Disaster Recovery Methods

After reading Karl Palachuk’s blog today – The Reality of Offsite Backup – I thought maybe the discussion could be expanded to include some best practices for disaster recovery. I’ll start off with some of what we’re doing for our clients and I look forward to seeing what kind of comments we get.

So, see if you can guess which technologies are used for the three “backup & disaster recovery” options below. By the way, I know there are many more ways to provide this service. These three solutions seem to be prevalent in our area though.

Option A – Standard file-based backup solution with regular offsite storage of media:

1. Execute disaster recovery steps to call everyone and initiate access to an alternate worksite if necessary. 4 hours
2. Replace any lost/damaged equipment. 2-4 business days minimum depending on requirements and vendors.
3. Bring offsite media to recovery location. 2-4 hours
4. Reinstall operating system on replacement equipment. 2-4 hours
5. Restore of the offsite data to the replaced/fixed equipment. 2-4 hours

Access to recovery systems: 3-5 days
Total disaster recovery time: 3-5 days

Option B – Offsite image replication without local recovery equipment:

1. Execute disaster recovery steps to call everyone and initiate access to an alternate worksite if necessary. 4 hours
2. Replace any lost/damaged equipment. 2-4 business days minimum depending on requirements and vendors.
3. Request for offsite data to be shipped to recovery location on BDR appliance. 24-48 hours
4. Virtualize failed server until permanent recovery equipment is available. 1 hour
   
5. (After Hours) Restore of the offsite data to the replaced/fixed equipment. 2-4 hours
   

Access to recovery systems: 2-3 days
Total disaster recovery time: 3-5 days

Option C – Offsite replication with offsite mount of image as virtual machine and RDP access:

1. Execute disaster recovery steps to call everyone and initiate access to an alternate worksite if necessary. 4 hours
2. Initiate offsite backup image mount on a hosted virtual machine. Provide RDP access to systems as necessary either in existing location or an alternate disaster recovery site. 1-2 hours
   
3. Replace or fix any lost/damaged equipment with local parts, rental servers or cold spares. 2-4 hours
   
4. Request for offsite data to be shipped to recovery location on BDR appliance. 24-48 hours
5. (After Hours) Take remote virtual machine offline and allow BDR appliance to replicate deltas. 2-4 hours
   
6. (After Hours) Restore of the offsite data to the replaced/fixed equipment. 2-4 hours

Access to recovery systems: 4-6 hours
Total disaster recovery time: 3-5 days

Option A uses BackupExec (or similar) with a rotation of weekly offsite drives. Since you’re not doing a CDP you’ll be losing any data since your last offsite transfer. How critical is that data? It would be better to take media offsite daily, but you’re still losing a lot of data in a disaster potentially. Besides busy clients don’t like to mess with tapes or USB drives every day and you end up with missed backups as a result.

Option B is the standard Zenith BDR solution. It’s better than “we’ll back up the most critical 24 GB of data” but has some drawbacks. The first is the amount of time it takes to get replacement equipment. The second is the loss of control over offsite data.

We’ve learned that Option C is better and doesn’t require much more investment. Option C could be a Datto Viridian solution if you don’t have your own offsite hosting facility. If you have somewhere to replicate data to over the internet you might choose a Zenith Arca or SonicWall CDP solution. The trick with Option C is providing immediate access to systems via mounting the backup images. Of course good internet connectivity is required for the RDP to work. This is relatively new technology and you’ll find Sun Virtualbox in this space among others.

In addition, you need to consider whether you'll have personnel available to perform critical functions, where you're going to put them, access to critical services (phone, fax, internet, etc.) in the recovery location and a slew of other items.

It rapidly gets complicated so you absolutely have to plan it out, develop a solution and not just slap together a few technologies and hope it works.

So, how are you doing disaster recovery? Are you providing DR consulting in addition to data backup & recovery? Do you do the DR consulting yourself or outsource it? Do you have a better backup option than we’ve discussed? Maybe you just have comments on one of the technologies I mention.

Let us know your thoughts!

Scott Cameron
Redwood Network Services, Inc.
www.redwoodnetworks.com