Throughout the world of enterprise storage, the world is divided between file storage, block storage and object storage. While traditional file storage and block storage still dominate most of the enterprise storage market, with the birth of object storage, a new type of storage, its inherent advantages have allowed it to eat into the enterprise storage market at a rapid pace. This article is a brief overview of the ins and outs of object storage.
Over the past two decades, the Internet has rapidly taken over all areas of human work and life with an explosion of Internet companies, which has led to the emergence of a variety of Internet businesses that need to handle geometrically increasing amounts of data. The drawbacks of traditional file storage have also become apparent with the processing of huge volumes of files, and for hundreds of millions or even billions of files, file storage has become a strain, so a new form of storage - object storage - was born. In 2006, Amazon launched the S3 (Simple StorageService) service, which is an epoch-making service and the de facto standard for object storage, and in a way, object storage and S3 can be equated.
Having said all that, what is object storage anyway? Don’t worry, we’ll take you into the world of object storage.
In a nutshell, object storage combines the high-speed direct access disk features of SAN with the distributed sharing features of NAS, making it ideal for Internet applications. The most basic unit of data in file storage is the file, while the object is the most basic access unit in object storage. Compared to file storage, object storage has the following main features, briefly listed as follows. ● Flat space: no traditional file storage with a complicated directory structure, no need to consider the path when accessing when dealing with large amounts of data
● Horizontal Scaling: When capacity performance bottlenecks occur, the capacity and performance can be increased nearly linearly by expanding storage nodes horizontally online without affecting the front-end business. Object storage scales better than distributed file storage because of the flat data organisation architecture
● K/V access: Each object has a specific Key corresponding to it, so that access to a specific object is only required to access its Key, improving the development efficiency of business systems
● Resource consumption: In the access process of traditional file system, if the client does not actively release the connection, the connection keeps occupying the resources, which is called long connection, while the object storage uses HTTP short connection to access the data, which can automatically release the related resources after the operation, so that the system can carry more load, and also can greatly improve the system stability
● Lifecycle management: For data with a time limit such as surveillance video, the lifecycle is automatically set and deleted when it expires.
● Version management: If the same data is written multiple times, each copy of the data can be saved as a historical version, so that changes to the data can be tracked throughout the process efficiently, and the data can be prevented from being maliciously tampered with, which is a more logical data protection mechanism than snapshots
Rich metadata: Each object can have more useful metadata attached to it, and metadata can be retrieved efficiently. The object storage can carry more complex business
Object storage was initially born along with the development of the Internet and was first applied to the network disk service, but with the advent of the era of big data, more and more traditional enterprises showed an explosive growth of unstructured data, the drawbacks of file storage then appeared, so also began to gradually migrate to object storage. After dedicated research and development, Shanghai Storage Xun’s object storage software CX-CLOUD-OBJ is a one-stop solution for enterprise data storage worries. -OBJ not only provides an object storage interface, but can also access object data with traditional file interfaces (e.g. NFS, FTP, etc.), providing a more compatible management solution as well as a data migration solution. In addition, CX-CLOUD-OBJ also provides a comprehensive solution to facilitate the migration of customers' data between private and public clouds.
Unstructured data such as images and videos are best suited to be stored in object storage systems with massive capacity and near infinite scalability. A typical example is the scalable video surveillance data access solution in the security industry. The figure below shows a security company video surveillance cluster, configured with 1PB of available capacity, the front-end camera through the application server will be deposited into the video stream through the S3 interface CX-CLOUD-OBJ, can simultaneously meet the concurrent needs of multiple high-bandwidth video. The infrastructure can be deployed within a large customer’s enterprise to provide localised services to key customers, or deployed in an IDC to provide proprietary cloud services directly to SMEs, where clients can access the video data in CX-CLOUD-OBJ directly through the public network.
The company’s main business is the development of new businesses such as remote card processing, electronic ticketing, and Internet finance, which generate a large amount of image, audio and video data, which need to be preserved for a long time on the one hand, while most require off-site dual-active, cross-IDC multi-live and other enhanced protection measures. Such application scenarios also give object storage a stage to shine, with its reliable HTTP access interface, flat management space, versioning, lifecycle management and a series of advantages over traditional file storage that can be brought into full play.
The era of big data has arrived, and for enterprises, the most important asset is not only people, but also data. Object storage is the future of data storage, and Storage Xpress will follow the trend of the times and do its utmost to protect our customers' business and build a rock-solid, golden fortress of data.