Disaster Recovery 201: An advanced course
There’s a very good piece by Esther Shein over at Computerworld today called “Disaster recovery 101: What you need to know”.
Shein does a great job of reviewing the pressing need that all businesses have for DR, and cites some scary statistics about the paltry uptake of DR among SMBs.
The most important part of her story covers the need for companies to have a DR plan. This is great advice, and comes before selecting a technology provider to help implement that plan. We couldn’t agree more.
Shein also repeats our mantra, “Test, test test.” It’s good advice, and bears repeating because DR testing is the kind of eat-your-spinach advice that is easy to ignore. But if you don’t test your DR plan, then you don’t have a DR plan.
I called this post “Disaster Recovery 201” because I’d like to add a few things to Shein’s piece that are a bit more advanced.
if you don’t test your DR plan, then you don’t have a DR plan
First, I want to point out that using a cloud storage or cloud backup solution is NOT disaster recovery. There are many scenarios where all you will need to do is recover some data, but true DR means being able to spin up servers that end-users can connect to, to keep working, and the cloud backups stuff can’t do that.
Second, Shein quotes someone in the story as advising triaging what gets backed up by file type. Again, this advice, while well meant (prioritizing what to protect is important), doing so at the file type level is the wrong level of granularity.
What corporate IT needs to is classify their servers as tier 1, 2 or 3, based on the impact to the company if a server becomes unavailable to the business and its employees. Email is tier 1 for almost all businesses. ERP is for manufacturing firms; document management for law firms; web servers for e-commerce, etc.
Once you determine the tiers of servers, and protect them accordingly (RPO and RTO that meet the needs of your business in a disaster), then you implement a data protection / snapshotting technology that captures everything on the server – the OS, app, settings and data – and protects it. Fool around with file types and you will be lost in the weeds in no time.
- » Ela Osterberger, Deliveroo: On inspiration, injustice, and building great data science teams
- » Veeam to be acquired by Insight Partners in $5 billion deal
- » Securing distributed clouds with an integrated approach: A guide
- » VMware Cloud on AWS implementation best practice: How to accelerate benefits with upfront planning