Disk-based backup options multiply

Posted on December 01, 2002

RssImageAltText

By Heidi Biggar

End users are increasingly turning to disk for faster, more efficient, and more reliable backup and restore, and vendors are responding (see figure).

Over the past several months alone, the number of disk-based-backup announcements has increased significantly, resulting in a variety of new—and some would say unconventional—ways for end users to now safeguard enterprise data.

Among those introducing products in this timeframe were start-ups Avail Solutions and Avamar Technologies, online backup-and-recovery provider EVault, and tape giant StorageTek.

Avamar takes "pure" disk approach

End users at the recent Storage Networking World show got a first-hand view of a secondary storage appliance from Avamar Technologies that could, with the right OEM support, change the way users back up, restore, and archive data.

What makes Avamar's Axion different from competing products from vendors such as Nexsan and Quantum is the way it approaches secondary storage from a pure disk, versus tape, perspective. The appliance, which was engineered from the ground up to work with disk, backs up to ATA disk using its own software, not standard tape-based backup applications (e.g., NetWorker and NetBackup) that have been designed to work with sequential access media (i.e., tape).

Click here to enlarge image

"We re-designed the backup model on the assumption that there would be a disk archive rather than a tape archive, and from that flows architectural advantages," explains Kevin Daly, president and CEO of Avamar.

Specifically, Daly points to issues with scalability, cost, and, most importantly, efficiency. "Because it cost so much less than disk, tape historically didn't have to be efficient," he explains. "And it didn't matter that backup/restore processes were generating archives that were many times the size of the data being protected."

But efficiency and total cost of ownership (TCO) are important issues today. "One of the big differences between Axion and products like Connected's TLM and EVault's InfoStage, which were also designed from the ground up to go to disk, is the way Axion can identify redundant data," says W. Curtis Preston, principal consultant for The Storage Group consulting and integration firm, in Oceanside, CA.

Preston is referring to the appliance's "commonality factoring" capability, which enables it to identify logical sequences in the storage space that have a high probability of recurring. Avamar claims that this technique can result in a 10x to 20x reduction in archive data, making the archive approximately the same size as the data being protected.

"When backing up to disk, this is very important," says Preston. "It should require a much smaller amount of disk space, which will allow users to operate at a much lower TCO than with competing solutions."

As for scalability, Daly says that users don't have to anticipate how big their secondary storage requirements will be over time. The system's content-addressed storage scheme reportedly allows for easy and infinite capacity scaling. "It spreads data uniformly through the storage space and allows users to add capacity on-the-fly."

The Axion platform—Advanced Content Store (ACS)—is implemented in a redundant array of independent nodes (RAIN) cluster of standard 1U servers with attached ATA disk. Each 1U node, or module, provides about 0.5TB of disk capacity. The base configuration has seven active nodes, for about 3.5TB of capacity. Avamar recommends implementing the nodes in a RAIN-5 configuration for optimum fault-tolerance. The units can be geographically dispersed for disaster-recovery purposes.

The company has developed "data-protection agents" for Windows, Solaris, and Linux, as well as for Oracle. Pricing starts at $175,000.

Avail goes beyond backup and recovery

At Comdex last month, Avail Solutions unveiled Integrity, a data-protection software product that at first glance might have end users thinking hierarchical storage management (HSM).

However, while Integrity has definite HSM-like attributes, it is very different from traditional HSM. Unlike existing products, it has been engineered from the ground up to meet new data-protection needs.

"We took a look at the data-protection market and saw that some of the products were old and couldn't take advantage of new technologies without Band-Aids," explains Randy Thornburn, vice president of marketing at Avail.

In particular, Thornburn says Integrity addresses two long-standing data-protection problems: the inability to back up to a number and a variety of storage devices (e.g., disk, tape, and network-attached storage [NAS]), and the painstaking process of having to make manual duplicate copies of backup media for long-term safe-keeping.

null

Click here to enlarge image

Integrity actually automates the entire data-protection process—from initial backup to replication and migration, to data restore—regardless of the data's location (on-site, off-site, or offline). Resource Location Modules (RLMs) keep track of storage resources within the network as they are registered.

The software performs the initial backup, pushing backup data to first-tier storage devices (typically, RAID, NAS, or JBOD) that meet the initial performance and storage longevity requirements of the data, and then dynamically directs (i.e., replicates, migrates, or archives) data to second-tier resources according to pre-set user-defined rules.

