As we introduce our cloud backup software to more and more people we have discovered an area of data protection that is causing some confusion, in this post we shall try and make the differences and implications clearer.
When it comes to data protection, you could say there are the following approaches:
- Do nothing
- Backups
- CDP
- Replication
- Near CDP
Option one obvious is not providing any protection, but scarily is pretty common. Especially if you take the view (as we do) that if you don’t test your data protection system regularly then you may as well be doing nothing at all.
Backups are good, take a copy of all the data you wish to protect and store somewhere. They are generally daily and allow you to restore any file to the state it was in when backed up.
Continuos Data Protection (CDP), is when every change to every piece of data is backed up. You are able to restore a file to any point in time
Replication, is when the data you have is kept in two places and all changes are written to both locations either simultaneously or near to. No restores as such, you have an exact copy of the data in two locations, if you lose one data store you can use the other.
Near CDP, is when all the changes at short intervals are backed up. You have point in time backups to restore from.
Each method (and there are others) has its on pros and cons. Backups for example have a big disadvantage in that they typically are run once a day and as such you can only recover back to the point. If you backup at midnight, and you lose a file at 3pm the next day; when you recover the file from the backup you will have lost all the changes you made during the day. In this example the window of risk for your data is up to 24 hours.
That said, backups are simple and robust. The methodologies for good protection with backups are well understood ensuring that data is recoverable for a long time.
CDP is terrific for being able to recover a file that is lost or corrupted at a known time, you can restore a moment before the error and little (if any) work is lost. Lose that file at 3pm and you can recover it from 2:59pm (or perhaps 2:59.30 even). The window at which you data is at risk is virtually nil. However, CDP requires quite a overhead in terms of processing changes and also needs greater storage requirements to store all changes. Typically CDP covers data only for a specific number of hours or days, otherwise the disk storage requirement becomes considerable. When used in conjunction with traditional backups it can be a good solution.
Replication provides arguably the best protection against hardware failure, specifically disk failures. Most replication solutions will switch from the primary store to the replica automatically if there is a failure. Unlike backups or CDP however, replication is not good for recovering files that are lost or corrupted, this is because all the changes to the storage are replicated to both locations. Lose a file and you lose it on both data stores/disks.
Near CDP, this is very similar to CDP, except that rather than every change being tracked and recoverable from, “snapshots” at specific intervals are taken perhaps every minute or 5 minutes or hourly. This allows you to minimise the window of time that data is at risk but also decreases the overhead required to protect your data.
When looking at data protection solutions you need to consider several factors:
Time to recover files
Windows of risk
Cost
Overhead
Length of time data is protected
Disk requirements
Each method of protection fits into these areas differently and there is a business decision to be made in regard to risks versus investment. For example replication of data will mean that hardware failures should not cause any downtime and users should not be affected. However, if a file becomes corrupted, that corruption will be replicated to both stores.
CDP allows you to recover to a point just prior to the corruption in our example, but the process often adds considerable overhead and does not protect as well as replication against a disk failure.
Backups provide a robust long-term protection of data, typically you’ll be able to recover a file from days, weeks, months and years in the past if required. However, the window of risk for losing data between data can be an issue.
Near continuous data protection is, we are finding a very good balance for most situations. Especially for laptop users, or organisations where use of the C: drive is common. The overhead is less than CDP as is the data storage requirement. The window of risk is also quite small and cost is often lower than the other solutions.
Of course we are biased, our cloud backup software does exactly this, you can have the changes to the files on your laptop backed up at very regular intervals. Typically every 5-15 minutes which when protection work or excel files is often adequate for 90% of data loss. Our software of course also allows the user to restore their own files to any of the points in time, no need for delays by involving the IT team.
Of course the other big thing with cloud backup is that you do not have to worry about storage of the backups. You don’t have to calculate how much disk/tape storage you’ll need as you have unlimited capacity on the cloud. Cost is another factor, buying all the tapes or disks to implement a solution is expensive, especially as you need to buy enough to cover expected future use; with the cloud you pay only for what you are actually using.
We are very happy talk through your data protection plans and strategies and to give advice on what will work for you; we generally find a mix of near CDP and traditional backups is a good mix. It is not often a single solution will address all you needs.
Please do take a look at our cloud backup pages, we have video demonstrations and cost estimates etc. We genuinely believe that it is a great solution that can merge into most organisations smoothly.
