Brief Analysis of Cloud Computing Elements

Because of the need to be able to quickly add requests when computing resources, people have developed technologies that allow for the provision of resources, but do not need to know specifically what the resources are and where resources are located. This flexible environment, to provide the necessary resources supply technology, called cloud computing. One of the main reasons why we agree with this change in calculation method is the potential to use a large proportion of fast, large-scale resources in smaller, restricted environments.

Due to the widespread adoption of virtual computing, this new computing paradigm is currently commercially viable. Just as virtual machines can build on relatively little consideration of the underlying hardware (in addition to capacity), virtual computing allows resources to be provisioned quickly. (It has been possible to physically allocate the required new resources such as processors physically, at least technically, for many years already provided by the vendors; however, a viable fee has not yet been established for the use of these technologies Mode and acceptable.)

Cloud computing elements

In a cloud computing environment, the level of abstraction and lack of control depends on the particular solution. However, most solutions use the same elements to build the environment without much thought and actually give the user more control. These elements include: Application Programming Interfaces (APIs), networking, processing cycles (CPUs), and storage.

API: API elements, often referred to as client interfaces, provide developers with control over specific abstract elements (such as Java features, storage, etc.).

Network: Network elements are usually not visible to cloud users. However, the network element is the foundation for interconversion of transactions between the public Internet and private cloud resources. Network elements are usually virtualized, allowing the network to be resilient failover while off-line giving network addresses to systems and possibly network routers and switches.

CPU: Depending on the solution provided, the processing cycle may or may not be visible to cloud users. Under some circumstances, it may be necessary to provide resources based on the number of instructions provided by the system and create a corresponding list. At a higher level, these processing cycles are central processing units, or through virtualized partial central processing units.

Storage: Storage is well suited for cloud computing because of its relatively high cost, as well as the difficulty of planning for growth and the corresponding provisioning. Depending on the solution, storage may be available to users. For example, EMC's Mozy Backup solution provides accessible storage for backing up PCs.

Open source cloud computing using Eucalyptus

Eucalyptus cloud computing software, on an open source platform (which is also commercially available), provides an abstraction of these resources. Eucalyptus source is open. There are packages available for CentOS 5.3, OpenSUSE 11, Debian 5, and Ubuntu 9.04.

Eucalyptus chose Xen and KVM as the hypervisor. As of this publication, the released version is 1.5.

Eucalyptus Components: Logically, the Eucalyptus Cloud includes four components: the client, the cloud controller (referred to as cls or "cloud"), the cluster controller (cc), and the node controller (nc). The client depends on the solution, but may come in the form of a browser script, a user space application, or even a kernel module. The cloud controller is the interface that goes into the cloud from the client and provides the cloud with logical decisions. Cloud Control performs the service of authenticating the client and translating the client's request into a transaction. The cluster controller is a collection of node controllers. It is responsible for status information and interaction with the provided virtual machine. However, the nodes themselves interact with online and offline virtual machines through a hypervisor (Xen or KVM). A node controller exists as a stand-alone instance on a separate machine. However, multiple node controllers make up a specific cloud.