Question

1. Explain the job of the hypervisor. Include how it is similar and how it is different between Type 1 and Type 2. Explain how the hypervisor helps maintain Popek & Goldbergs properties for virtualization. (Min: 500 words)

Answer 1

Hypervisor:

A hypervisor could be a hardware virtualization technique that enables multiple guest operational systems (OS) to run on one host system at constant time. The guest OS shares the hardware of the host laptop, specified every OS seems to possess its own processor, memory and different hardware resources.

The term hypervisor was initially coined in 1956 by IBM to talk to computer code programs distributed with IBM RPQ for the IBM 360/65. The hypervisor program put in on the pc allowed the sharing of its memory.

The hypervisor put in on the server hardware controls the guest software running on the host machine. Its main job is to cater to the wants of the guest software and effectively manage it specified the instances of multiple operational systems don't interrupt each other.
Different between Type1 and Type2:

The main difference between Type 1 ANd Type 2 hypervisors is that Type 1 runs on vacant metal and Type 2 runs on high of an OS. Each hypervisor type also has its own professionals and cons and specific use cases. Virtualization works by abstracting physical hardware and devices from the applications running thereon hardware.

Similarities between Type1 and Type2:

Both sort one and kind pair of hypervisors use hardware acceleration support however to variable degrees. sort one hypervisors deem hardware acceleration technologies and usually do not operate while not those technologies accessible and enabled through the system's BIOS. sort a pair of hypervisors area unit usually capable of using hardware acceleration technologies if those hardware options area unit accessible, however, they'll usually fall back on computer code emulation if native hardware support is not accessible on the pc.

The properties of Popek and Goldbergs:
The Popek and Goldberg criteria area unit a collection of conditions decent for a laptop design to network virtualization with efficiency. They were introduced by Gerald Popek and Henry M. Robert Goldberg. The paper establishes 3 essential characteristics for system computer code to be thought of a VMM:

Equivalence/Fidelity: The computer code running underneath the VMM ought to exhibit a behavior basically clones of that incontestable once running directly on equivalent hardware, exclusion temporal order effects;
Efficiency/Performance: a colossal majority of machine directions should be dead by the hardware while not VMM intervention;
Resource Control/Safety: The VMM should be in complete management of the virtualized resources.

Maintaining the properties of Popek and Goldbergs:

To satisfies the properties of Popek and goldberg we need to follow these three conditions in hypervisors.

Privileged directions

Those that entice if the processor is in user mode and don't trap if it's in system mode (supervisor mode).

Control sensitive directions

Those that plan to amendment the configuration of resources within the system.

Behavior sensitive directions

Those types of directions depending on the configuration of resources (the content of the relocation register or the processor's mode).

The main results of Popek and Goldberg's analysis will then be expressed as follows.

Theorem 1.

• For any conventional third-generation laptop, an effective VMM is also created if the set of sensitive directions for that laptop could be a set of the set of privileged directions.
• Intuitively, the concept states that to make a VMM it's decent that each one direction that might have an effect on the right functioning of the VMM (sensitive instructions) perpetually entice and pass management to the VMM. This guarantees resource management property. Non-privileged directions should instead be dead natively (i.e., efficiently). The holding of the equivalence property conjointly follows.
• This theorem conjointly provides a straightforward technique for implementing a VMM, called trap-and-emulate virtualization, a lot of recently called classic virtualization: as a result of all sensitive directions behave nicely, all the VMM should do is entice and emulate each one of them.
• A connected drawback is that of etymologizing decent conditions for algorithmic virtualization, that is, the conditions underneath that a VMM that may run on a duplicate of itself are often designed. Popek and Goldberg gift the subsequent (sufficient) conditions.

Theorem 2.

• A conventional third-generation computer is recursively virtualizable if it is virtualizable and a VMM with none temporal order dependencies is often created for it.
• Some architectures, just like the non-hardware-assisted x86, don't meet these conditions, in order that they can't be virtualized within the classic method. however architectures will still be totally virtualized (in the x86 case that means at the mainframe and MMU level) by mistreatment totally different techniques like binary translation, that replaces the sensitive directions that don't generate traps,
• which area unit typically known as vital directions. this extra process but makes the VMM less economical in theory.
• But hardware traps have non-negligible performance value still.[citation needed] A well-tuned caching binary translation system might win comparable performance, and it will within the case of x86 binary translation relative to initial generation x86 hardware assist, that just created sensitive directions trappable.
• Effectively this provides a theorem with totally different sufficiency conditions.

Theorem 3.

• A hybrid VMM is also created for any third-generation machine during which the set of user sensitive directions area unit a set of the set of privileged instructions: