7.2.1. Definition

Virtualization includes a range of technologies offering IT resources that can be used flexibly and on demand, independently of the underlying hardware infrastructure. It can be seen as a layer that separates the operating system from the hardware so that the various resources can operate more efficiently. This layer helps to create an environment suited to specifications instead of adapting specifications to fit within the limitations of the hardware environment.

A piece of IT hardware is initially designed to make a single operating system and several applications function. Virtualization makes it possible to bypass this limitation and run multiple operating systems on the same hardware support, thus optimizing the use of hardware resources.

Virtualization can affect the entirety of the hardware environment, that is, the processor, memory, hard disk, peripheral devices and network components, so that each virtual machine has its own virtual resources. Consequently, every virtual machine is completely independent and will not be disturbed or disrupted by the functioning of other virtual machines (shutdown, startup, installation of programs, etc.) that may coexist on the same physical machine.

7.2.2. Benefits of virtualization

Virtualization technology offers a number of benefits:

– pooling of resources: the setting up of multiple logically separated systems on a single physical machine optimizes the use of these systems;

– ease of configuration, deployment, upgrading and migration of virtual machines;

– ability to test a new system without disrupting the base operating system;

– reduction of operating energy consumption: using less hardware results in lower electricity consumption and fewer upkeep and maintenance operations;

– improved security: network security is increased because guest operating systems are not visible to attackers;

– improved availability: a cluster can be formed of two physical servers with virtual machines, which can be automatically copied from one server to another in the event of problems. Moreover, the snapshot or instant capture function of a virtual server makes it possible easily to reinstall a server configuration;

– improved prevention and management of malfunctions as well as system activity resumption thanks to automatic recovery. Virtual servers can be simply restored in the event of problems due to their portability, since the physical server sees them as a series of files;

– scalability via the rapid deployment of new application servers;

– the supervision function is greatly simplified and centralized; a single monitoring software program supervises all virtual servers.

7.2.3. Areas of application

7.2.3.2. Virtualization of servers

The principle here is to make multiple virtual servers function on a single physical server in order to optimize the use of its capacity. Studies have shown that servers use an average of 15% of their capacity. This technique enables a business to save in terms of physical infrastructures, specifically with regard to space occupied, electricity consumption, climate control and staff needed for administration.

7.2.4. Categories of virtualization

Virtualization, a way to supply a number of IT resources independent of the hardware platform, can be broken down into four categories according to Figure 7.2.

7.2.4.1, Virtualization by isolation

This category can pertain to applications or processes and is based on confining the execution of applications to well-defined contexts using a specific software program, making it possible to run several instances of the same application. Open VZ provides an example of process virtualization.

7.2.4.2. Paravirtualization or type 1 hypervisor

A hypervisor functions directly on hardware, thus avoiding using a complete host operating system and limiting itself to the basic kernel services of common operating systems, in this case the management of virtual machine memory and scheduling. In this situation, the guest operating system is modified and acknowledges that it has been virtualized and then relies on the hypervisor to access the hardware resources of the physical server. This category of virtualization shows high performance but is costly. Examples include VMsphere(ESXi), KVM and Microsoft Hyper V.

7.2.4.3. Complete virtualization

This type of virtualization is also called a type 2 hypervisor. It is based on an emulator running on a classic operating system used to manage the memory and scheduling of virtual machines. The hardware is completely simulated. Other operating systems can be installed; these will think they are communicating directly with the hardware. This type of virtualization enables multiple systems to cohabitate on a single physical machine while remaining partitioned from one another. However, it is not very high-performing. Guest operating systems are not aware that they are virtualized. Examples of this approach include VirtualBox, Oracle VM and HyperV.

7.2.5. Limits of virtualization

Despite the significant benefits of virtualization technology, there are some disadvantages.

The physical machine or server hosting various virtual machines must be extremely powerful and equipped with state-of-the-art technologies. It must possess several processors, consistent memory and high storage capacity. These characteristics are, of course, very costly.

The major issue with virtualization is the possibility of the physical server encountering problems or breaking down. If this happens, the virtual machines become unavailable.

Performance may be reduced by the use of virtualization, which can hinder applications from running, as virtualization itself takes up around 30% of the machine load. However, this drop in performance will have more or less of an effect depending on the type of virtualization adopted.

A security breach in the hypervisor exposes all virtual machines to security risks. Virtual systems become preferred targets for attackers, especially if they are not regularly updated and patched.

7.3.1. Definitions

Wikipedia defines cloud computing as a concept that consists of siting IT processes on remote servers that are traditionally executed on local servers or at a user workstation.

NIST (National Institute for Standards and Technology) defines cloud computing as access via a telecommunications network, on demand and by self-service, to shared, virtualized IT resources. It is a relocation of a business's IT infrastructure.

Cloud computing represents an economic IT model that enables the use of software and hardware resources involving servers that are not owned by a company via the Internet, generally by payment-on-demand and according to the company's needs. The goal is to share IT resources including computational power, storage space and network functionalities.

7.3.2. Leverage factors and generic principles

Several factors have contributed greatly to the emergence and success of the cloud computing paradigm.