As for data recovery, Integrity restores data directly from its current stored location, not re-traced through its hierarchical journey. Additionally, end users can set multiple retention policies so that data that has been moved to second-tier storage (e.g., tape) can be actually restored from its initial first-tier storage location (e.g., tape) for a period of time. This allows for quicker restore times in certain situations.

Integrity currently supports Windows NT/2000, Solaris 7/8, and Linux 6.8 and higher operating systems. Pricing starts from $295 for a single-server/disk/tape backup configuration. A bundled solution (Integrity software/NAS device) is available from Promicro Systems for less than $8,000.

Evault targets mid-tier

EVault last month announced that it has begun selling InfoStage, a component of its core data-protection-services business, direct to mid-tier companies.

Mid-tier users typically have limited IT resources, are price-sensitive, and have multiple operating systems with which to contend. InfoStage provides end users with a "tactical" solution for these and other data life-cycle management issues, says Ray Ganong, chief technology officer at EVault.

InfoStage works by sending backup data, which has been encrypted, from various distributed sites over a WAN or LAN to a QuickRecovery disk vault in an InfoStage Director (any direct-attached SCSI or storage area network [SAN] storage subsystem, preferably in a RAID-5 configuration). The vault serves as a storage pool for each server being backed up. It also keeps a history of all backup activities related to each protected server (see figure).

The InfoStage Director manages the disk, catalogs the backup files, and monitors the backup processes at local and remote sites. InfoStage Agents, in turn, scan protected servers for changed data. Management is done via the InfoStage CentralControl console.

EVault says it is able to minimize the size of the storage pools or vaults by compressing the data using a delta-processing algorithm that recognizes new files and modified portions of existing files. Ganong claims that by using this algorithm it is able to compress initial backups to 50% of their true size and subsequent backups (only specific data changes are transmitted after the initial backup) to just 2% of the initial backup.

The software currently works in Windows XP, NT, NT Server, and 2000, Netware, Solaris, Red Hat Linux, IBM AIX, and HP-UX environments and with Exchange application. Support for SQL Server and Oracle is slated for first quarter. The company also plans to introduce migration software early next year. The software will reportedly allow end users to migrate secondary data to archival storage.

StorageTek targets backup, fixed-content

Addressing customer demand for ways of improving access to fixed-content data and protecting business-critical data, StorageTek recently introduced BladeStor, an ATA-based disk subsystem.

The company says the technology fits between its disk and tape families, but admits some overlap with these technologies (in particular, with its low-end disk and high-end tape products).

Target markets include fixed-content distribution (e.g., e-mail, surveillance, healthcare, broadcast), disk-based backup and restore, mirroring and data replication, and general-purpose secondary disk. In backup environments, the appliance works in conjunction with traditional backup applications.

BladeStor consists of an ATA-based disk array, dual active-active RAID controllers, and storage management software (the controller and the management software are supplied by LSI Logic). Each 6U-high disk array can be configured with up to 10 storage blades (each blade contains five, 160GB parallel-ATA drives) for a current maximum capacity of 8TB.

Because BladeStor uses native ATA disk-drive technology with a dual-ported Fibre Channel front-end, the system behaves as if it were a general-purpose Fibre Channel disk array. The value of this is that it makes integration into DAS and SAN environments," says Mike Koponen, marketing manager for StorageTek's disk business unit. "You don't have to write to proprietary APIs to take advantage of this system."

Koponen contends that other fixed-content systems, such as EMC Centera, are comparatively difficult to manage and implement because they require APIs to be written for each application.

BladeStor supports RAID levels 0,1,3, and 5 and can be used in Windows, Solaris, HP-UX, IRIX, AIX, Netware, and Linux Redhat 6 environments. On-board storage management software allows for dynamic capacity expansion and RAID configuration. The array can be partitioned (up to 64) for security and resource-sharing purposes.

A 4TB base unit costs about $85,000, an 8TB unit about $150,000.


Comment and Contribute
(Maximum characters: 1200). You have
characters left.

InfoStor Article Categories:

SAN - Storage Area Network   Disk Arrays
NAS - Network Attached Storage   Storage Blogs
Storage Management   Archived Issues
Backup and Recovery   Data Storage Archives