Cheap Modern Backup Strategy for Home Users and Startups

Saturday August 02 2014 - , , , ,

I’ve never liked the idea of online backup. Specifically, because it is billed by the Gigabyte, this acts as a massive disincentive to back everything up. Backup services typically cost in the order of £1 per Gigabyte. Sometimes it is more or less, or sold in chunks of 10 Gigabytes or some other multiple. Nevertheless, there is a direct correlation between the amount backed up and the cost. This results in users backing up as little as possible and certainly omitting the operating system and anything that may be perceived as a ‘system file’. The result is less like a backup and more like an off-site copy of a few important files. Anyone who has ever had to recover a crashed system from the ground up will understand what a huge mistake that can turn out to be. There is a great deal of data that users never think about as being ‘their files’ which will never be protected in a pay-per-gigabyte online backup; this is only discovered when it is too late. In my view, backups must include a full snapshot of the entire system. Online backup turns out to be a poor compromise that leads to a false sense of security. To explain why, I need to explain a bit of backup terminology.

What Makes An Effective Backup?

Backup TapeAn effective backup strategy depends on two things: the backup resolution (how often files are backed up) and the backup horizon (how long old files are kept in the history).

Backup Resolution

The lower the backup resolution, the more information you stand to lose in a system crash. If you back up once a day, then your backup has a resolution of one day. If your system crashes just before the next backup runs, you’ve lost a whole day’s work. If you back up every hour, you can lose a whole hour’s work. The best possible backup resolution is that every file is backed up every time it is changed.

Backup Horizon

The backup horizon determines how long you can go without noticing a problem and still be able to recover the file. For example, if you accidentally delete a file and don’t notice your mistake until a week later, if your backup horizon is shorter than a week then that file will be gone from the backup history. You might be editing a cash-flow spreadsheet and notice that last year’s figures are missing. Where they present in the document when it was backed up a year ago? If your backup horizon is more than a year, you will be able to recover the old document and check.

The Trade-Off

Ideally, you want a high resolution (back up often) and a long horizon (keep the backups for a long time); the trade-off is the cost of storage and the time taken to perform the backups and manage the media.

Traditionally backup has been in the form of magnetic tapes because they were arguably cheaper and supposedly more robust than hard drives. Typically some sort of rotation scheme would be adopted such as Grandfather-Father-Son, to balance the need for a long backup horizon against the cost of purchasing and managing a large number of tapes.

How Does Online Backup Stack Up?

The problems, as I see it, with online backup are as follows:

  • Typically there is no full-system image because the online storage is too expensive.
  • It takes a long time to back up, especially over a slow internet connection. Backup resolution is therefore inherently limited and typically is 1 day.
  • The backup horizon is typically the same as the resolution: 1 day. Again, due to cost-induced downward pressure on storage capacity, even if there is an option to keep a backup history, it will be short in order to save money.
  • Typically, there is only ever 1 copy of any given file (short backup horizon, poor redundancy).
  • If your internet connection is down, you can neither back up nor restore anything.
  • If your backup company stops trading, your backup goes with it.
  • While backup can be optimised (only backing up changed files), a restore of all files can take several days and several attempts.
  • Most people never consider what happens when they need to restore their files.

The Rise of the DiskDropbox

I have always felt that the perceived benefits of tape storage and online backups were largely an illusion, but whether or not you agree with my position it is certain that in recent years  the widespread availability of large, cheap hard drives has turned backup on its head. Various tape-disk hybrids have been tried over the years, but hard drive based backup is now commonplace, for good reasons:

  • Hard drives are fast, which lets you have an extremely high backup resolution. Some modern backup approaches dispenses with the concept of a fixed resolution altogether and back up every file whenever it is changed.
  • The backup horizon is limited only by the available storage capacity which can be many months or even years in some cases. I have one customer that has a hard-drive based backup going back more than 4 years.

Most operating systems have some sort of disk based backup built in these days – but there are problems with this approach.

  • No off-site copy unless you use multiple drives and take one off-site occasionally
  • One backup drive which can fail. A lightning strike could toast your hard drive and your backup drive at the same time.

OneDriveNevertheless, despite these problems, there are compelling advantages to hard drive based backups. For businesses, it may be acceptable to keep several drives and rotate them off site, but that would not be a viable solution for most home users and even many micro-businesses.

