Redefining Near-Online Storage

New types of data and new business requirements are driving change for the typical storage infrastructure. Traditionally, storage infrastructure has consisted of three distinct layers: primary storage for workloads that require high-volume access to data such as online transaction processing (OLTP) applications, backup and recovery applications, and archiving applications. Primary storage has been disk-based; backup and recovery and archiving have been tape-based.

In some cases, enterprises have data that they aren't ready to archive but don't want to maintain on disk drives. Many businesses in this situation have opted to store this data near-online on either optical media or quick-loading tape drives. Although using those technologies can drop access times significantly, it can also increase the cost of storage. According to some studies, the lifetime total cost of ownership (TCO) for disk-based systems for near-online storage can be as much as 30 times greater than the alternatives.

During the past year or so, however, lower-cost disk-based products for near-online storage have begun to gain a toehold in the market. Five areas of need are fueling the interest in disk-based near-online storage. The most obvious is the growth of email. According to research from StorageTek, which released an innovative solution to the near-online storage problem, Instant Messaging (IM) and email data will account for half of all stored corporate data by 2004. Email has distinctive storage requirements: On the one hand, most users don't regularly need to access vast amounts of stored email; on the other hand, when users do need to retrieve messages, they want to do so quickly. In addition to offering an email storage solution, near-online storage is a good solution for audio, image, and video databases, which often require random access to significant amounts of storage capacity but aren't subjected to high I/O demands.

Business continuity and disaster recovery are the last two areas of need driving the search for a better near-online storage alternative. Companies are taking a hard look at how quickly they can recover after a disaster. Often, it takes too long to restore mission-critical data from tape.

Low-cost IDE ATA disk arrays are one solution for more cost-effective disk-based near-online storage. IDE ATA disks are the workhorses of the PC marketplace. According to the ATTO Technology white paper "Using ATA Disk Technology and ADXT to Power Next Generation Storage Arrays" ( http://www.attotech.com/diamond/pdf/adxtwhitepaper.pdf ), approximately 85 percent of all primary PC disk drives sold have an IDE ATA interface. And although IDE ATA drives entered the marketplace as a low-cost, low-speed, low-reliability technology, their performance and reliability have improved significantly.

Network Appliance (NetApp) has packaged an array of ATA disk drives with its Data ONTAP OS in a product called NearStore ( http://www.netapp.com/products/nearstore ). NearStore is targeted to applications that demand quicker random access than tape systems can provide but don't have the high volume demands associated with primary storage. EMC has built what it calls a "content address storage" (CAS) system around an array of ATA disks ( http://www.emc.com/products/networked/cas/index.jsp ). And EMC argues that its Celerra product (http://www.emc.com/products/networking/celerra.jsp )is the first storage system designed to meet the storage requirements of high data volumes such as email.

StorageTek recently unveiled what might be the most innovative near-online storage solution yet. The company has packaged IDE ATA disk arrays into what it calls "storage blades." The storage blade concept is similar to the server blade concept. A server blade is a single circuit board populated with components such as processors, memory, and network connections that are usually found on multiple boards. Server blades are designed to slide into a chassis to increase the computer power of a server. StorageTek's BladeStore System (http://storagetek.shareholder.com/news/20021021-92950.cfm )is a group of servers with attached IDE ATA hard disks. Each blade contains 5 disks and a CPU; a 6U (10.5") chassis can hold 10 blades. Consequently, the system can scale from 4TB to 160TB of storage. Storage controllers simply see 10 high-capacity Fibre Channel disks.

Disk-based near-online systems are finding their place in the enterprise between primary storage and archiving. IDE ATA disk arrays are changing the way storage managers make economic and infrastructure decisions by increasing their storage options.