06-09-2016, 10:07 AM
1453119183-Introduction2.docx (Size: 749.69 KB / Downloads: 6)
Introduction
Cloud computing also referred as the cloud (due to the internet’s representationin flow diagrams) is an on-demand computing modelwhichconsists of independent, networked hardware and/or softwareresources.Cloud computing applies the concept of virtualization for optimal usage of hardware and/or software resources. In simple terms cloud computing is the virtualization of a pool of resources, under data centres for hosting cloud applications, which are made available to everyone on subscription basis. In cloud computing the hardware and software resources are made available by the providers for different users or clients. Service providers offer cloud services with predefined quality of service (QoS) terms through the Internet as a collection of easy-to use, scalable, and economically feasible services to the clients. The cloud services fall under three categories: Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS).
The characteristics of cloud computing have attracted many IT giants like Amazon, Google, Microsoft, Sharepoint, VMware etc. Amazon is the leader amongst all cloud providers. The two most common services they provide are Amazon S3 a Simple Storage Service and Amazon EC2 Elastic Cloud Computing, both belong to IaaS service model.S3 and EC2 both work on a concept of pay-as-you-go model. Therefore, the number of people using these services are exponentially increasing with the increase in deployment of new applications on the cloud as well.
Though cloud computing has the capability of handling many of users through virtualizationconcurrently,its power consumption and carbon emission has become a major environmental concern.
GREEN CLOUD COMPUTING
Green computing deals with processes of designing, maintaining and disposing of computer devices without doing any harm to the environment. Due to the growing concerns of increase in power consumption and carbon emissions by the IT industry the concept of green computing was realised. Most IT companies realised that taking a step towards green computing would not only reduce the carbon emissions and but would also cut total costs of the system from a business perspective. Moreover, the probability of hardware failures increases due to higher power consumption as the heat dissipation also increases with it. The most commonly used green computing technologies are: Virtualization, Green Cloud Computing, Power Optimization,Green Data Centre and Grid Computing
Green computing can efficiently use virtualization so as to improvethepower efficiency of data centres by assigning the tasks of multiple virtual machines (VMs) to a single server. Server virtualization helps in workload consolidation by processing the tasks and turning off idle physical machines there by lowering the consumption of energy.By using virtualization technology, multiple applications can be hosted and executed on the same server in isolation which can process more tasks with the same power usage.
Another way for green computing to reduce power consumption is through service level agreement SLAs. SLA is an agreement between the service provider and the consumer which takes place before any allocation of resources. The SLA service can be related to storage space, bandwidth, and power consumption.
RELATED WORK
Several researchers have introduced various models and/or methods to conserve energy.
Mr. Louis Rillingetal proposes a virtual infrastructure optimization solution using the ant colony optimization algorithm for finding better paths through graphs. The most common approach while performing workload consolidation is that the workload is allotted to a physical machine(eg CPU) and those resources which require excessive provisioning are converted into a lower power state.
Mr. Osvaldo Adilson de Carvalho Junior etal proposes the use of a function that can ensure the most appropriate behaviour to the principles of GreenIT but not the quality of service. For this he proposes the use of GreenMACC (Metascheduling Green Architecture) and its module LRAM (Local Resource Allocation Manager) to automate the execution of all scheduling policies implemented in the Scheduling Policies Module so as to provide Quality of Service in Cloud Computing and determine its flexibility.
Task consolidation is an efficient method which is used to reduce power consumption by increasing the resource utilization but due to task consolidation resources may still draw power while being in the idle state. Mr. Young Choon Lee etal has introduced two algorithm to maximise the utilization of resources of the cloud. The two algorithms are ECTC and MaxUtil. ECTC works on the premise of calculating the energy which is being used by a particular task when there are simultaneous tasks running parallely with it, and then it is compared with the optimal energy which is required. MaxUtil focuses more on the mean usage of a particular task when it is being processed.
Mr.DzmitryKliazovichetal presents a simulation environment for data centres to improve their utilizaiton of resources. Apart from working on the distribution of the tasks, it also focuses on the energy used by the data centre components. The simulation outcomes are obtained for various architectures of data centres.
Mr. Robert Basmadjianetal proposes the use of proper optimization policies reducing the power usage and increasing the resource utilization without sacrificing the SLAs. He developed a model which worked on incrementing the capability of the processor to process tasks.
Mr Zhou Zhou et al proposes a Three Threashold Energy Saving Algorithm [TESA] which has three threasholds to divide hosts between heavy load , light load & middling load.Then based on TESA 5 VM migration policies are suggested which which significantly improves energy efficiency.
Mr Dung H Plan et al proposes GreenMonster protocol which improves renewable energy consumption while maintaining performance by dynamically moving services across IDCs. GreenMonster uses Evolutionary Multiobjective Optimization Protocol [EMOA] to make service placement and migration decisions.