Data Sharing: Storage as a Strategic Tool
By Scott Hansbury
Today`s skyrocketing volumes of corporate data traffic present a potential gold mine of competitive opportunity. As storage technology evolves to incorporate advanced capabilities such as data sharing, it is emerging as an unexpected strategic tool for tapping this opportunity.
True data sharing is an important component in the evolution away from traditional storage architectures toward intelligent storage networks. In these environments, data sharing enables the seamless sharing of information between heterogeneous systems. In addition to saving costs and streamlining operations, data sharing provides investment protection by enabling a new class of capabilities that use existing storage equipment. Most importantly, it enables companies to use data already flowing through the corporation as a competitive resource for generating new revenues and increasing customer satisfaction.
"The center of computing is moving to content. When you become content- or data-centric, you`re focusing on organizing data, protecting and authorizing access, and finally supplying the data in a variety of formats and speeds," says Robert Gray, an analyst at International Data Corp., a research firm in Framingham, MA. "Applying the storage version of Metcalfe`s law, the potential business value of this data is proportional to the square of the number of people who have shared access to it."
As networks become the primary source for business communication and information distribution, data sharing will enable customers to do more than just keep pace with increasing volumes of data traffic. Efficiently integrating storage systems and anticipating and adjusting to increased capacities, peak usage periods, and points of failure across the network will help make data available as a strategic competitive resource.
While analysts are predicting a return to a centralized model of data storage, it is clear that today`s approach must redefine storage for the network age. Yesterday`s "glass house" data centers must be replaced with centralized storage tailored to meet the needs of distributed networks. This means creating intelligent storage networks that enable heterogeneous computing and storage resources to work together, giving users access to any information, anytime, anywhere, and facilitating the managing, distributing, and protecting of business-critical information across networks. Storage systems must scale to match the power of the data server, and disk and memory performance must keep pace with improvements in network and processing speeds.
Intelligent storage networks enable heterogeneous consolidated storage for multiple types of Unix, Windows NT, and mainframe systems. As part of intelligent storage networks, data sharing merges the benefits of network computing with data center reliability.
With data sharing, customers can buy a single large storage device and connect multiple, heterogeneous servers to one piece of storage equipment, streamlining operations and lowering total cost of ownership. This saves precious processing cycles that can be used for high-value tasks, while ensuring that enterprise information is always current and easily accessible.
In addition, data sharing allows different levels of privilege. For example, it can allow only the owning platform to modify (write) the data, or it can allow read and write by all platforms one at a time, or it can allow "concurrent data sharing" where all platforms can either read or update data simultaneously.
Data sharing enables data created by any host to be immediately accessed for faster, more accurate decision making. And because the shared data is current rather than copied, customers save time and money by eliminating transfers and duplication of data.
In contrast, traditional storage architectures do not have the capacity to make the best use of corporate data. For example, with traditional storage architectures, resources such as disk arrays and tape libraries are attached directly to servers that own, and do not easily share, large amounts of data between heterogeneous platforms. These platforms can be worked together--but only in piecemeal fashion, and mainframes and associated storage resources often reside completely outside of this network.
The new model of centralization will exploit underlying distributed computing architectures with "network-intelligent" storage solutions that transfer greater data loads, streamline data access and distribution, and handle more complex information types, such as digital video. For companies with large investments in mainframe-based data centers, intelligent storage solutions provide the best of both worlds--the performance and efficiency of a distributed computing environment and the reliability and manageability of centralized shared storage.
With storage requirements expected to continue growing rapidly in the foreseeable future, storage solutions must have built-in functionality for advanced data management. Data storage becomes even more critical with the increased use of the Internet and the thin-client computing model. Remote and dataless clients must be able to search for and retrieve data quickly, with consistent and acceptable response times.
As users continue to distribute information across networks, the demand for reliable, scalable, intelligent storage solutions will continue to grow. By enabling heterogeneous computing and storage resources to work together, data sharing technology will be a major factor in the role of intelligent storage networks. And in today`s data-intensive corporations, this capability will be a key factor in remaining competitive.
Scott Hansbury is director of storage software marketing and engineering for Sun Microsystems` Storage Business Unit, in Mountain View, CA.