Cloud Computing- Research Issues and Challenges

Transcription

1 Proceedings of the Global Engineering, Science and Technology Conference December 2012, Dhaka, Bangladesh Cloud Computing- Research Issues and Challenges Nazia Majadi Cloud computing uses the internet and central remote servers to maintain data and applications. It allows consumers and businesses to use applications without installation and access their personal files at any computer with internet access. This technology allows for much more efficient computing by centralizing data storage, processing and bandwidth. The appearance of cloud computing has made a tremendous impact on the Information Technology (IT) industry over the past few years. Recently IT industry needs Cloud computing services to provide best opportunities to real world. Cloud computing is in initial stages, with many issues still to be addressed. The objective of this paper is to explore the different research issues and challenges of cloud computing and identify important research opportunities in this increasingly important area. In this paper, there are different challenges categorized under design, performance, and security challenges. Field of Research: Information Technology (IT) I. Introduction According to the National Institute of Standards and Technology (NIST) definition [1] Cloud computing is a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction. The name comes from the use of a cloud-shaped symbol [2] as an abstraction for the complex infrastructure it contains in system diagrams (see Fig. 1). Cloud computing entrusts remote services with a user's data, software and computation. Cloud computing [3] recently received considerable attention, as a promising approach for delivering Information and Communication Technologies (ICT) services. With the fast development of processing, storage technologies, the sensation of the Internet, and computing resources have become cheaper, more powerful and more universally available than ever before. This technological trend has enabled the realization of a new computing model called cloud computing. The traditional cloud computing service provider is divided as the infrastructure providers who supervise cloud platforms and lease resources according to a pay-per-use based model, and service providers, who rent resources from more infrastructure providers to serve the end users. From the past few years, the cloud computing has made a tremendous impact on the Information Technology (IT) industry, where large companies such as Google, IBM, Amazon and Microsoft struggle to provide more powerful, reliable and cost-efficient cloud platforms, and business enterprises seek to find new paradigm in their business models. Nazia Majadi, Lecturer, Military Institute of Science and Technology (MIST), nazia_majadi@yahoo.com,

2 Indeed, cloud computing provides several compelling features that make it attractive to business owners, like less up-front investment, lowering operating cost, highly scalable, easy access, reducing business risks and maintenance expenses [12]. Figure 1: A logical diagram of Cloud Computing [2]. This paper is organized as follows. Section II discusses the motivation of cloud computing. The services provided by cloud environment are given in Section III. Section IV explains the research issues of cloud computing categorized under design, performance and security issues. Section V discusses the research challenges of cloud computing. Finally, Section VI concludes the paper. II. Motivation Cloud computing promises enormous opportunities for organizations to cut costs and speed time to market while eliminating the heavy investment in IT capital and operating expenses traditionally associated with running a business. The range of solutions available today including publically hosted clouds, private internal clouds or a hybrid environment, ensure that the user- whether it be an enterprise, service provider, entrepreneur or developer- can find the best cloud computing solution to meet their needs and quickly gain the advantages that enable them to become more agile [7]. Over the years, Chief Information Officers (CIOs) have been continually challenged with how to enable their organizations to quickly and cost-effectively deliver infrastructure and services to support end-user applications. To initially address this need, innovation in systems, applications and infrastructure were quickly delivered based on new networking capabilities and large network performance. While some argue that cloud computing is simply an evolution in the model of network computing, one that builds on and extends existing frameworks 2

3 and architectures in the datacenter. Cloud computing delivers higher efficiency, limitless scale and faster and easier deployment of new services and systems to the end-user, thereby changing the economics of the datacenter by shifting the delivery of IT resources to an on-demand model and opportunities for both enterprises IT and service providers. The tremendous success of companies like Amazon, ebay, Google, Yahoo! has demonstrated that building out cloud infrastructure produces cost savings an order of magnitude larger than a traditional enterprise can achieve with the same types of commodity hardware and open source software. However, it is when we dive into the difference in models that we start to understand that cloud computing is not just about the market cost of individual components like compute, network and storage, but the sum of those components built up in a smarter, more efficient lightweight manner that provides higher utilization and elasticity over that of traditional datacenter models. While these are all important points to note, it is equally relevant to the discussion to consider that this new model brings not only a wealth of benefit, but also introduces new set of technical challenges. III. Cloud Computing Services Cloud computing services are available across the entire computing spectrum. The basic services of cloud have been considered as the following. Figure 2: Basic three cloud computing services with cloud clients [2]. Software as a service (SaaS): SaaS reassign programs to millions of users all the way through browser. For user, this can save some cost on software and servers. For Service provider s, they only need to maintain one program, this can also saves space and cost. SaaS provider naturally hosts and manages a given application in their own or leased datacenters and makes it available to multiple tenants and users using the Web. 3

