Cloud Computing and Virtualization

Virtualization abstracts physical machines into flexible, isolated virtual resources, and cloud computing builds on it to deliver elastic, on-demand computing as a service.

Definition

Virtualization creates abstract, isolated instances of computing resources—whole machines or process environments—decoupled from specific hardware; cloud computing uses virtualization to provide measured, on-demand, elastically scalable computing and storage to many tenants over a network.

Scope

This topic covers machine virtualization (hypervisors and virtual machines) and operating-system-level virtualization (containers); the cloud service and deployment models (IaaS, PaaS, SaaS; public, private, hybrid) as standardized by NIST; elasticity, multi-tenancy, and resource isolation; and the cluster managers and orchestration systems that schedule virtualized workloads across data centers.

Core questions

• How do hypervisors and containers provide isolation and resource sharing, and how do they differ?
• What do the IaaS, PaaS, and SaaS service models offer, and to whom?
• How are virtualized workloads scheduled and isolated across a data center at scale?

Key theories

Machine virtualization

A hypervisor multiplexes physical hardware among multiple virtual machines with strong isolation; paravirtualization, as in Xen, achieves near-native performance by presenting a modified hardware interface to guest operating systems.

Cloud service and deployment models

The NIST framework defines cloud computing through five essential characteristics, three service models (infrastructure, platform, and software as a service), and four deployment models, giving a shared vocabulary for the field.

Cluster management and orchestration

Large-scale cluster managers pack diverse virtualized and containerized workloads onto shared machines to drive utilization while honoring isolation and priority, as exemplified by systems like Borg.

Clinical relevance

Virtualization and orchestration are the operational foundation of every public and private cloud; they determine how efficiently data-center hardware is used, how workloads are isolated from one another, and how services scale elastically with demand.

History

Hardware virtualization, pioneered on mainframes, was revived for commodity servers by Xen and VMware in the early 2000s; the NIST definition (2011) standardized cloud terminology, container technology and cluster managers such as Borg then drove the shift to fine-grained, orchestrated deployment that defines modern cloud operations.

Debates

Virtual machines versus containers

Virtual machines provide strong hardware-level isolation at higher overhead, while containers share the host kernel for lightweight, fast startup with weaker isolation; the trade-off between isolation strength and efficiency drives ongoing choices in deployment.

Key figures

• Paul Barham
• John Wilkes
• Michael Armbrust
• Peter Mell

Seminal works

• armbrust2010
• barham2003
• mell2011

Frequently asked questions

What is the difference between a virtual machine and a container?

A virtual machine virtualizes an entire computer, running its own guest operating system on a hypervisor, giving strong isolation. A container virtualizes at the operating-system level, sharing the host kernel, which makes it far lighter and faster to start but with weaker isolation between tenants.