The Closing Window
“Protecting” data means making multiple copies after determining how many are needed, where to put them and how often to update them. Since the dawning of the mainframe era and until recently, it was acceptable to bring servers to their knees nightly in a “race to sunrise” to duplicate data on tape backup before users returned to their desks in the morning. Tapes generally were delivered offsite for storage and retrieved only if an unexpected outage required the data reconstruction.
Such traditional backup solutions, obviously, left at least a day’s worth of data unprotected from a server or disk crash, with rebuilding the server from tape taking many hours to days. Access and recovery gaps widen further with weekly backups or combination strategies of weekly full backups with incremental or differential backups run daily. Such data loss and recovery delay was acceptable for many companies, but competitive pressures and increased data load on corporate networks now make it increasingly less acceptable. Recreating word-processing documents or spreadsheets still may be acceptable, but losing a day’s worth (or more) of productivity of a software-programming team, for example, is probably seen as financially and developmentally too expensive.
The typical backup window has been inexorably closing. Evolutionary business-process changes, accelerated by Internet-based communications, mean distributed employees, partners and customers are logging into servers and databases globally 24x7. E-commerce enterprises require always-on systems and continuous data protection for order processing. Even a delay of e-mail, now generally considered “mission critical” at most companies, can severely impede the ability of a company to operate.
More complicating, however, is that tape-backup technology – despite increases in megabytes-per-minute speed – has not kept up with the growth in server size, causing total restoration times to continue to increase. Not so long ago, a typical 50-gigabyte server could be restored in about eight hours at a rate of around 10 MB/min with traditional tape-backup technology – if everything worked well. For a one-terabyte server today, restoration at a 900 MB/min or 52 GB/hour rate will take 19 hours, even with the best tape-backup available – during which time the server and application would be unavailable to users (and potentially a threat to service level agreements (SLA) delivery).
Enhancing tape backup with disk-based, continuous replication technology can accommodate this dual reality of beefed-up installed-disk capacity and the shrinking backup window because a live replica on disk means simply using a replica instead of having to restore, per se.
Opening The Window
Disk-based continuous replication technologies open the backup window as wide as needed because the production server continues to perform unaffected. Because backup tapes can be made of the disk-based replicas, offloading the burden of periodic tape backup from multiple production servers to a dedicated backup server allows for truly centralized tape backup.
Continuous replication can be performed at either the file level (replication) or at the disk or disk block level (mirroring). Unlike traditional tape back-up, file-based replication uses the smallest amount of network bandwidth possible because it captures and transmits only the changes made, rather than sending the entire file across the network each time. By contrast, disk mirroring (usually operating in a synchronous mode) maintains a replica of databases and/or file systems by capturing and transmitting changes at an entire disk block level. Its synchronous nature also delays the I/O process of the primary server to accommodate the update of the remote site.
The big advantage of asynchronous replication is that it significantly reduces the need for network bandwidth, particularly for large files like databases, video captures and activity logs. Because file replication software is flexible, users have the option of selecting for continuous replication just those files or directories they deem mission critical, typically only five to 10 percent of total data. This greatly trims the cost of off-site protection, compared with mirroring applications.
The elegant beauty of continuous replication is it allows a consistent copy of an application environment to be accessed by software utilities (or other applications) while the production copy stays online. The dedicated high-availability (or disaster-recovery) server can be accessed even if the production server is down. Every file on the target server, regardless of its state on the source, is closed and available for backup at any time. Changes to a file are queued on the target until backup is completed, then automatically applied. Moving the backup process to the secondary server has zero impact on production-server performance and generates no additional network traffic.
While tape backup remains the most popular method of protecting enterprise data from unplanned outages, disk-based replication solutions are gaining wider market acceptance as its uses expand beyond disaster recovery to a variety of operational functions – including helping fulfill SLAs. According to Gartner Group forecasts, fully 75 percent of large enterprises by 2003 will combine disk-based data replication and tape-based technology for rapid application recovery.
As an affordable, scalable and easily managed option for enhancing tape backup systems, disk-based continuous replication technologies can provide both “instant” protection and fast recovery via on-line disk. Tape backup then maintains archival, point-in-time copies of different file versions for the record. And continuous replication facilitates a variety of critical business needs, including high-availability and failover capability, disaster recovery, Web-content replication, data distribution among multiple servers, platform migration, and change management and testing.
IT Challenge
Unfortunately, CIOs with fiduciary responsibility for protecting company data may be tempted to overestimate the ability of tape-based systems to keep pace with server growth and user-access demand. While the Y2K compliance effort was a boon to business continuity and data recovery capabilities funding, complacency has since crept in. If left to fester, apathy could wind up shutting down critical business systems where Y2K failed to do so. Inability to readily access business critical data can adversely affect productivity, competitiveness, customer relations and revenue.
“Apathy appears to play a much greater role in the general lack of readiness than most industry professionals are willing to admit,” according to a post-Y2K Gartner Group study. Failing to embrace a rapid-recovery capability that enhances tape backup, the study concluded, is “like playing Russian roulette with the enterprise.”
The Gartner Group study said most IT managers publicly support the criticality of business continuance, but reported that apathy set in when budgets started tightening and there were no recent catastrophic failures. Turnover of IT personnel often relegated business continuance to a second-rate problem for “my successor” to deal with, echoing arguments in the mid-to-late 1990s for delaying Y2K renovations. “If they don’t feel the pain,” one Gartner Group symposium exhibitor said, “they won’t buy the cure.”
“Like any form of insurance,” the Gartner Group report said, “the decision ultimately boils down to risk evaluation and how much is acceptable to the leadership.” Cost is no longer a viable excuse, as there are many products that focus on the vulnerable server-based environments that are amazingly cost-effective to acquire and deploy. Considering that most of the outages occur with these distributed platforms, protecting your enterprise has never been less expensive.
“Many IT managers place data recovery and business continuance planning in the same category as buying insurance for the company car. ‘Buy as little as possible and keep the expense to a minimum – we’re never going to need it anyway.’ This can be a dangerous approach. With instantaneous online and Web-based transactions driving real-time revenue, any IT manager not prepared to recover mission-critical systems is risking not only their career, but potentially the survival of the enterprise. Such behavior is nothing short of fiduciary malpractice.”
David J. Demlow, vice president of product management, joined NSI Software in March 1997. He is responsible for defining the company’s impressive product roadmap and positioning. This includes identifying market opportunities and requirements from existing and potential customers, performing competitive analysis and monitoring overall industry trends and events. Demlow has more than 10 years of product marketing and management experience in the networking and storage industry.
