Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure
Maintaining a backup and recovery system that supports an organization's business objectives has always been a big challenge facing IT departments.
Issues such as weighing the risk of vendor lock-in against the potential benefits and risks of new technologies, the emergence of continuous data protection (CDP) and other new backup and recovery technologies, and the changing roles of tape and disk in the storage arena are all vying for the attention of storage users as they craft a backup and recovery plan. The rapidly shifting storage landscape brings with it plenty of confusion and uncertainty but it also offers opportunities to improve storage processes.
Selecting a platform remains one of the most vital elements of any business solution. Many users say they must balance the risk of vendor lock-in against the potential benefits of new technologies. Unfortunately, there are no easy answers to this question.
"Storage vendors can't help them with this; no one can," says David Scott, president and CEO of 3PAR. "Because as soon as a solution is developed that adds value in comparison with competitive offerings, it is by definition unique and differentiated."
Scott says the only choice a customer has is at what level in the storage solution architecture they choose to be locked in. The lock in is always determined by software: whether that software operates at the host server level, in a network switch or appliance, or in the storage array itself.
"The cost of moving from that software stack to another should be compared with the benefits delivered by that software stack in its expected lifetime," says Scott. If the benefits outweigh the expected transition costs, go for it. If they don't, choose another solution.
The Rise of CDP
Perhaps no new backup technology has created more buzz than continuous data protection. The first continuous backup products were deployed more than a year ago, but the recent entry of such heavyweight contestants as Microsoft and IBM have brought the technology into clear focus. The solutions promise a lot, and users expect them to collapse their recovery point objectives from hours or days to a matter of minutes.
Paul Schoenau, Ciena's TSG senior product manager, says users can expect lower and lower recovery point objectives (RPOs) for backups. When the backups are completed remotely, it is critical to ensure that the MAN/WAN network between the primary and backup location meets these tight RPOs and recovery time objectives (RTOs), he says.
"Enterprises need to balance the cost of the networking services with the performance needed," says Schoenau. "Some of the key performance considerations are the throughput and latency."
Schoenau believes that a seamless, extensive management solution is required to understand end-to-end performance and quickly isolate fault within the data center the MAN/WAN network.
Continuous back-up products are just emerging, says Brian Biles, vice president of marketing for Data Domain, so it's easy to imagine that there will be more extensive solutions over time.
Disk Assumes a New Role
Industry experts expect data loss windows to decline from 24 hours to mere seconds. Experts also believe that the number of bytes flowing to tape (the preferred archiving mechanism) won't change, but the number of recovery bytes flowing from tape will diminish greatly.
Customers will increasingly recover from disk-based solutions, assuming high disk utilization efficiency will make it affordable for users, says Scott. This will minimize the flow of bytes back from tape, leaving tape as the long-term archiving solution.
Data Domain's Biles believes that due to the advent of capacity optimization technology, the use of tape in general will decline over the next five to 10 years. He predicts that the decline will occur in three phases:
- Tape library sales will decline, as more on-site retention becomes disk-based;
- Tape will lose favor as a vehicle for sending data off site, as capacity-optimized replication diminishes the cost of WAN bandwidth required to get data off site; and
- Tape will lose favor as an archiving vehicle because once offline, it can't be easily quality-controlled.
"I do not foresee tape disappearing anytime in the near future, but its function within the data center will shift to one of archiving rather than that of a primary backup and recovery mechanism," says Skeffington.
Analysts expect disk to become the primary recovery medium, and in order to create a self-service mechanism down to the end-user level, storage users will have to develop recovery processes based on the capabilities of their backup products.
"The driving factor will be reduced recovery point objectives, reduced recovery time objectives and the improved fidelity of the data recovered," says Scott.
He also predicts that self-service mechanisms will become increasingly important as the volume of stored data increases. "It will simply be impractical for a storage administrator to be piggy-in-the-middle for recovery of unstructured data files associated with individuals," says Scott.
The question of continuous backup is not based on if or when, but whether a company can stomach the high cost, because the capabilities are already available today, says Skeffington.
"The majority of restores performed by backup administrators today are requested by the end users directly, and are usually limited to single files or home directories," he says. "The protection of continuous backup will ultimately make its way down to the end user, where all created files and subsequent iterations will be saved as a matter of routine."
That said, Skeffington believes that the solution will be phased in, with the most critical data and files, or those of the "C" level, protected first.
For more storage features, visit Enterprise Storage Forum Special Reports