IBM publishes software-defined storage technology for the large data age

"Tenkine Server channel May 21", IBM has released a series of software-defined storage products that not only improve the economy but also enable businesses to access and process any type of data stored on any device anywhere in the world. One of these technologies, called resilient storage, delivers unprecedented performance, unlimited scalability, and the ability to reduce storage costs by as much as 90% by automating the movement of data to the most economical storage devices.

This groundbreaking new technology was born in the IBM Institute, which allows businesses to leverage (and not just manage) The vast array of data generated by countless devices, sensors, business processes, and social networks. New storage software is ideal for very high data-intensive applications that require high speed access to large amounts of information, such as seismic data processing, risk management and financial analysis, weather modelling and scientific research, and determining the next best action in real-time retailing.

"In today's business world, data has become the basis of competitive advantage, but the rapid growth of digital information, the number of large, so that the traditional data storage and management has been unable to deal with." "IBM Senior vice president, System and technology general manager Tom Rosamilia pointed out," We provide customers with innovative storage technology, speed, scalability and cost of unprecedented upgrade, will make customers more competitive advantage. ”

Software-defined storage is a series of software features that automate the management of local and global data, providing breakthrough data access, simpler management, and the ability to expand the technology infrastructure faster and more economically as the volume of data increases. In addition, these improvements can provide both automated and virtual storage with any of the company's storage systems.

Techniques for changing the rules of the game

The IBM Watson system employs resilient storage technology. IBM Watson used 200 million of pages of structured and unstructured data, including Wikipedia, in the "Danger Edge" competition. By using flexible storage, 5TB's Watson "knowledge" (equivalent to 200 million pages of data) is loaded into the computer's memory in just a few minutes.

One of the major reasons Watson chose the software in the danger Edge competition is scalability, which can extend the architecture to up to thousands of YB. 1YB (yao byte) equals 1 billion pb (bytes), which is equivalent to the amount of data stored in the data center of 1 million city blocks, and the sum of 1 million blocks equivalent to the area of Delaware State and Rhode Island State.

The IBM Institute has demonstrated that resilient storage can successfully scan 10 billion files in just 43 minutes with a single system, providing unmatched performance for customers who exploit business insights by analyzing massive amounts of data.

At its core, resilient storage is based on IBM's GPFS to provide online storage management, scalable access, and integrated data governance tools capable of managing large volumes of data and billions of of files. For example, in contrast to a standard SAS disk, resilient storage can achieve a six times-fold performance boost with flash storage built into the server. This feature identifies Flash storage within the server and automatically uses it as a cache memory to improve performance.

Resilient storage virtualization allows multiple systems and applications to share common storage pools. This helps to achieve transparent global data access without modifying the application and eliminating the need for additional or frequent storage management changes. Because resilient storage does not rely on centralized management to determine file location and layout, customers can ensure continuity and high availability of data access in the event of a software or hardware failure.

The growing volume of data is part of the DNA of the National Center for Atmospheric Research's Computing and Information Services Laboratory (CISL). The agency's Wyoming and Colorado Center store and manage more than 50PB of information, relying on resilient storage technology to enable researchers to quickly access a wide variety of data.

"We provide computing, education and research data Services for Earth science to over 1,000 users in more than 200 different locations," said Pamela Gillman, manager of the CISL data Analysis service. "IBM Global file system software supports scalable, reliable, and fast access to information." This has significantly improved the performance of departments and the organization as a whole. ”

A key element of resilient storage is its ability to automate and intelligently move data to the most strategic and economical available storage systems. For example, through policy-driven features and real-time analysis, resilient storage can automatically move infrequently used data to less expensive low-cost tape drives, while storing frequently accessed data on high-speed flash systems for faster access. These policy-driven features can provide up to 90% of the cost savings.

In addition, the software provides native encryption and security erasure features that ensure data security retention in order to comply with regulations such as the health Insurance Circulation and Liability Act (HIPAA) and the Sarbanes-Oxley Act (Oxley).

By supporting OpenStack Cloud management software, resilient storage also helps customers achieve global data sharing and collaboration across private, public, and mixed cloud data storage, management, and access. In addition to supporting OpenStack cinder and swift access, resilient storage supports other open APIs, such as POSIX and Hadoop.

For traditional storage systems, transactional and analytical applications must move data to different storage areas for optimal results, but resilient storage can support different application workloads, including Hadoop based analysis, by automatically balancing resources. This significantly accelerates the analysis and eliminates the high cost and time-consuming process of generating duplicate copies of data.

Flexible storage software will be launched as IBM SoftLayer Cloud service later this year.