Managing and protecting all enterprise data


Remote-office backups made easy

With plenty of economical disk-based backup products and cloud-based services available, remote offices can be brought back into the fold.

With plenty of economical disk-based backup products and cloud-based services available, remote offices can be brought back into the fold.

Rick Cook

If backups are a headache for storage managers, backing up data at remote offices is a migraine. The traditional method of handling remote backups is to have a tape system on site so that local staff can run regular backups. This has a number of obvious drawbacks, not the least of which is reliability because of the need for human intervention.

To address the issue of the manual nature of tape-based backups, more and more companies are replacing tape with other solutions in their remote offices. The trend is nascent, but catching on fast.

Remote offices are generally characterized by relatively small amounts of data that have to be backed up, as well as a lack of technically adept staff. In an office with just a couple of gigabytes to back up, tape can be relatively expensive. While tape still has the lowest cost per gigabyte, that's only true of the system as a whole above a certain threshold. Tape drives aren't cheap, and tape loaders and libraries are even more expensive. If the backup volume isn't big enough, the cost of the tape system, especially hardware, dominates the economics. If you use a system that costs several thousand dollars to back up a few gigabytes' worth of files, the economics are seriously skewed.

The problem with tape

But the biggest hitch in a remote tape backup scenario, and the one most likely to jeopardize effective data protection, is that most remote offices don't have technically knowledgeable people on staff who can maintain a tape system. Backups are often handled by an administrative assistant or clerical staffer with little or no training in the tape technology. Even a grandfather-father-son rotation is hard to handle consistently and more complex rotation schemes like the Tower of Hanoi are just about impossible. A multi-tape loader or library mitigates the problem, but those devices are more expensive than a single drive and don't completely eliminate the difficulty. Even a grandfather-father-son scenario requires rotating the oldest tape out of the loader or library for offsite storage on a regular schedule.

"We often had problems," says Edward Ruffolo, IT director at Miron Construction Co. Inc., a Neenah, Wisc.-based firm that's in the process of converting its tape backup to an ExaGrid Systems Inc. disk-to-disk storage device. "You'd think something was on this tape, go to restore it, and there was something wrong with the tape or the drive was dirty," says Ruffolo. "Tape management is a problem; we felt a lot of time and energy was being spent daily just managing tape.

"A receptionist or someone has to change tapes daily," adds Ruffolo. "They mean well, they're trying hard, but they're busy." There have also been cases where someone has left the same tape in the drive all week and, in one case, an employee locked the tapes in a safe and lost the key.

Miron Construction has four remote offices spread across the upper Midwest, some of which are four or five hours from its Neenah headquarters. That makes it hard for the IT staff in Neenah to handle problems.

It could be worse. Grey Healthcare Group Inc., a New York City-based healthcare advertising firm, was very nearly a worst case for tape in remote offices. The company was using a combination of LTO-1, LTO-2 and LTO-3 to back up its remote offices, and a DLT library to back up 8 TB to 13 TB of data at its central location. Not only was its tape capacity maxed out, but the tape systems were old and growing unreliable, requiring constant maintenance efforts by the company's small IT staff to keep them running. Even when the tape systems were running, they weren't running very well. The worst part, says Chris Watkis, Grey Healthcare's director of information technology, was that the backups were unreliable and couldn't be counted on to recover lost data.

Grey Healthcare replaced its tape system with two FalconStor Software Inc. Enterprise Storage Appliance virtual tape libraries (VTLs) and now backs up its other remote offices over the WAN to the libraries. Grey Healthcare Group has four remote sites; one of the FalconStor VTL appliances is in its headquarters and one is in a remote office.

Still a place for tape

Despite some of the drawbacks of tape, it's still the most common method of backing up data at remote offices. The big reason, says Subodh Kulkarni, vice president of global commercial business at Imation Corp., a maker of tape as well as tape-, optical- and disk-based backup systems, is cost. "Tape continues to have the lowest cost per terabyte," he says. "A terabyte cartridge costs $50 to $70, much less than disk."

Although we tend to assume that remote-office staffs lack the skill or time to manage tape, that's not always true. "I may have someone who knows how to deal with tape," says David Hill, principal at the Mesabi Group, a consultancy that deals with backup issues.

And the assumption that remote offices have modest data storage needs doesn't always hold water. Remote sites such as engineering field offices or laboratories can produce very large data sets that have to be archived, something tape is often ideal for. Even if a site has modest storage needs, the data may need to be archived. "Dentists want to keep their X-rays for a long time, but they don't want to keep their business data for 20 years," says Hill. That tends to lead to a hybrid solution with tape as part of the mix.

