Never underestimate the bandwidth of a station wagon loaded with backup tapes. This was thrown on the table during a recent customer meeting in the context of getting data offsite and onto disaster recovery sites. What a great visual I thought to myself. The customer continued, FedEx is an amazing network they have high bandwidth but also high latency. They move millions of packages every day…phenomenal bandwidth. But in the always-on economy 24 hours to ship a backup tape is the epitome of high latency.
I talk to customers regularly about their regular backup rituals and disaster recovery plans. Like clockwork, FedEx (or other overnight carrier) comes up as the preferred network transport to ship tapes to the salt mines. But I had to ask myself is putting your company’s most valuable data on trucks the best option. If it’s getting backed up and sent offsite, it has to have at least some value, right? While I appreciate the speed and efficiency that FedEx offers, to me it seems that scenario introduces cost and risk that companies should think twice about.
Being a chronic instigator of controversy I challenged that customer to stop putting tapes on trucks and take to the AIR. No, I’m not suggesting they put their tapes on airplanes. And yes, I recognize overnighting tapes involves airplanes.
When I refer to AIR I mean the Auto Image Replication feature in NetBackup. Auto Image Replication allows customers to automatically create copies of mission critical backups at a remote location that forms part of a separate NetBackup domain. This domain could be another data center belonging to the same organization, a dedicated disaster recovery site within the organization or even a third party disaster recovery facility.
The growth in the use of disk storage (particularly deduplicating disk storage) for backup has highlighted a limitation of disk storage when it comes to site disaster recovery. Tapes can easily be sent to offsite storage and then to a disaster recovery site in the event of a site outage.
Disk doesn’t offer this flexibility and, while many disk technologies do have the ability to replicate their contents to a compatible array at a remote location accessing the backups with NetBackup is not always a simple matter. The use of Storage Lifecycle Policies and optimized duplication allows disk-based backups to be replicated between devices under NetBackup control. Thus to achieve a site disaster recovery capability with optimized duplication requires a NetBackup domain that spans at least two geographically remote sites.
(This may range from two rooms in the same building/complex to two data centers in separate cities –obviously the greater the separation between the sites the greater the type and severity of disaster you can protect against.)
Electronic off siting in this way allows the backup to be duplicated to an off-site location as soon as the backup has completed and without the need for user intervention. It also means that the duplicate copy is available at the disaster recovery site as soon as the duplication has completed. This offers a significant advantage of traditional tape based off-siting in which backups must be duplicated to tapes that are then removed from the tape library and shipped to the disaster recovery site, a time consuming and costly activity with a risk that the tapes could be lost or stolen in transit.
So, I offer you the same challenge – It’s time to modernize your DR strategy -- stop putting tapes on trucks and take to the AIR.