Survey On Resource Scheduling In Cloud Computing

Survey On Resource Scheduling In Cloud Computing

Survey On Resource Scheduling In Cloud Computing

Discuss about the Survey on Resource Scheduling in Cloud Computing.



Currently, cloud computing has received a substantial attention, as an assuring approach for delivering ICT (information and communication technologies) services. With rapid development of processing, internet sensation, storage technologies and computing services have become more powerful, cheaper and more globally available than ever [1]. The resources such as storage and CPU are offered as general utilities to the users connected to the internet. The impact that cloud computing technology has had on the IT industry is immense that the providers like Google, IBM, Amazon, and Microsoft are working to ensure that they deliver a reliable and cheaper cloud-based solutions to better serve their clients. Undoubtedly, cloud computing offers many fascinating features that attract business owners such as minimizing operating expenses, easy access, less upfront investment, highly scalable, minimizing maintenance costs and business risks, among others. The context will discuss service offered by cloud environment, security concerns in cloud environment, data issues, and issues of performance in cloud environment.

PaaS (platform as a service): PaaS is an application deployment and development platform offered over the Web as a service to developers. Equipment of a middleman can be utilized to build programs and send it to the consumers through servers and internet. The complexity and cost of deployment and development of applications can be minimized extensively by developers by utilizing PaaS [2]. Therefore, developers can minimize the expense of purchasing and decrease the complexity of managing the needed infrastructure. The platform comprises of a database, development tools, infrastructure software and middleware.

IaaS (infrastructure as a service): deliver related services both hardware and software that are related and interdependent. Hardware include storage, server and network, and related software are virtualization technology, operating system and file system.

SaaS (software as a service): SaaS reallocate programs to multiple users thoroughly via the browser. Users are able to reduce some cost on servers and software. Service providers, on the other hand, need only to manage a single program, therefore, cost and space is saved [3]. SaaS provider naturally manages and host a particular application in their own and makes it available to many users and tenants using the Web.

Availability of service: considering that many firms are cautious of cloud computing and there are adequate computing services, all available products of SaaS have a high standard. High accessibility from cloud platforms is expected by users. It is very appealing for many customers with business-continuity plan to migrate to cloud computing in vital situations [4]. They can accept various firms to offer them independent software stacks, but it is not easy for one firm to maintain, justify and develop more than one stack in pursuance of software dependability.

DDoS (Distributed Denial of Service) attack is another availability problem. Attackers utilize large botnet to minimize the SaaS provider’s profits by rendering service unavailable. It is quite a challenge to track and manage a botnet attack, but it can be easily identified if it lasts for a longer time.

Data security: this is a critical opportunity for the cloud providers to research into because they are not authorized to access physical data storage facilities. To acquire data security service providers should rely in providers of the infrastructure [5]. The service provider can only point out the specific security context in a virtual private cloud environment, and it is not clear how those are fully deployed. During that time, the infrastructure must consider the following goals: secure data transfer and access, application security interruption, auditability, and confidentiality.

Cryptographic tools are used to achieve the aspect of confidentiality while remote authentication methods are used to attain auditability objective. Remote authentication usually needs a TPM (trusted platform module) to create systems summary that are non-forgeable as the system security proof. In virtual environment, virtual device can move from one location to another. It is challenging to develop mechanism of trust in each structural cloud layer [6]. Migration of virtual machine should only occur if both the destination and source servers are dependable.

Data Lock-In: stacks of software have the ability to use and exchange data between channels, but customers finds it challenging to extract their data and programs from one location and manage it on the other. Some firms are concerned about acquiring data from cloud because they have not fully embraced cloud computing [7]. Users of cloud computing fear such things as consistency issues, increase in price and the probability of providers leaving the business. Developers of SaaS could deploy data and services on many cloud computing providers to avoid the customer data from being affected in case one company fails. The cloud pricing problem can be addressed by convincing users that quality is more important than price and justifying that it concerns data lock-in.

Bottlenecks of data Transfer: Transfer of applications across cloud boundaries may complicate data transport and placement. Cloud users and providers should reduce cost on the traffic concept and placement implications at every level of system. One method of overcoming the high bandwidth transfers cost is to transfer disks. If data is already in the cloud, it may not cause hindrance in enabling a new service. If data documents are in the cloud, it is possible to sell cycles of cloud computing with the new services [8]. Another opportunity is reducing the cost on network bandwidth faster. It is approximated that the cost of fiber is one-third while two-third of WAN bandwidth cost is the high-end routers’ cost. Some analyst are investigating lower expense with centralized control routers rather than the high-end distributed routers. If the technology was implemented in WAN, then the costs of Wan would drop quickly.

Survey On Resource Scheduling In Cloud Computing