If you find yourself on mailing lists from backup vendors, channel partners and media portals (or watch twitter for backup/storage topics) … then you have likely seen advertisements like:
Five Easy Steps to Disaster Recovery
Don’t miss this revolutionary, exciting, innovative, and informational …
Yada yada yada, blah blah …
Those webcasts/whitepapers aren’t bad … and in many cases, they may offer new ideas for you to consider.
But let’s be honest — DR is not that revolutionary or innovative. There may be newer, faster, cheaper tools for accomplishing the technical aspects of it, but from an introductory perspective, let me boil the “easy” part down to 1 Step:
STEP 1 (there is only one step) – Get Your Data Out of the Building, Now
From an IT perspective, almost nothing else matters
Get Your Data Out of the Building (GYDOOTB) – those seven words may very well make the difference between your company surviving a crisis and not. They were the first words and the last words in my book, and they are the first thing that I tell anyone who is thinking about revisiting their Backup/HA/DR/BC strategy.
“HOW” — As for the “how” in “GYDOOTB Now”, here are two suggestions:
A) If your workloads (applications) can natively replicate data asynchronously (and often with a failover mechanism built in), do it! E.G. Exchange 2010 DAG, SQL Database Mirroring, Windows Distributed File Systems. This takes you from DR to BC in one clean step, and often with the least infrastructure impact of any technical mechanism.
B) If not, start immediately with either the disk-to-disk component of your backup solution, or a host-based replication solution
If you need to, consider downloading an evaluation of an industry-leading (read: reputable) replication software and just get the data moving. By using host-based methods (app-aware or file-replication), you can typically receive the data on older servers that may not match your current production servers, using cheaper and slower storage than what your production farm may use.
Notice that I did not immediately suggest hardware-based mirroring or anything synchronous. I like those technologies, but those are harder to implement “now”, due to either a larger capital expense or some significant bandwidth commitments. My goal in “Step One” is to protect your data NOW, this weekend, or at least by the end of this month.
“WHERE” is ‘out of the building’
If you have a second data center, go there for sure – and send everything. If you don’t have two data centers, add some extra disk to your favorite branch office and send it there.
There is a bigger question of “how far is far enough away”, but that shouldn’t all be addressed in Step One. More on that later.
This might be a good excuse to look at cloud-based backup. But for that, you will likely look at replicating only a portion of your data, since those services tend to bill by used-capacity. Moreover, your goal in Step One is tactical, not strategic. And embracing cloud-based protection is much like augmenting your storage for distance-based replication – it will likely be part of your long-term DR/BC strategy, but may not qualify for “easy to do now” in your environment — maybe it does. For some of you, it might be as simple as a credit-card and a long-weekend of initial replication to get some key workload covered to the cloud.
Here is the “WHY”
I push on “now” — because too often, I see companies that struggle with “Planning Paralysis” during DR/BC preparation
The reality is that DR/BC takes more than “5 Easy Steps”. It is more about people and process than it is about technology. It involves understanding your business workflow, as well as interdependencies of your data and systems. It involves planning for a wide variety of crises from failed-storage to regional power-failures, each with quantifiable business impact, as well as probability assessments to determine the real necessity. By the way, a lot of the quantifiable metrics (RPO/RTO, BIA/RA, ROI/TCO) are covered in the 2nd chapter of my book, which is available as a free download from www.DataProtectionBible.com.
But all of that stuff takes time. And what too often happens is that you may launch a committee … and then you seek executive sponsorship … and then you start assessing technologies … and … and … and … you start building out this grandiose plan. But while you are still holding committee meetings and printing up big binders of documentation, you actually experience a crisis. When the crisis actually comes, too many companies are left with some well-informed and well-meaning participants and 2/3 of a binder … but no data.
By the way, not all crises are floods or fires. Small ones happen all of the time, but almost all of them rely on data for their recovery.
So, get your data out of the building now. The method may not be what you will eventually roll out. And if you do have well-managed storage, you will eventually look at alternative means of replication. You will also almost assuredly (or should) investigate cloud-based solutions. But by ensuring that your data exists outside the building now, you have many more tactical options – and you have more time.
– Now, you can go watch those webcasts and read those whitepapers
– You can have your committee meetings and start doing the real work of DR/BC planning
– You can start looking at more advanced technologies, justifying capital expenditures, etc.
– And if a crisis happens; you may only have 2/3 of your plan finished, but you will have your data.
That data, along with some elbow grease, will give you a much better chance of recovering.
Thanks for reading…