Cloud Storage in Cloud Computing

Total Page:16

File Type:pdf, Size:1020Kb

Cloud Storage in Cloud Computing

CLOUD STORAGE IN CLOUD COMPUTING

Mr. Pancham Singh Asst. Professor, AKGEC. Vasundhra Tyagi [email protected] B.Tech (3rd year) Sudhanshu Gupta [email protected] B.Tech (3rd year) [email protected]

Abstract— computing lets you access all your applications and As an emerging technology and business documents from anywhere in the world, freeing paradigm, Cloud Computing has taken you from the confines of the desktop and making it commercial computing by storm. Cloud easier for group members in different locations to computing platforms provide easy access to a collaborate.(1) company’s high-performance computing and storage infrastructure through web services. Cloud computing platforms provide massive Providers such as Amazon, Google, Salesforce, scalability, 99.999% reliability, high performance, IBM, Microsoft, and Sun Microsystems have begun and specifiable configurability. These capabilities to establish new data centres for hosting Cloud are provided at relatively low costs compared to computing applications in various locations around dedicated infrastructures. This article covers the the world to provide redundancy and ensure key technologies in Cloud Computing and Cloud reliability in case of site failures.(3) Since user Storage, after the introduction of the Cloud requirements for cloud services are varied, service Storage reference model. providers have to ensure that they can be flexible in The architecture of cloud storage is based on it their service delivery while keeping the users conception and the proposed hierarchical and the isolated from the underlying infrastructure. Recent discussed key technologies involving data advances in microprocessor technology and organization, virtual storage, Data Duplication, software have led to the increasing ability of security, etc. With the development of cloud commodity hardware to run applications within computing and global data growth, it will get Virtual Machines (VMs) efficiently. VMs allow more attention and will be developed well. both the isolation of applications from the The paper also concentrates on analyzing and underlying hardware and other VMs, and the discussing Storage Management. Storage customization of the platform to suit the needs of Management Control optimized is an effective the end user. Providers can expose applications method that will reduce the working time to large- running within VMs, or provide access to VMs scale data Storage management. Combining themselves as a service (e.g. Amazon Elastic Storage devices and control management software Compute Cloud) thereby allowing consumers to will provide system data sharing and system high install their own applications. One of the primary applicability. With Cloud Storage Management uses of cloud computing is for data storage.(1) control mechanism being applied, business With cloud storage, data is stored on multiple third- enterprises could benefit from this. party servers, rather than on the dedicated servers used in traditional networked data storage. When Keywords – cloud computing,storage,architecture storing data, the user sees a virtual server—that is, it appears as if the data is stored in a particular I. INTRODUCTION place with a specific name.(2) The user’s data could be stored on any one or more of the Cloud computing portends a major change in how computers used to create the cloud. The actual to store information and run applications. Instead of storage location may even differ from day to day or running programs and data on an individual even minute to minute, as the cloud dynamically desktop computer, everything is hosted in the manages available storage space. But even though “cloud”—a nebulous assemblage of computers and the location is virtual, the user sees a “static” servers accessed via the Internet.(2) Cloud location for his data—and can actually manage his storage space as if it were connected to his own PC. office applications such as word processors. This Cloud storage has both financial and security scenario is called Software as a Service (SaaS). associated advantages. As for security, data stored in the cloud is secure from accidental erasure or hardware crashes, because it is duplicated across multiple physical machines; since multiple copies of the data are kept continually, the cloud continues to function as normal even if one or more machines go offline.(3)

II. KEY TECHNOLOGIES OF CLOUD COMPUTING

Depending on the type of provided capability, there Figure 1: Cloud computing service types with are four scenarios where Clouds are used as examples showed in Fig.1:

1) Infrastructure as a Service IPs manage a large set of computing resources, III. CLOUD STORAGE ARCHITECTURE such as storing and processing capacity. Through virtualization, they are able to split, assign and Cloud storage is a model of networked computer dynamically resize these resources to build ad-hoc data storage where data is stored on multiple virtual systems as demanded by customers, the SPs. They servers, generally hosted by third parties, rather deploy the software stacks that run their services. than being hosted on dedicated servers. Hosting This is the Infrastructure as a Service (IaaS) companies operate large data centers; and people scenario. who require their data to be hosted buy or lease storage capacity from them and use it for their 2) Platform as a Service storage needs. The data center operators, in the Cloud systems can offer an additional abstraction background, virtualize the resources according to level instead of supplying a virtualized the requirements of the customer and expose them infrastructure, they can provide the software as virtual servers, which the customers can platform where systems run on. The sizing of the themselves manage. Physically, the resource may hardware resources demanded by the, execution of span across multiple servers. the services is made in a transparent manner. This is denoted as Platform as a Service (PaaS). A well- known example is the Google Apps Engine.