Hill also suggests that backing up to a central repository may not be ideal. "I may not want to depend on a central site to restore," he says. "If something happens and I need to restore, I may not feel I can restore from the central site because of bandwidth reasons."

Some businesses may want to move their remote sites away from tape, but economic constraints come into play. "It also comes down to the timing of decisions," says Stephanie Balaouras, principal analyst at Forrester Research, a Cambridge, Mass.-based market research firm. "Corporate policy often says the useful life of IT equipment is five years. They have to use what they have for five years."

This is especially true in today's economy. Not everyone is satisfied with tape, but those who are don't see much point in changing. And tape can be extremely reliable. Drives can work flawlessly for years, so many companies see no reason to replace them. Kulkarni notes that Imation is still making cartridges for obsolete tape formats like Travan because there's still demand for them.

Kulkarni agrees that tape has some disadvantages. "On the negative side, tapes do need some knowledge, especially of the tape identification scheme," he says. "It's not quite an off-the-shelf product."

Running "dark"

Switching to another form of backup won't magically produce someone with the training to run remote-office backups. But most of the alternative technologies require much less operator intervention at the remote site, and in some cases none at all. Vendors and their customers are aiming for remote backup that runs "dark," without any human intervention at all. Some of the replacement technologies, especially remote backups over the WAN or into the cloud, are dark technologies as far as the remote office is concerned.

Many backup vendors provide the ability to manage their products remotely, often from a Web-based console. This allows IT staff to monitor and control backups without intervention at the remote office. "This is just a box hanging on your rack," says Ruffolo of Miron Construction's ExaGrid system. "You don't have to worry about it."

Restoring from backup is usually not quite as simple, but can still be managed from the central office in many cases.

Watkis says Grey Healthcare gained a number of advantages with its new backup system. "It has reduced the cost, the amount of system administration time, storage media requirements, the recovery timeline, our service warranty guidelines and improved security because the tapes are encrypted," he says. Deduplication wasn't part of the original specification, but it came with the FalconStor products and has proved to be a major advantage, adds Watkis.

The cost savings realized by switching to offsite VTLs can be substantial. "We were spending over $60,000 or $70,000 on tape media itself," says Watkis. "That's now down to $12,000 and maybe lower." But the company's savings went well beyond the cost of tape media. Staffing costs were cut, and they were even able to avoid some storage system purchases. "We were spending $5,000 to $7,000 a month for people to stay late; that's gone away," he says. "We're no longer purchasing any external storage for our SAN. Deduplication freed up more server space. We budgeted $60,000 for additional storage and that wasn't done."

Miron Construction is in the process of switching to disk-based backup. It installed an ExaGrid system in its data center a year ago and intends to switch the remote offices to disk backup this year.

"The first thing you need to be certain about is to make sure you have a reliable WAN connection," says Watkis. "You need a dedicated switch connection into the internal network for the VTL. Don't try to share a connection."

Watkis says that most central offices already have these kinds of connectivity if they have remote sites. However, having enough bandwidth is vitally important for doing remote backup.

"The VTL method is new and I took a risk in implementing it," says Watkis. "We were really putting business at risk and the solution had to work with an immediate return on investment."

Speed matters

Another important remote backup issue is speed. Specifically, any remote backup planning has to consider how long it takes to do the actual data backup and how long it will take to restore data.

"We found out the hard way that the amount of data being stored matters," says Victor Liu, president at Link High Technologies Inc., a Denville, N.J., reseller specializing in backup technologies. "On a pure Internet solution, more than 50 gigabytes is just not practical." Link High Technologies' Liu says one customer with a large amount of data stored over the Internet took more than four days to do a complete restore.

There are several ways to deal with this. One way is to keep several of the most recent backups on local disk and transmit them to the remote site. Another, which is available from some vendors such as Link High Technologies, is to have the data transferred to USB drives and overnight expressed to the customer needing the restore. With large quantities of data, the difference between reloading from a local disk and reloading over a network can be considerable, even if you take into account the time it takes to ship the disks.

Online backup alternatives

One of the more intriguing backup alternatives for remote sites and smaller offices is cloud-based, or online, backup.

"I wouldn't say it's becoming popular yet," says Eric Burgener, a senior analyst and consultant at Taneja Group, Hopkinton, Mass., "but I think it will over time." Burgener notes that unlike the storage service providers (SSPs) of a few years ago who tried to provide main storage for corporate data, Storage-as-a-Service (SaaS) companies using the Internet "cloud" are concentrating on backing up and archiving data where there are fewer performance constraints.

