Continuous uptime, once a requirement for banks, hospitals and casinos, is becoming a necessity for businesses of all types.

In the late 1960s, when businesses began to deploy mainframe and ‘mini’ computers to handle their information management demands, they relied on batch backups to magnetic tape as a means to themselves from all too frequent hardware and software failures. As transaction volumes grew, so did the window necessary for periodic saves. Because faith in these systems hadn’t yet been established, organizations maintained established paper-based business systems as a fallback. While tape backup systems aren’t terribly advanced, most companies still perform saves to tape and store the media off-site.

Technology has evolved with business requirements, and gone are the paper-based ledgers that businesses kept as a failsafe. Access to corporate data systems is no longer reserved for those who work within the four walls of an organization. Web 2.0 is reality and people in 24 time zones are connected to some server somewhere via desktop, laptop, notebook, or smart phone, while organizations reach out to their markets through on-line catalogs, order entry systems, and other types of innovative applications.

Historically, when it came to information technology, management was most concerned with the cost of development and complexity of new applications. While IT budgets are still carefully scrutinized, many are now focused on system availability given the way business computing has pushed beyond the cubical. Recovery strategies that encompass software, data and hardware are high priorities because uptime has become a serious competitive lever.

As the demand for availability has escalated, so has the challenge associated with its attainment. Computing environments are decentralized, with heterogeneous servers dedicated to specific tasks.

It is intuitive that electronic banking, airline reservation, POS, and hospital management systems for instance, require uninterrupted resiliency. But, as Web 2.0 based- business strategies trickle down from large organizations with lots of bricks and clicks resources, to small, single facility, single server shops, 24/7 access to applications and data is essential.

One way to dramatically improve system uptime is by deploying real-time redundant servers. High-availability (HA) clusters improve the availability of applications and services by eliminating single points of failure. Redundant applications, storage and hardware can ensure continuous data access around-the-clock.

HA differs seismically from traditional tape-based backup methodologies because traditional backups can only restore data accumulated in large intervals representing the time frame that tape saves were performed.  HA, on the other hand, allows you to restore in much smaller intervals.

At minimum, two nodes are necessary to form a cluster and facilitate redundancy. Twenty years ago, the hardware, software, and bandwidth necessary to facilitate an HA cluster required a serious financial commitment. Today, inexpensive computers, bandwidth and HA software systems are placing continuous uptime well within the reach of small and mid-sized organizations. 

Beyond HA lies Continuous Data Protection. CDP provides the widest range of recovery points for restoring data with fine granularity. If data is lost either accidentally or maliciously, CDP enables an organization to return to the point in time just prior to the loss for easy, instant recovery.

CDP preserves a record of every transaction that is committed. If a file becomes corrupted and the problem is not immediately discovered, CDP makes it possible to recover the most recent clean copy of the affected file. It’s important to understand that transactions committed in the interval between the damaging event and the restoration will not be recoverable, and therefore it’s imperative to catch the problem quickly.

While HA involves complete system redundancy and ensures system availability through frequent intervals of planned maintenance, unplanned service, or improbable disasters, CDP specifically refers to the recoverability of data. HA tools that facilitate CDP deliver the salient benefits of both disciplines in a highly integrated system. When CDP is a feature of an HA product, tools are available to ensure synchronicity between databases, and easily facilitate data restoration though a common user interface.

Another key distinction between HA and CDP is with HA, transactions are journaled and moved to the mirror processor in intervals that can range upwards from 15 minutes. With CDP, synchronization occurs on an “every write” basis.

Tony’s Fine Foods in northern California is an example of how a small organization uses HA with CDP to keep their core business applications accessible regardless of the circumstances. Tony’s uses IBM’s Power System running IBM i to support a distribution system another one running AIX to handle logistics. Tony’s delivers, on average over a million dollars worth of merchandise to retailers every day. They implemented a DR environment in a co-location facility, and have separate virtualized servers or partitions in the parlance of IBM i5/OS and AIX operating environments. 

“Whether it’s keeping processes moving at our warehouses, having Web sites up for online orders, or dozens of other business requirements, we want our systems to be seamless from the perspective of our customers and partners and our own sales representatives,” according to Mark Geery, IT director. “You can’t predict the future, and that’s why we need the strongest possible backbone.”

As to how organizations will incorporate HA with CDP into forthcoming availability initiatives, in a recent study entitled “Continuous Data Protection: When and Where,” Dave Russell from Gartner, Inc. said, “CDP will play a role as part of the overall recovery methodology and, just as tiered storage exists for capacity, a hierarchy of recovery that includes traditional backup to tape, backup to disk, CDP, snapshot and replication will emerge.”

Once reserved for those who had a profound exposure to loss if their servers failed, HA, now with CDP features is now within the reach of small and mid-level organizations. Traditional tape backups will probably never disappear, but the value associated with fast and granular restoration capabilities is eye opening.

Doug Piper is the director of product strategy for Vision Solution's high availability and systems management products. He has a B.S. in computer science from the University of California Irvine and 26 years of experience in the systems software industry.