Online data backup comes in many flavors, but the most common architecture involves the transfer of data across...
a WAN segment to a service provider that then stores your data. There are many misconceptions about online backup and points to consider when deciding if online backup is right for your business.
Misconception No. 1: Online data backup is only for smaller businesses.
There are many online backup solutions targeted at small to midsized businesses (SMBs), but online backup can have a role in the enterprise, and it's not just to protect laptop data. A service provider armed with the right software can protect the enterprise. How can you tell? While there are many variables to consider. These include multiple sites; what the providers' backup and disaster recovery (DR) plans are; and whether their service-level agreements (SLAs) are inline with your expectations. The problem is that these things are difficult to evaluate. You will only know if their DR plan works if they have actually gone through a disaster; even a report of a test can only give you some assurance that the plan will actually work. SLAs can commit to just about anything. It's their actual execution that can be interesting.
Second, look at the software that the service provider is going to use. Most, if not all, online backup applications support very granular (block-level) backups. You have to dig deeper to look for backup support of common enterprise applications like Oracle, SAP, Exchange and VMware. Companies like AmeriVault Corp., EMC Corp. with its MozyEnterprise, IBM Corp./Arsenal Digital ViaRemote, Iron Mountain Inc. LiveVault and Seagate Technology eVault are some examples, among many others.
Misconception No. 2: I will get locked into a single vendor.
The challenge is that if you have been backing up to the same provider for a number of years and then suddenly the provider starts delivering bad service, switching may be complicated or even impossible because the original service provider has all of your historical information. The software that the service provider is using should be able to support the migration of data to another provider that uses that same software. As is always, address a migration plan in the initial contract.
It's critical to be aware of suppliers using a software application that is only available from them. If there is no one else using their software application and you grow dissatisfied with their service, how can you switch? You can't. Some might say that you could recover everything and then back it up to another service provider. Under this scenario, after backing up several years' worth of data to a provider, the likelihood of actually recovering all this data and its historical copies over a WAN link -- even a fast one -- is functionally impossible.
Being realistic, without a migration capability, you would have to keep a contract in place with the offending provider until all your data expires, while beginning the process all over again with a new provider. Software providers like eVault and Asigra Inc. (the software that AmeriVault uses) have many service providers from which to choose.
Be careful to explore any contractual issues about data ownership. Make sure that you have ownership of the data and that you have the right to move or delete that data when you choose to.
Misconception No. 3: The service provider will be able to see all my data.
Will the service provider be able to see all of your confidential data? It shouldn't. Most applications send and store the backup sets in encrypted form, some even offer key management, which allows for storage of the encrypted keys at the remote vault, in case you loose them. Look for a solution that provides AES 256-level encryption and one that encrypts data during transmission across the wire and when that data is at rest on the remote vault. Most online service providers today offer encryption but, as stated earlier, verify who owns the encryption keys and who's allowed to use them.
Misconception No. 4: I have to have a huge WAN connection to send all this data.
Most of the online backup solutions leverage data deduplication and compression and are typically transmitting less that 1% of the total data set across the WAN. As a result, the WAN connection can be pretty modest. This deduplication is tracked across sites, so if a file in site A is similar to a file in site B, after the file in site A is backed up only the unique parts of the file in site B are sent to the remote vault. Most of the aforementioned companies support this capability.
That said, most enterprise customers tend to use an existing DS3 or better to link to the service provider. The service provider should be able to help you determine what kind of communications link you'll require.
Misconception No. 5: Online data backup gets more expensive over time.
Ironically, the traditional tape vaulting services have this problem more than the electronic vaulting services. Even though the likelihood of using a tape that is three years old decreases with each passing day, traditional tape vaults still charge the same rate to store it. Again, selection of the software is critical but a service provider with the right software solution can tier your data on different types of storage. Some of the software in use by service providers will allow for local storage of most recent data (for very fast recoveries and good for server rebuilds). Tiers of storage can be set up based on speed, capacity and cost at the remote vault.
Providers that support local caching of data and use a tier-pricing model include AmeriVault, Digitiliti and DS3 DataVaulting. While there are many variables and pricing really needs to be done on a case-by-case basis, the market pricing for the "older" backup data should be approximately 30% of the price of the younger/fresh backup data. Typically, this type of pricing is offered when a customer signs a three year to five year storage deal, with an option to renew. So, if someone is paying $2 per GB per month, they would be paying less than $0.60 per GB per month to keep that data for five years. In most cases, the software that the service provider is using can be billed based on the storage tier being used.
Misconception No. 6: It doesn't matter what hardware or software the service provider uses.
Understanding what software and hardware the service provider uses is critical. There are many capabilities to look for in the software, but the hardware is equally critical. If you select storage hardware that efficiently stores data, is easy to manage and scales to very high capacities, then the service provider can lower its costs and pass those savings on to you.
Look for name tier one or tier two storage manufactures. While the service provider may survive an economic downturn, you want to make sure the storage platforms they have selected will, too. Look for advanced features like storage virtualization, thin provisioning and automated tiered storage (either through the software or the hardware).
Why are these important to you? Understanding the type of hardware that the service provider uses gives you some indication as to whether or not the service provider is going to be able to meet its SLA commitments. While a broken SLA gives you something a reason to initiate legal action, it doesn't get your data back.
Misconception No. 6: It's hard to get started with online data backup.
It isn't difficult to get started with online backup; it can be pretty easy, especially with the solutions that are agentless. The best practice is to select an area of the enterprise that's been causing the most challenges to the backup process, see if that works well with the online backup approach, then as you build confidence in the solution, add more to it.
Initial backup can be handled in two ways. First, you can simply start the backup and let it churn through the process. Depending on your data and the size of your WAN connection, this can take three days to four days to complete. This method is especially viable with service providers that provide a local copy of the backup. Backups can be protected locally and data can then "trickle" to the service provider over time. The other method is to have the service provider provide a local drive that's then shipped to the main site and copied into the vault as the baseline backup.
Is online data backup right for your enterprise? As always, the answer is "it depends." There are a variety of ways to improve your satisfaction with the data protection process, and certainly online backup is worthy of consideration.
About the author: George Crump, founder of Storage Switzerland, is an independent storage analyst with more than 25 years of experience in the storage industry.