4 Platform as a Service (PaaS): PaaS is an application development and deployment platform provided as a service to developers over the Web. Middleman s equipment can be used to develop programs and transfer it to the end users through internet and servers. The cost and complexity of development and deployment of applications can be reduced to a great extent by developers by using this service. Thus the developers can reduce the cost of buying and reduce the complexity of managing the required Infrastructure. It provides all of the services essential to support the complete life cycle of building and delivering web applications and all the services entirely available from the Internet. This platform consists of infrastructure software, a database, middleware, and development tools. Infrastructure as a Service (IaaS): IaaS is the delivery of associated software and hardware as a service. Hardware like server, storage and network, and associated software like operating systems, virtualization technology and file system. It is an evolution of traditional hosting to allow users to provide resources on demand and without require any long term commitment. Different PaaS services, the IaaS provider does very little management other than keep the data center operational and end-users must deploy and manage the software services themselves-just the way they would in their own data center [12]. IV. Research Issues In this section, some of the research issues in Cloud Environment are described. A. Design Issues The issues involved in designing cloud computing are described as follows: (i) Energy Management- Improving energy efficiency is a major issue in cloud computing. It has been estimated that the cost of powering and cooling accounts for 53% of the total operational expenditure of data centers [8]. Infrastructure providers are to reduce energy consumption under enormous pressure. Current research on energy-efficient network protocols and infrastructures [4, 11, 15]. To achieve a good trade-off between energy savings and application performance is key challenge. Some of the researchers [8, 10] have recently worked on the solutions for performance and power management in a dynamic cloud environment. (ii) Software Frameworks- Cloud computing provides a persuasive platform for hosting significant data-intensive applications. In this, Hadoop for scalable and fault tolerant data processing by using applications leverage MapReduce frameworks concept [9]. The design challenges include performance modeling of Hadoop jobs in all the possible cases, and dynamic conditions in adaptive scheduling. Some researchers are still working on a trade-off between performance and energy-awareness [9, 13, 15]. (iii) Novel Cloud Architectures- Large data centers and operated in a centralized fashion is implemented as commercial clouds. In this design achieves high manageability and economy-of-scale. But there are some limitations in large constructing data centers such as high initial investment and high energy expense. Some researchers [6, 14] suggest that small size data centers can be more advantageous than big data centers in many cases: a small data 4

5 center consumes less power. So, it does not require a powerful and high expensive cooling system. (iv) Software Licensing- Existing commercial software licenses usually control on which computer the software can run. Users can pay an annual maintenance fee and initially payment for the software. For cloud computing applications, this existing commercial licensing approach for software is not fine and many cloud computing providers are relying on open source software [5]. The primary solution for commercial software companies is to better fit Cloud Computing by change their licensing structure. The challenge is software companies to sell products into cloud computing by support sales services. (v) Client Incomprehension- We re probably not in the days where people used to think that clouds were just big clusters of servers, but that doesn t mean we re free of ignorant. We are aware of the fact that the cloud is moving forward. There are many misunderstandings about how private and public clouds work together, how virtualization and cloud computing overlap and also how to move from one kind of infrastructure to another and so on. A good way to clear these is to present users/customers with real-time examples of what is possible and why. B. Performance Issues The following performance issues need to be considered in cloud environment: (i) Virtual Machine Migration- Virtual Machine migration provides major benefit in cloud computing through load balance across data centers. It also provides robust and high response in data centers. Avoiding hotspots is major benefit of VM migration even though it is not straight forward. Initiating a migration lacks the facility to respond to unexpected workload changes and detecting workload hotspot. (ii) Server Consolidation- In a cloud computing environment, Server consolidation is an efficient approach is to minimize the energy consumption for makes best use of resource utilization. Server consolidation is not depend on application performance.it is known as the resource usage means individual VMs may vary time to time. (iii) Performance Unpredictability- Sharing I/O is complex in cloud computing while multiple virtual machines can share CPU and main memory easily. Virtualization of I/O interrupts and channel is one solution to improve the efficiency of operating system and improve architecture. (iv) Scalable Storage- Cloud computing important properties are: infinite capacity on-demand, no up-front cost, short-term usage. This is an open research problem. (v) Bugs in Large-Scale Distributed Systems- Another challenge issue in cloud computing is removing errors in these large scale distributed systems. The debugging of these bugs have to be done at large scale in the production data centers as these bugs cannot be reproduced in smaller configurations. 5

