or emulator that you can run old Nintendo games on

  • Full Virtualisation - Tries to trick the guest into believing that it has the entire system

  • Paravirtualisation - VM does not simulate hardware. It offers a set of hypercalls which allows the guest to send explicitly requests to the hypervisor (as a system call offers kernel services to applications)

  • Process-level virtualisation - The aim is to simply allow a process that has written for a different OS to run. Wine in linux to run windows applications; Cygwin to run linux shell on windows

Types of hypervisors​

  • Natives (Type 1) - Like an OS, since it is the only programming running in the most privileged mode. Its job is to support multiple copies of the actual hardware
  • Hosted (Type 2) - Relies on a OS to allocate a schedule resources, very much like a regular process

VMMs (Virtual Machine Management)​

Native Virtual Machines (Type 1 Hypervisors)​

  • Hypervisor installs directly on hardware
  • Hypervisor is the real kernel
  • OS runs in user mode
    • Virtual kernel mode
    • Privileged instructions need to be processed by the Hypervisor
    • Hardware VT technology will be necessary
  • Acknowledged as preferred architecture for high-end servers
  • Paravirutalised-based VMs are typically based on type 1 hypervisors

Hosted Virtual Machines (Type 2 Hypervisors)​

  • Installs and runs VMs as an application on an existing OS
  • Relies on host scheduling. May not be suitable for intensive VM workloads
  • I/O path is slow because it required world switch
  • Process-level virtualisation will reply on type 2 hypervisors. Needs an OS.

What needs to be​

Hypervisor must virtualise:

  • Privileged instructions
    • Not safe to let guest kernel run in kernel mode. VM needs two modes; virtual user modes and virtual kernel mode. Both run in real user mode
    • When guest OS executes an instruction that is allowed only when the CPU really is in kernel mode(map virtual pages to physical pages):
      • Type-1 hypervisors: In CPU without Virtual Technology the instruction fails, and the OS crashes
      • Type-2 hypervisors could work without VT - privileged instructions are emulated (binary translation)
    • Trap-and-emulate - When attempting a privileged instruction in user mode causes an error(trap). VMM gains control and executes it. Returns control to guest in user mode
  • CPU
    • Need to multiplex VMs on CPU. Can do this with:
      • Time-slice - Each VM will time slice its OS/Applications during its quantum
    • For type-1 - Need a relatively simple scheduler
  • Memory
    • The hypervisor partitions memory among VMs; Assigns hardware pages to VM, Needs to control mappings for isolation
    • In each VM, the OS creates and manages page tables. But these tables are not used by the MMU
    • For each VM, the hypervisor creates a shadow page table that maps virtual pages used by the VM onto the actual pages the hypervisor gave it
  • I/O devices
    • Guest OSs cannot interact directly with I/O devices - but the guest OS thinks it 'owns' the device
    • VMM receives interrupts and exceptions
    • Type 1 hypervisors run the drivers
    • Type 2 hypervisors - Driver knows about hypervisor/VMM and cooperates to pass the buck to a real device driver on the underlying host OS