Late one day in September, I was out jogging not far from my home in Apollo Beach. The turnaround point of my run is a small state park that is right on Tampa Bay. To my knowledge it is one of the only places in the Tampa area where you can simultaneously see downtown Tampa and downtown St. Petersburg. As I arrived at the park, off in the distance, a cloudburst had just started over Tampa. I took out my trusty iPhone and snapped a pic. The west coast of Florida in summer is known for frequent short, intense thunderstorms that produce cloudbursts like this; they often look as though a trapdoor has opened in the bottom of a cloud and water is pouring out. The good news is I got a great pic of it. The bad news is the storm followed me home and by the time I got back I was soaked!
Witnessing this cloudburst got me thinking about the nature of outages and incidents that can create data loss for businesses. When we think of preparing for disaster we often think of the big things – fires, floods, hurricanes, etc. Statistically, however, data loss is much more often caused by small “cloudbursts” of trouble. Sooner or later, everyone who manages an IT shop has to deal with them, and there are hundreds of different kinds, from system issues such as corrupted databases to “ID10T” issues like accidentally deleted files. Nightly backups are all well and good as a base-level form of protection against these small but potentially crippling forms of data loss, but are nightly backups good enough?
This is a good place to introduce (or review) the related concepts of RPO and RTO:
- RPO, or Recovery Point Objective, is the point in time to which you will have to return in order to get a good restore of your data. For example, if you are backing up nightly, your RPO is potentially as long as 24 hours because the last backup copy of your data is from last night. In this example a data-loss incident at the end of a business day could potentially cost you an entire day’s changes because your latest backup is from last night.
- RTO, or Recovery Time Objective, is the amount of time it will take you to recover from a data-loss incident. RTO is frequently overlooked when thinking about data protection. Think about it; if you lose data at the end of a business day, how long will it take you to recover it? If it takes you until the end of the following day to complete the recovery, you’ve now lost two days changes. One of the dirty secrets of the traditional backup business is that, although backups can be slow (sometimes taking all night or longer), recovery jobs can be even slower.
So, what to do to insure that data is backed up in such away as to minimize both RPO and RTO? It helps to know that not all data is created equal. Business data is usually a mix of transactional data such as databases, email, etc., and file data such as Office documents, pictures, PDFs, etc. Transactional data tends to change quickly and file data typically does not. Therefore the RPO for each type should be different, and the best way to deal with the differences is to treat them separately when backing up. Because transactional data changes more frequently, you may want to consider backing it up at much higher frequency than file data, giving it a shorter RPO. Treating transactional data and file data separately has another benefit – smaller backup jobs. Instead of running one big job once a night, you can run several smaller jobs. Why is this significant? Because a smaller backup job creates a smaller backup file, which is quicker and easier to restore, thereby reducing RTO.
So now we know that in order to reduce both RPO and RTO, we should split our backup jobs into smaller, sometimes more frequent backup jobs. Sounds like more work, doesn’t it? Actually not. Most good backup software allows users to schedule automatic backup jobs for specific data sets, so it simply comes down to building and scheduling separate backup jobs. In addition there is a whole new world of storage and backup solutions from vendors such as Nimble Storage, Unitrends and (in partnership) Veeam & NexGen Storage, all of which offer the ability to non-disruptively take frequent “snapshots” of data at intervals as short as every 15 minutes, which would mean a 15 minute RPO. They can also recover data from snapshots extremely quickly; reducing RTO to just a few minutes.
So the key to weathering the cloudbursts of trouble that cause data loss is to find ways to reduce both Recovery Point Objectives (RPO) and Recovery Time Objectives (RTO). Fortunately it’s easier now than it ever has been. Call us – we’ll be happy to review your data protection needs in detail and make specific recommendations.