Finally, Zisman says that the storage systems, like other computing systems, need to be “autonomic” or self-managing. This breaks down into four elements that occur automatically:
- Configuration – adding and or changing features, servers and software can take place without bringing the system down. Other parts of the system recognize these changes and adapt accordingly minimal human intervention.
- Healing – The system recognizes a failed component, takes it off line and repairs or replaces it. For example, if a file becomes corrupted, the system can locate a copy on a mirror site and replace the damaged file. If a server goes down, it routes traffic to the backup server.
- Protection – The system monitors who is accessing resources. It blocks and reports any unauthorized access attempts.
- Optimization – This is not something that should be done once when installing a new piece of hardware. Autonomic systems constantly monitor system conditions and tune storage, databases, networks and server configurations for peak performance.
A SAN-Ready File System
IBM has a crew working on each of the above points. In February, it incorporated the Storage Networking Industry Association’s new Storage Management Initiative Standard into its Enterprise Storage Server equipment to improve interoperability between its own and other vendors’ equipment. Its autonomic computing project, formerly called eLiza, is gradually adding self-management features to a number of IBM products.
The big news, however, is in the storage virtualization arena. In 1997 a team at IBM’s Almaden Research Facility started developing a new file system to unite all the distributed devices in a heterogeneous network and scale up to hundreds of servers holding billions of files containing petabytes of data accessed by thousands of users.
“Those of us who came from mainframes expected that a lot of machines should be able to get at and share the same storage resources, but in the open systems that had not happened,” says David Pease, Manager of Storage Software at IBMs Almaden Research Facility in San Jose, CA. “SANs give the hardware infrastructure to do that kind of sharing and resource centralization, but it was clear to us that it would take a major new file system infrastructure to fully realize all the benefits of SAN.”
This new file system doesn’t replace any existing file systems, but supplements them.
“In a Windows environment it appears as just another drive letter, Drive S, while in Unix it gets mounted at a mount point and is part of your file space,” Pease explains.
Called the TotalStorage SAN File System, it will be available as an integrated hardware and software package starting in December 2003, extending the storage virtualization’s function. Among other things, it improves data sharing and the ease and efficiency of access to data as well as reducing the application downtime caused by storage and data management tasks. All of which produces the benefit of lightening the load on the IT staff.
“The new system improves productivity and reduces the pain for IT storage and server management staff,” says Zisman.
Discuss this article in the Enterprise IT Planet storage forum.