Many cloud providers, such as Carbonite Inc., are currently targeting individuals and very small businesses. But cloud backup is well suited for remote offices because it can handle dark backups -- automatic backups that don't require manual intervention. If a remote office has a relatively modest amount of data to back up, a cloud service may be a good fit; however, bandwidth may still be an issue. At the least, a broadband connection is required, and consideration should be given to bandwidth requirements for large restores.

Reliability could also be an issue. There have been some well-publicized outages at large cloud providers, but a connection that's performing poorly or not working at all is more likely to be a problem.

10 tips for using cloud backup services

1. Check your bandwidth. You need to know how much data you expect to back up to the cloud service and if your current bandwidth is adequate not just to handle backups in a reasonable time, but for restores.
2. Ensure reliability. A cloud backup service, like any online service, can experience outages. Check on the service's record, noting how many outages they've had and how long they've lasted.
3. Tally the costs. Because services have different fee structures, it's important to know how much data you'll ship to their site, how frequently you'll run backups and how often you expect to restore data. With that information in hand, you'll be able to make accurate cost comparisons.
4. Evaluate access controls. You may want your users to be able to do their own restores, but access to backup data should be controllable to limit unnecessary backups/restores and to protect the data, especially if access from anywhere is allowed.
5. Make sure your data is safe. Ask what measures the service provider takes to safeguard your data. They should have backup data centers and offer encryption for data in flight and at rest. If encryption is an option, get it and make sure it's turned on.
6. Stop and resume. A cloud backup service should allow you to stop a backup in progress and then restart it from the point it was interrupted. Having to rerun an entire backup is costly and time consuming.
7. Big restores. If a disaster strikes and you have to restore your entire backup data set or a large part of it, online transmission will likely be impractical. Find out how the service handles these requirements.
8. Protect desktop and laptop data. If all of your company's user data is stored on servers, you don't have to worry about desktop or laptop PCs. But if you have a mobile workforce or allow local storage, ask if the service provider can also protect the data on those systems.
9. Agents and other software. Many services require an agent to run on the servers you're backing up. Find out if the agent will affect the servers' performance or interfere with other applications, and if they can be managed centrally.
10. Continuous or scheduled backups. Some services can back up your servers and other systems continuously (or nearly), while others do backups on a regular schedule. Make sure your provider offers the types of services that best fit your company's environment.

Costs vary widely among cloud backup providers, as do the methods of calculating them. Some offer "unlimited" amounts of storage for a monthly fee. Others, such as Amazon's Simple Storage Service (S3), charge a few cents per gigabyte per month to store data, with additional charges to upload or download data. Other vendors have sign-up fees or minimum monthly charges.

A major advantage of backing up to the cloud is that, in most cases, you can recover to any computer. This restore-anywhere feature may sound a bit risky, so most cloud services offer encryption as part of their packages.

Because most cloud backup services rely on the Internet to transfer data, there are practical size limits on how much data you can effectively back up. By its nature, cloud backups are slower than most LAN backups, although the speed depends almost entirely on the bandwidth of your connection. The other determining factor is the size of the backup window. With a fairly typical 6 Mbs connection speed and a weekend backup window, 1 TB or 2 TB of actual data transmitted may be pushing the limits. But most transmissions are far smaller, as differential or incremental backups only ship the changed data. The first backup, which is a full backup, takes far longer than subsequent backups.

And just because someone else is handling your backups doesn't mean you can forget about them. Analysts say it's important to trust your data to vendors who have the capability to protect it.

"Consider the class availability and disaster recovery capability," says Stephanie Balaouras, principal analyst at Forrester Research Inc., Cambridge, Mass. "Are they backing up to another site? If I were an enterprise-class company, I'd want to make sure my data was further protected."

Finally, consider the legal implications of where the data is being stored. "For European operations, you need to ask where the data center is located," says Balaouras. "Are you in violation of any national privacy laws by transmitting the data offsite?"

There are still plenty of things to like about tape, so it's unlikely it will disappear from remote offices completely. But it will undoubtedly become much less common as vendors develop replacement technologies at lower costs that enterprises can easily integrate into their existing backup infrastructure and operations.

BIO: Rick Cook specializes in writing about issues related to storage and storage management.

Article 2 of 7

Dig Deeper on Remote data protection

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.

Get More Storage

Access to all of our back issues View All