Tuning maid storage for backup and archive data

By | August 18, 2005

Companies are generating data faster than ever before. This growing amount of data needs to be secure and easily accessible to be of value. In addition, performance demands are greater than ever before. Backup windows are under pressure and therefore favor high bandwidth. Shorter recovery time objectives (RTOs) are needed for fast restores and quick random access to archived data.

Archive data have long lifetimes, so an archive storage system needs to ensure data retention over long periods without disrupting availability. The total cost, including acquisition and management, needs to be reasonable.

Disk systems are well-suited for high-performance access and reliability through RAID, but do not fare well on cost and scale when compared to tape. Automated tape libraries are favored in cost and scale but are severely limited in access performance and in the cost-effectiveness of data redundancy.

COPAN Systems has solved the tape/disk conundrum by providing complete storage solutions that unlock the value of long-term data by introducing a new tier to the storage hierarchy. COPAN Systems Revolution 200T introduces a patent-pending MAID-based storage system platform that delivers reliable and massively scalable disk-based storage with a VTL front-end at a price that is competitive with tape. COPAN Systems Massive Array of Idle Disks (MAID) architecture spins only those disks that are required for read/write access. Because most disks are idle most of the time, more disks can be packed into a smaller space. This platform not only takes less room per terabyte of storage, but also reduces costs associated with power consumption and cooling. Because disks spin only when needed, they also last up to four times longer than disks in an all-spinning system.

Performance analysis of active archive storage using COPAN Systems Revolution 200T suggests that the file throughput will approach that of an all-spinning disk system. With its performance, data reliability, storage scalability, and affordability benefits, COPAN Systems is an intelligent choice for archive storage.

To increase bandwidth for backups, COPAN Systems uses multiple parallel RAID controllers, more specifically, controllers managed by COPAN’s POWER MANAGED RAID™ software. The company’s 3-tier architecture allows application servers or hosts to connect to a common virtualization system that provides I/O load balancing and different storage personalities (tape, file, or disk) across the MAID system.

Data reliability, as measured by the probability of data loss in a RAID disk set, is increased by 14 times over all-powered RAID systems. With proactive disk health monitoring provided by COPAN Systems’ DISK AEROBICS™, reliability at least quadruples over all-powered RAID systems. This benefit supports long-term retention of archival data.

Archival application workloads can be characterized as write once/read occasionally. Active archive applications are more demanding than deep-vaulted archives that have modest random data access frequencies measured in numbers per hour.. Data in an active archive must always be accessible preferably within seconds. MAID systems are ideal for such archival data. COPAN Systems’ flagship product, the Revolution 200T, is the first full implementation of the MAID architecture. It is being used by large organizations today for data archiving via a variety of data management software products that provide direct access to files on disk, all presented as virtual tape.

A 2005 study by Enterprise Strategy Group (ESG) concluded that many large companies see the benefit of keeping more data on disk. End users see disk-based solutions as offering a greater degree of access while also reducing the complexity and increasing the reliability of their overall storage strategies. “While tape will continue to be a viable choice for long-term archival of information,” ESG explained, “we believe a new breed of emergent disk-based systems which provide users with the means to retain massive amounts of data online, will challenge users’ perceptions about the true cost of tape.”

ESG noted that organizations are already moving to disk and that many more will go that direction in the next few years. The study found that IT professionals see the value of moving more data to disk and would like to do so. “End users recognize the need for new technologies that will enable their organization to better meet changing business requirements such as the need for faster, more reliable backups and restores, improved business continuity and disaster recovery protection, and highly available on-line archives to support regulatory compliance initiatives.”

Leave a Reply