3) Storage as a Service Commonly known as Storage as a Service (StaaS), it facilitates cloud applications to scale beyond their limited servers. StaaS allows users to store their data at remote disks and access them anytime from any place. Cloud Storage systems are expected to meet several rigorous requirements for maintaining users’ data and information, including high availability, reliability, performance, replication and data consistency; but because of the conflicting nature of these requirements, no one system implements all of them together.

4) Software as a Service Finally, there are services of potential interest to a wide variety of users hosted in Cloud systems. This is an alternative to locally run applications. An example of this is the online alternatives of typical Figure 2: A typical Cloud Storage system architecture storage offerings today. This can be done with an IV. CLOUD STORAGE REFERENCE adapter to their existing proprietary interface, or by MODEL implementing the interface directly. In addition, existing client libraries such as XAM can be The appeal of cloud storage is due to some of the adapted to this interface as show in Figure 3. same attributes that define other cloud services: pay This interface is also used by administrative and as you go, the illusion of infinite capacity management applications to manage containers, (elasticity), and the simplicity of use/management. accounts, security access and monitoring/billing It is therefore important that any interface for cloud information, even for storage that is accessible by storage support these attributes, while allowing for other protocols. The capabilities of the underlying a multitude of business cases and offerings, long storage and data services are exposed so that clients into the future. can understand the offering. The model created and published by the Storage Networking Industry Association™ , shows multiple types of cloud data storage interfaces able V. KEY TECHNOLOGIES OF CLOUD to support both legacy and new applications. All of STORAGE the interfaces allow storage to be provided on demand, drawn from a pool of resources. The A. Data Organization of Cloud Storage capacity is drawn from a pool of storage capacity Currently, the data storage unit based on cloud provided by storage services. The data services are classification, cloud storage can be divided into two applied to individual data elements as determined categories: Block Storage and File Storage. by the data system metadata. Metadata specifies the data requirements on the basis of individual data (I)Block Storage: elements or on groups of data elements Block Storage of data will write a (containers). different single hard disk, in order to get a larger As shown in Fig 3, the SNIA Cloud Data single read and write bandwidth, Its advantage is Management Interface (CDMI) is the functional the single read and write data quickly, disadvantage interface that applications will use to create, is high cost, and cannot solve the real mass file retrieve, update and delete data elements from the storage. cloud. As part of this interface the client will be able to discover the capabilities of the cloud storage (2)File Storage : offering and use this interface to manage containers File Storage is based on the file-level and the data that is placed in them. In addition, storage, it is to a file on a hard disk, even if the file metadata can be set on containers and their is too large split, they put the same hard disk. The contained data elements through this interface. disadvantage is that a single file read and write performance will be a single hard drive limit, the advantage of a multi-file, multi-user system, the total bandwidth can be increased with the expansion of the storage node, its structure can be unlimited expansion, and low cost. File Storage suitable for the occasion are as follows: a. Large file, the total read bandwidth-intensive - such as Web sites, IPTV; b. Write multiple files simultaneously - for example monitoring; c. Prolong storage of files - such as file backup, storage or search.