6 (vi) Scaling Quickly- Pay-as-you-go certainly applies to network bandwidth and storage on the basis of used bytes count. Depending on the virtualization level, computation is slightly different. Cloud computing providers already get low overhead and careful accounting of resource consumption. By imposing to pay attention to encourages programmers in the concept of efficiency are utility computing, per-hour and per-byte costs, development inefficiencies, and allows more direct measurement of operational. (vii) Latency- Latency is a research issue on the Internet. Any performance in the cloud is going the same meaning of the performance of the result on the client. The latency in a cloud introduces not to be tedious. The latency is compressed back for understand how and where they re running with both smartly-written applications and an intelligently planned infrastructure. In future, cloud computing capacity and cloud based applications are rapidly increases and latency is also increases. Cloud computing latency must be improved in the desktop PC is largest bottlenecks in the memory and storage. C. Security Issues The following security issues need to be addressed in cloud computing: (i) (ii) Availability of Service- Most of the organizations is some cautions of cloud computing are existing in utility computing services. In this view, all available SaaS products have a high standard. Users expect high accessibility from cloud facilities; it is very attractable for large customers with business-continuity opportunity to transfer to Cloud Computing in critical situations. It is possible to accept for different companies to provide independent software stacks for them, but it is very hard for single organization to justify maintain and create more than one stack in the name of software dependability. One of the most available problem is Distributed Denial of Service (DDoS) attacks. These attacks are shifts by cloud computing to the Utility Computing provider from the SaaS provider. In this, who can more willingly absorb it and it also maintains DDOS protection in this competency. Data Security- Data security is another important research topic in cloud computing. Since service providers does not have permission for access to the physical security system of data centers. But they must depend on the infrastructure provider to get full data security. In a virtual private cloud environment, the service provider can only specify the security setting remotely, and we don t know exactly those are fully implemented. For data security, the infrastructure provider must reach the objectives, like confidentiality, auditability, whether applications security setting has been tampered or not. Remote attestation typically requires a trusted platform module (TPM) to generate non-forgeable system summary as the proof of system security. v. Research Challenges Some of the research challenges in cloud computing are discussed in this section. 6

7 A. Elasticity, Scalability and Reliability While the model and infrastructure for how IT services are delivered and consumed may have changed with cloud computing, it is still critical that these new solutions support the same characteristics that have always been important. Whether the cloud serves as a test bed for developers prototyping new services and applications or it is running the latest version of a hot social gaming application, users expect it to be up and running every minute of every day. Thinking about both availability and reliability, the cloud needs to be able to continue to operate while data remains intact in the virtual datacenter regardless of a failure in one or more components. B. Interoperability and Portability Interoperability and portability must be considered up front as part of the risk management and security assurance of any cloud program. Large cloud providers can offer geographic redundancy in the cloud, hopefully enabling high availability with a single provider. Nonetheless, it is advisable to do basic business continuity planning, to help minimize the impact of a worst-case scenario. The following challenges need to be addressed for interoperability and portability: Data Portability o Move data or applications across multiple cloud environments at low cost and minimal disruption. Service Interoperability o The ability of cloud consumers to use their data and services across multiple cloud providers with a unified management interface o The capability of users to communicate among multiple clouds- clouds. System portability o The migration of a fully-stopped VM image from one provider to another, or migrate services and their contents from one provider to another. C. Service Management To productize the functionality of cloud computing, it is important that administrators have a simple tool for defining and metering service offerings. A service offering is a quantified set of services and applications that end users can consume through the provider whether the cloud is private or public in nature. Service offerings should include resource guarantees, metering rules, resource management and billing cycles. The service management functionality should tie into the broader offering repository such that defined services can be quickly and easily deployed and managed by the end user. D. Visibility and Reporting The need to manage cloud services from a performance, service level, and reporting perspective becomes paramount to the success of the deployment of the service. Without strong visibility and reporting mechanisms, the management of customer service levels, system performance, compliance and billing become increasingly difficult. Datacenter operations have the requirement of having real-time visibility and reporting capabilities within the cloud environment to ensure 7

