• Ei tuloksia

2. THEORETICAL BACKROUND

2.2 Cloud computing

2.2.1 Cloud computing concept

Cloud computing can be comprehended as all the services which takes place out of enti-ties own premises and are accessed over the internet [23]. In addition National Institute of Standards and Technology (NIST) defines the cloud computing to be a model in which an universal access is based on the commission and where computing resources can be configured and changed, including storages, servers, networks, services and

ap-plications [24]. Premises in here can be sorted out to mean individual persons’ homes or facilities for some particular company or business. To be more precise cloud compu-ting covers all the activities that takes place over the internet which incorporates the use of devices, services or, more anonymously said the use of resources located on provid-ers web servprovid-ers [25]. Reference to the modern cloud computing can be found from the history. Cloud computing is a combination over the grid computing, parallel computing and distributed computing [26]. The basis of the thesis is to concentrate on the cloud computing resources used by businesses but for the wider audience it can be mentioned that most of ordinary people use cloud based solutions in every day basis. For contact-ing their friends and family or when uscontact-ing web based bankcontact-ing solution. Most of the people never realize that the usage of electronic mail is also a use of cloud computing [25].

Cloud computing is based on large data centers which are maintained by the cloud pro-vider [27]. Within these data centers, vast amounts of physical resources are running simultaneously. These physical resources are then applied by multiple virtual machines (VM’s). Each of these VM’s can represent one ecosystem whereas this ecosystem are the virtualized locations of cloud consumers. Cloud consumer can have one or many of these VM’s and so forth use the cloud as on-demand. Scalability, on-demand resources, resilient computing, recovering from disaster and extensively high performance are the main features for justifying the use of cloud based computing. Major players on cloud computing field are also providing payment methods where you only pay for what you use. [27; 28] For small and medium sized enterprises, this creates a significant asset.

Costs for using the cloud are substantially lower when compared to the technology where computing power is maintained on private servers [25; 29; 30]. All this also has another side; cloud data centers are remarkably sophisticated infrastructures. Orchestra-tion of cloud resources makes the cloud soluOrchestra-tions both, vulnerable for security and reli-ability, yet accessing cloud resources require additional expertise through the lack of standardized interfaces [27; 28; 31].

Cloud computing theory holds three different models for describing the level of services whom cloud providers are offering. Frost & Sullivan addresses these level as the legacy levels, for the reason that new business insights are on the horizon [10]. The lowest lev-el of basic service is Infrastructure as Service (IaaS). IaaS is a service where service provider offers only the physical resources accessed by the consumer. Consumer must deploy their own operating system (OS), data storage methods, software’s and network connections. [3] Platform as Service (PaaS) represents the middle level out of these three service stages. PaaS is a realization of physical resources been submitted for one virtual machine. Used resources can be distributed over various data centers; however, virtual machine acts as one frontend entity. Consumers exploit this platform as one en-vironment for deploying their own services and computational applications. Service consumer can maintain virtual machine(s) via internet browser based portal and through

this portal changes can be made for platform within the limits of service provider. [3;

26] On the highest level locates the Software as Service (SaaS). Method and ideology for using SaaS diverts totally from two mentioned ones. The principle is that SaaS is accessed on-demand bases and through any device, that has an access to internet. Ser-vice consumers possess only minor possibilities to customize the serSer-vice. SaaS acts as an individual entity that is used to change the state of the consumer or provide a new thread for the consumer to continue its actions. [3; 26]

These three layers forms the basis for cloud technology, although multiple variations exist. Singh et al presents a new form of PaaS named Plat Serve derived from Platform as PaaS [26]. Plat Serve stands for the paradigm where all the operating systems are installed on the central server and user only picks up the one, which is required at a time. This gives the advantage over traditional problems for operating system updates.

All the updates are always activated through this one master OS, gaining the user an access to the latest features. [26] Additionally, variations of different services can be combined and illustrated with equal basis. These variations include, among others, Communications as Service (CaaS) [3] and Materialized View as Service (MVaaS) [32]. Consulting company Frost & Sullivan introduces a model of Everything-as-service (XaaS) in their report [10] of new business opportunities in cloud services. Frost & Sul-livan states that new services will emerge and more increasing amount of services will be offered as cloud based.

Dialogic portrays four different cloud location models in their Whitepaper [3]. These four models are also addressed by Duan et al in their article for a Construction Method and Data Migration Strategy for Hybrid Cloud Storage [7]. Private cloud is addressed as cloud located entirely within locations firewall. Private cloud can be maintained either the operator itself or some third –party operator. [3; 7] Private cloud has the advantages for storing vast amount of data with high reliability in terms of availability [7]. Com-munity cloud portrays a cloud deployment model where multiple consumers share the same cloud infrastructure [3]. Usually these consumers have similar requirements for the cloud and so forth the usage of same deployment is conceivable. Consumers may also hold a demand for allocating rather modest amount of financial resources for de-ploying their function at cloud. In these circumstances, a public cloud comes in ques-tion. Public clouds are commercial versions of cloud based computing and can be ac-cessed with rather modest payments for the provider. Financial model in these clouds are based on pay-as-you go type of invoicing. Hybrid cloud is a composition over two or more of these three other explained cloud types. [3; 7] When deploying a hybrid cloud two main issues should be covered. Usage of two different cloud types should be invisible to the end user, and at the same time implementation should hide the com-plexity of the structure behind the multiple cloud system. Gaining these two aspect at the same time is much more challenging than commissioning a one model based cloud system. [7]