Cloud-based disaster recovery is to conventional disaster recovery what sealed hard drives were to removable hard drives way back before the turn of the century.
Early in my career my responsibilities included maintenance and repair CDC hard drives in customer data centers. Those drives were kind of like small top-loader washing machines. Instead of opening the lid to throw in some dirty laundry, you put a stack of disk platters inside and when the drive started up read-write heads shot out over the platters. Occasionally one or more heads crashed into the disk surface. It was my job to fix a hard drive after a head crash. That meant tearing down the drive, cleaning debris, replacing the absolute filter, putting in a new set of read-write heads and aligning the heads. The twenty heads had to be precisely aligned so that disk packs could be interchangeable between drives. You didn’t do that with a plumb line. Head alignment required a field test unit to position the read/write head assembly, a special disk pack called a CE pack, a head alignment tool, and an oscilloscope to look for the correct waveform as each head was properly aligned. Repairing drives after a head crash was a difficult and stressful job. When I tried to align heads it seemed they didn’t want to move at all or they jumped all over the place. My back would get sore from bending over the machine and I struggled to control my temper all the while hoping not to crash a new head into the CE pack. It usually took me most of a day to repair and align a drive after a crash (I’m sure there were people who were better at it than I was). When we started using sealed disk drives all we had to do was replace the failing drive. What a relief.
Now consider disaster recovery testing. When I worked as an IT services manager, we would plan an annual sojourn to a large metropolitan area to test our recovery procedures in a contracted recovery center. Even setting aside the travel and associated costs, the day spent testing recovery was arduous at best. To begin with, recovery was from tape. How many times have you started a tape recovery and after an hour or two you asked yourself, “Should it be taking this long”? The follow-up question to that is of course, “Should I start over?” Then, “Should I try a different tape?” Before you know it, 20 hours have crawled by and you end up with mixed results. Well, that was just after the turn of the century and we’ve come a long way since then. Today we can virtualize data centers in a cloud environment. With data deduplication we can take frequent snapshots of servers in a production data center and store the data offsite almost immediately. That dramatically improves recovery times and recovery points. It also enables the testing of DR plans without sending personnel, tapes, and documentation on the road. Easier, cheaper, faster, what’s left? Security. A good cloud services provider will follow stringent security standards (e.g. SAS-70 Type II) and employ a number of procedures and technologies to ensure customer data is protected. A service-level agreement should specify the level of security provided. When I look at this recovery-as-a-service (RaaS) technology I can’t help but be reminded of the feeling I had with the advent of sealed disk drives. What a relief.