Managing storage in a data-driven world
One of the most reliable indicators in technology is Moore's Law, which states that the number of transistors on a chip doubles roughly every 18 months. An equally impressive trend, one to which almost all of us contribute, is the burgeoning amount of digital data. And, like the increases underlying Moore's Law, there is no near-term end in sight.
While no one has yet announced a law, people have calculated where the data is. Currently, the Internet and attached intranets contain about 300 terabytes (300 trillion bytes). But ten thousand times more data — 3,000 petabytes, or 3 exabytes — is accessible in online files, and an additional 20 exabytes of digital data is stored offline. Even if no new digital content were created, the 300 exabytes of existing analog data — much of which is being digitized — points to a healthy growth for the storage industry.
But it also points to a growing challenge for companies today. A sizable fraction of that data consists of information that is not only vital to their operations, but is often the most valuable part of the business. Without rapid, reliable access to core data, a corporation can find its operations grinding to a halt.
The primacy of data is driving new approaches to managing storage. In particular, distributed network storage is coming to play a critical role. By caching data near to end users, distributed storage avoids the need to continually access a central location over networks subject to bottlenecks or broken connections. This, in turn, is fueling new business models, notably storage utilities — an area that represents a huge opportunity for IBM.
IBM intends to be a major player in the storage utilities marketplace. In this highly competitive arena, market leadership will depend on technology leadership, and we are rapidly advancing the underlying technology — storage-area networks (SANs) and network- attached storage (NAS) — that such utilities will need to succeed. Already, enhancements that greatly improve the performance and capabilities of storage networks are running in our labs.
For instance, we have developed a novel approach to managing SANs that allows a server to access storage without involving the other application servers, thereby reducing delays and removing a major drain on computing resources. Furthermore, the new management software can perform multiple functions automatically, resulting in greater efficiencies and lower operating costs. Our researchers have also pioneered new features for NAS, including load balancing across multiple NAS boxes and approaches to scalability and security that will enable storage utilities to extend their services to Internet users.
There is much still to be done, but there is no doubt that we'll be ready for your
first exabyte.
Paul M. Horn
senior vice president and director,IBM Research