Successive developments in IT systems, such as the increasing of computational power with new processors, the meteorically swift development of Web technologies, the opening up of companies to the Internet and the current trend toward decentralization have led logically to cloud computing. The current environment, which is highly competitive and based on cost reduction, the appearance of new, increasingly sophisticated terminals, etc., provides an ideal context for cloud computing.

Cloud computing represents a method of organizing and relocating the components that make up an IT system. It is based mainly on the concept of virtualization. The end user is able to execute applications according to his or her own needs without knowing where the servers or the data used are located.

Cloud computing is characterized by the following features:

– continuous and up-to-date availability of services;

– bespoke operation, with payment made according to consumption;

– dynamic allocation of capacity according to load for better management of traffic peaks and sharing of resources.

7.3.3. Architecture models

Cloud computing is generally broken down into three main service models that differ according to the degree of outsourcing adopted by the company's IT system. These three models illustrate three levels of virtualization and can also be considered delivery models.

The objective of the cloud concept is to set up a service between a client and a service provider while ensuring the continuity and quality of this service. To achieve this goal, methods in all levels involved have been developed, including:

– the application, used by the client;

– the platform, or the environment that executes the application;

– the infrastructure, which constitutes the platform support.

7.3.3.3. SaaS (Software as a Service)

Business applications are transferred to the cloud provider, meaning that all necessary software programs are functional without any prior installation or updating, and require no maintenance on the part of the client. This model supplies the highest degree of outsourcing and is based mainly on the principle of standardization of hosted applications. SaaS platforms are based on SOA (Service Oriented Architecture); examples include Gmail and Office 365.

The cloud is a solution that provides space for virtual placement of server or network infrastructures, development or execution platforms, applications, etc. A cloud supports different layers of the service model, from infrastructure to the end client.

To further explain cloud service models, an analogy called “Pizza as a Service” is often used as a tangible visual illustration.

7.3.4. Types of cloud

Two classification criteria are used to categorize the types of cloud:

–Criterion one: who manages the cloud, the cloud provider or the company?

– if the company manages the cloud with its own resources, it is an internal cloud,

– if the cloud provider manages the cloud with its own resources, it is an external cloud.

– Criterion two: who benefits from the service provided by the cloud?

- if the cloud is dedicated exclusively to the company's own needs, it is a private cloud,

- if the cloud is open to the public or to other companies, it is a public cloud.

Four typologies have been established:

– internal private cloud: the company sets up its own infrastructure and offers its services for its employees;

– external private cloud: the company operates an infrastructure rented to a cloud provider and hosts its professional services;

– internal public cloud: pertains to cloud providers. Internal services are offered within the company and to external clients;

– external public cloud: a cloud managed by third-party cloud service providers and made available to partners or clients of the company in question.

Another classification of the types of cloud is also widely used, specifically:

– public cloud: a cloud made available to the public at large. It is external to the company, accessible via the Internet and can be shared among multiple companies. An external service provider with corresponding infrastructures is responsible for managing it. The payment method is pay-as-you-go;

– private cloud: takes the form of an internal cloud structure proper to the company;

– community cloud: a particular type of private cloud that is open to partners of the company (service providers, clients, banks, etc.);

– hybrid cloud: combination of a private cloud and a public cloud that provides the benefits of both types.

7.3.5. Areas of application

Cloud computing is used today by both companies and private individuals. It is a technology that has rapidly become widespread and is experiencing great popularity. Several uses are currently proliferating:

– WEB2.0 applications: e-mail, web conferencing, collaborative work, CRM (Customer Relationship Management), development and testing environment;

– intelligence applications;

– simulation;

– grid computing;

– home automation;

– audiovisual media;

– automotive industry;

– industrial control of machines and procedures.

Many manufacturers and publishers have become involved in the field of cloud computing and have thus become cloud providers, such as Sun, HP, IBM, Oracle, SAP, Google, Yahoo and Amazon.

Table 7.1 some cloud technologies grouped according to service delivery model (SaaS, PaaS, IaaS).

7.3.6. Advantages and limitations

Cloud computing has numerous advantages for businesses:

– reduction of costs: usage determines costs. We only pay for what we have consumed. This approach enables a company to reduce its expenditures and to be more competitive. This automatic adaptation to demand is called elasticity;

– ease of deployment and innovation: elasticity enables companies to experiment and easily deploy new approaches at lower cost and with lower risk;

– contribution to Green IT by reducing electricity consumption;

– ease of collaboration: cloud computing enables a company to collaborate more efficiently with its clients and partners thanks to shared workspaces in cloud communities.

However, cloud computing is not without limitations. Some of these limitations include:

– dependence on a reliable, fast Internet connection as well as increased load on networks, which will be more heavily used;

– security issues: enormous quantities of data can be stored in a cloud provider's servers. This data, some of which is confidential, is transferred to the Internet network and can be stolen or hijacked, with potentially disastrous consequences for companies and cloud service providers;

– regulations can impose constraints, as both cloud service providers and businesses must be aware of and comply with the laws in effect, which can vary from one country to another.