The Future is Cloudy

The availability of cheap (sometimes free) cloud storage that is synchronised in real time is adding a new dimension to the available choices when considering a backup strategy. This is still an evolving market place but file storage in the cloud is becoming cheap (sometimes even free) and easy. Privacy concerns may rule this out for some businesses, but for the majority of businesses and home users, this represents a very effective and cost effective facility.

The big problem with online backup is that you pay over-the-odds for your storage, you have a fairly low resolution (usually 1 day) and you probably have a horizon of 1 backup. It starts to look like a synchronised copy of a subset of your important data. Cloud storage is cheaper and synchronises faster, in near real time. So why are we paying over the odds for a system that only synchronises once a day? Why aren’t we just using cloud storage? A good question! Let’s look at the pros and cons of cloud storage.

  • It’s cheap – sometimes even free. I currently have a 28 Gigabyte OneDrive and a 50 Gigabyte DropBox, neither of which is costing me a penny. Click here to get yours: OneDrive, DropBox.
  • It’s fast – files are synchronised in near real-time as they change
  • It’s distributed – your files are replicated across all your devices as well as keeping a copy in the cloud that can be accessed from any internet connection.
  • It’s error prone – if you make a mistake on one device, that mistake will instantly replicate to everywhere. Depending on the service you use, there may or may not be a file history available.
  • It’s not a backup! Backing up your entire system this way would be expensive and probably not work too well.

 

I tend to use services such as OneDrive (formerly SkyDrive) and DropBox to keep a synchronised copy of my documents in the cloud. I have a 28 Gb OneDrive and it isn’t costing me a penny. I have a 49.25 Gb DropBox that I built up by referring friends, which also isn’t costing me a penny. That’s almost 80 Gb of cloud storage that isn’t costing me a penny. So how can we integrate all this storage into a backup strategy?

My Recommended Backup Strategy

The ‘backup’ strategy that I am now espousing for users and businesses without a server is a hybrid strategy based on partly local and partly cloud-based systems. Let me explain.

  1. Get a free or paid OneDrive or DropBox account (if necessary, build up your free storage by referring to your friends and associates until you have enough for your needs). Store all of your documents in your OneDrive (or DropBox) folder and  have them synchronised to the cloud and all of your devices.
    • This gives you redundancy should your device fail, because there is always a copy on your other devices. If you have a single device, there is always a copy in the cloud.
    • There is always an off-site copy if your device (or backup) goes up in flames).
    • You can still work with your documents (via your web browser) while a restore is in progress or if you can’t get to any of your devices.
  2. Get a large hard drive that is at least double the size of the total used space on your computer. The ‘sweet spot’ in terms of price is currently about 2 Terabytes so there’s no point getting any less than that. As soon as you get the drive, use it to create a ‘system image’ or ‘recovery image’ onto the drive using the software built into your OS.
    • This gives you fast recovery in the event of a complete system crash.
    • Never underestimate the amount of ‘invisible data’ stored on your computer, things like user accounts, passwords, favourites, log files, cookies, email and so on.
  3. Keep this drive plugged in (or put your system image away in a safe place and use another drive) and enable the ‘file history’ or automated backup features built into Windows. This varies depending on your version of Windows, but will give you:
    • The best possible backup resolution. In Windows 8, the File History feature essentially backs up every changed file as it is changed. In Windows 7 and earlier, this will be in the form of a scheduled incremental backup that will run at least once a day.
    • A long backup horizon. The file history is kept and only overwritten when the drive is full and the space is required for new backups. This gives you your long backup horizon dictated only by the capacity of the drive. Modern software is very smart about how it uses the storage; I have a customer that has full backups of several laptops going back more than 4 years using only 1 Terabyte of backup storage.

This strategy gives you:

  • good redundancy (you always have an up-to-date copy in the cloud and on your other devices);
  • good accessibility (files always accessible via the web);
  • reasonably fast recovery in the event of a ‘bare metal restore’;
  • no ‘single point of failure’. If your backup drive fails, you have your online copy; if you accidentally propagate a mistake online, retrieve a copy from your local backup history.
  • fair to excellent backup resolution (depending on your version of Windows);
  • and an excellent backup horizon, from a few weeks to several years, depending on your backup drive capacity and how much ‘file churn’ you generate.
  • It’s cheap. A one-off £50 payment for a 2 Terabyte external drive will last you for several years.
  • It’s completely automatic. No more remembering to back up.