B. Storage virtualization Cloud storage in the large number of storage devices and distributed in many different areas, how different manufacturers, different models and even different types (such as FC storage and IP storage) among multiple devices logical volume management, storage management and virtualising multi-link redundancy management will be a huge problem, The Deployment of virtual technology is Figure 3.Cloud Storage reference model a method of computing resources, it will apply the system at different levels: hardware, software, data, It is expected that the interface will be able to be networking, storage, etc. each one to isolate, to implemented by the majority of existing cloud break the data center, servers, storage, networking, data and applications obstacles in the physical The index could be stored in data segment device, to achieve dynamic framework, and to reconstruction, and provides a reference list to centralized management and dynamic use of determine whether the data blocks in the repository. physical resources and virtual resources, improve In the copy operation, the index used to determine flexibility and improve service, manage risk and which data segment to be stored, which data other purposes. Virtual Storage is to enable segment to be copied. When the data de-duplication multiple storage device which looks like a storage software found in a block of data has been device, to achieve unified management, processed before, it will insert a link to the original deployment and monitoring. data set metadata in the data block pointers instead of storing the data block again. If the same block C. Thin Provisioning appears more than once, will generate more than a Thin Provisioning technology goal is to achieve pointer to it. Using variable-length data de- storage resources "according to his needs." System duplication technology can store multiple sets of to the application of virtual storage space, when the discrete metadata image. However each image actual physical space required to write data, the represents the different data set, but all images are system before the actual allocation of physical referenced shared memory pool contains data space and virtual space to physical space to blocks. complete the mapping, which are transparent to the application. Although Use of Thin Provisioning F. Load Balance and Data Migration technology, the actual allocation of physical space Load balances are to keep available storage spaces is small, however the application to see is the actual for later application in different storage devices in allocation of physical space than the larger virtual cloud storage system. Data migration of cloud memory space. With the application of data to storage means moving data in one storage system write more and more people, the actual physical to other storage system in different places. It aims space must be automatic and timely expansion in cooperation and keeping load balance in cloud followed, to avoid the lack of adequate physical storage system. Data migration is one effective space and downtime caused by application. mechanism for load balance. When the storage capacity is used over some threshold proportion D. Storage Security values, the data should be migrated into other cloud Cloud storage and distributed characteristics storage units and keep pointers in the old stored relative to traditional data services, cloud storage positions, or modify and update the metadata at the model for greater dependence on the remote server same time. clusters, cloud platform server cluster is running in However it may bring overhead workload to the network environment, the server cluster may network bandwidth and VO process, and it doesn't contain many user data, and these data may be relieve access bottleneck of concurrence clients. scattered in various virtual data centers, these data centers are not necessarily in the same physical G. Hierarchical Storage location, if access to these data, the control of Most of the cloud storage system is a "loose carelessness, it will face serious security and user cluster," which means that the performance of a data privacy issues, and when these problems arise single node will become a bottleneck, because the when different servers located in different physical data did not, and closely matched the same cluster locations. Cloud storage service providers access to are distributed to the node. As a result, if a file is data for the server access control must be strict. frequently gained access to, then the time it can only be read from a node. The solution is, copy this E. Data Duplication file to multiple cluster nodes, and then change the Data duplication is detected through duplicate data, application to see who else needs the document. In remove redundant files, data blocks of the process, addition, if the file access frequency lowers down, so that only unique data is stored in the system. you need to find a copy of this file and delete Data duplication technology through effective redundant. Often, the final step is rare, and this led reduction of redundant data storage system in the to a lot of wasted space. This requires store possession, use to solve the storage space managers to pay more additional management time. efficiency. A simpler and more effective solution is to add the Data duplication technology, the specific usage is: hierarchical storage management. Automatic the data set (in the backup environment, usually the hierarchical document will visit frequently (or backup data stream) is divided into blocks of data document fragment) or move to RAM-based solid and the data blocks written to disk target region. To state disk cache area. Then, when files are transmit data stream identification data block, data frequently gained access to when the system will be de-duplication engine for the data segment to create provided from the high-speed store the file. This a digital signature (like fingerprints), and the method does not require changes to the signature of a given repository to create an index. environment (or change limited), when files are frequently gained access to they can be identified systems. They use a different approach to data and when to migrate to high-speed storage. Then, resiliency, redundant array of inexpensive nodes, with the visit frequency decreased, the file will be coupled with object based or object-like file automatically migrated to the cache. Therefore, the systems and data replication (multiple copies of the memory becomes self-management and self- data). regulation can be stored. The paper proposes the architecture of the cloud storage, and discusses the related key technologies. VI. CONCLUSIONS AND FUTURE WORK Cloud storage is a new concept, its related products and research is still in the initial state. With the Cloud Storage with a great deal of promise, isn’t rapid increase of data, stored in network storage in designed to be high performing file systems but the cloud will become increasingly important, rather extremely scalable, easy to manage storage market demand will be more strongly. Emphasis should be concerned about the performance, storage and cloud computing for the next reliability, fault tolerance, ease of use, scalability generation of operating system development. and self-management capabilities, as well as cloud

VII. REFERENCES

1 . Qinlu He, Zhanhuai Li, Xiao Zhang, Analysis of the key technology on cloud storage

2010 International Conference on Future Information Technology and Management Engineering

2. Jiyi WU1,2, Lingdi PING1, Xiaoping GE3,Ya Wang4, Jianqing FU1, Cloud Storage as the Infrastructure of Cloud Computing

2010 International Conference on Intelligent Computing and Cognitive Informatics

3. ZHAN Ying, Cloud Storage Management Technology

2009 Second International Conference on Information and Computing Science

Recommended publications