Keeping two copies is the best bet. But that comes at a price, as building and maintaining this infrastructure means doubling your storage acquisition and operating costs.
Storage is expensive. That raises a question for anyone engaged in cloud backup or considering hosting a backup service for others: Should we keep one copy of data or two?
The immediate answer is two, but economics can enter in. Keeping a single copy of data will be the most cost-effective approach, but there is a downside: It leaves you without any insurance should something happen to the primary copy. During Superstorm Sandy, for example, a New York healthcare firm found itself badly exposed when its DR site in New Jersey suffered from flooding. It took it more than a week to get its systems up and running, at a colossal loss of revenue.
That’s why keeping two copies is the best bet. But that comes at a price, as building and maintaining this infrastructure means doubling your storage acquisition and operating costs.
Probably the simplest way to achieve this is to add tape storage and send tapes to an offsite facility. Tape cartridges, after all, are relatively inexpensive and readily available. But buying a tape silo comes at an expense similar to that of a disk storage subsystem. Further, an offsite storage company is needed to retrieve and store tapes. That’s an added expenditure. Since tape drives tend to be upgraded every 18 months, MSPs will have to keep legacy drives under contract should you have to restore data from older backups.
But there are other factors to consider in a tape-based system. In most backups, tapes are read sequentially, especially if you are trying to recreate a system from a full backup. If one of those tapes is bad, your disaster recovery efforts will be lost. Unfortunately, this happens more than one could imagine. Gartner analyst Dave Russell placed tape backup failure rates at anywhere from 10% to 25%. MSPs can’t afford to risk their reputation by relying on tape.
Another approach, and one favored by many reputable MSPs, is to create a copy of the data in a secondary storage subsystem. This allows for the MSP to have a warm backup of its data that can be accessed quickly in the event of a disaster. However, the cost of building and maintaining this infrastructure is significant.
So what is an MSP to do? Perhaps the best option is to hand this duty off to a cloud storage provider. The economics of doing so are typically favorable, but, more importantly, the service can be offered to existing customers without tying IT staff by doing a point-in-time recovery.
It’s as simple as selecting the date when the issue occurred and start the recovery. There’s no waiting for the tapes to arrive; no need to recover the data sequentially, tape by tape, from the last full backup; or even worry if the tapes actually work.
Telling a CEO that you will have the spreadsheet recovered in 30 minutes sounds much better than telling him or her that the spreadsheet will be recovered within the next 24 hours. Such challenges have caused some MSPs to drop backup from their repertoire. But if you could hand it over to a reliable backup partner, you could use its services to increase profitability while leaving IT staff free to work on strategic initiatives.
For more information, here’s an MSP tech brief on what you need to know about building out a private vs. public cloud.
Guest blogs such as this one are published monthly and are part of MSPmentor's annual platinum sponsorship.