8 compliance, security, billing and charge backs as well as other instruments, which require high levels of granular visibility and reporting. E. Cloud Federation Cloud federation is the practice of interconnecting the cloud computing environments of two or more service providers for the purpose of loads balancing traffic and accommodating spikes in demand. Cloud federation requires one provider to wholesale or rent computing resources to another cloud provider. Those resources become a temporary or permanent extension of the buyer's cloud computing environment, depending on the specific federation agreement between providers. Cloud federation offers two substantial benefits to cloud providers. First, it allows providers to earn revenue from computing resources that would otherwise be idle or underutilized. Second, cloud federation enables cloud providers to expand their geographic footprints and accommodate sudden spikes in demand without having to build new points-of-presence (POPs). F. Administrator, Developer and End User Interfaces One of the primary attributes, and successes, of existing cloud-based services on the market comes from the fact that self-service portals and deployment models shield the complexity of the cloud service from the end user. This helps by driving adoption and by decreasing operating costs as the majority of the management is offloaded to the end user. Within the self-service portal, the consumer of the service should be able to manage their own virtual datacenter, create and launch templates, manage their virtual storage, compute and network resources and access image libraries to get their services up and running quickly. Similarly, administrator interfaces must provide a single pane view into all of the physical resources, virtual machine instances, templates, service offerings, and multiple cloud users. VI. Conclusion Cloud Computing emerged as a major technology to provide services over the Internet in easy and efficient way. The main reason for possible success of cloud computing and vast interest from organizations throughout the world is due to the broad category of services provided with cloud. The cloud computing is making the utility computing into a reality. The current technology does not provide all the requirements needed by the cloud computing. There are many challenges to be addressed by the researchers for making cloud computing work well in reality. Some of the challenges like security issues are very much required for the customers to use the services provided by the cloud. In this paper, the challenges in terms of design, performance and security challenges. The possible solutions to these problems even though lot of work is needed to be done in this regard are also provided as well. 8

9 References [1] Available from: Last visited on the 12 th November, [2] Available from: Last visited on the 15 th November, [3] Bechtolsheim A. (2008). Cloud Computing and Cloud Networking. Talk at UC Berkeley. [4] Brooks D. M. et al. (2000). Power-aware microarchitecture: design and modeling challenges for the next-generation microprocessors, In the Proceedings of IEEE Micro. 20 (6). p [5] Cheng D. (2008). PaaS-onomics: A CIO s Guide to using Platform-as-a-Service to Lower Costs of Application Initiatives While Improving the Business Value of IT. Technical Report of LongJump, a Service of Relationals, Inc. p1-12. [6] Church K., Greenberg A., and Hamilton J. (2008). On delivering embarrassingly distributed cloud services. In the Proceedings of Hot Nets. [7] Cloud Computing: Elastic, Scalable, On-Demand IT Services for Everyone. Cloud.com White Paper. (2010). [8] Hamilton J. (2009). Cooperative Expendable Microslice Servers (CEMS): Low Cost, Low Power Servers for Internet-Scale Services. In the Proceedings of CIDR. 1. p1-8. [9] Kambatla K., Pathak A., and Pucha H. (2009). Towards Optimizing Hadoop Provisioning in the Cloud. In the Proc of the 2009 conference on Hot topics in cloud computing. 22. p1-5. [10] Krautheim F. J. (2009). Private Virtual Infrastructure for Cloud Computing. In the Proceedings of the 2009 conference on Hot topics in cloud computing. 5. p1-5. [11] Rao K. T., Kiran P. S., Reddy L. S. S., Reddy V. K., Rao B. T. (2010). Genetic Algorithm for Energy Efficient Placement of Virtual Machines in Cloud Environment. In the Proceedings of IEEE International Conference on Future Information Technology. 2 (2). p [12] Reddy V. K, Rao B. T., Reddy L. S. S, and Kiran P.S. (2011). Research Issues in Cloud Computing. Global Journal of Computer Science and Technology. 11 (11), p [13] Sandholm T., and Lai K. (2009). MapReduce Optimization using Regulated Dynamic Prioritization. In the Proceedings of the 11 th International Joint Conference on Measurement and Modeling of Computer Systems. p

10 [14] Valancius V, Laoutaris N. Massaulie L., Diot C., and Rodriguez P. (2009). Greening the Internet with Nano Data Centers. In the Proceedings of the 5 th International Conference on Emerging Networking Experiments and Technologies. p [15] Vasic N., Bristis M., Salzgeber V., Kostic D. (2009). Making Cluster Applications Energy-aware. In the Proceedings of Automated Control for Datacenters and Clouds. p