A few years back, it became very hip to talk about continuous data protection (CDP).  CDP is a backup job that always runs. It is always monitoring disk writes and anytime there is a file saved, that file is immediately backed up.  In a lot of ways, it is great.  As soon as a change is saved or a file is committed, boom, the file is backed up.  However, it tends to be very resource intensive.  It tends to be expensive.  For very important data, this can still be quite useful, and worth the overhead.  Cash registers, a doctor’s laptop, the computer of a struggling screenwriter who left his wife and kid back in Kansas so he could go to Hollywood and get a movie made and move back home as a hero but who has been distracted by the cheap drugs and constant parties and for a moment forgot his own name after trying smack at a club on Sunset and was thrown out by the bouncer after vomiting on the stage, but the adreneline of being ejected and scraping his elbows on the cement cleared his head and shone a light on an idea that has been sitting right in front of him the whole time so he rushed home and has been typing a fully formed story for 19 hours straight – these are all good candidates for CDP.  Mirroring a database is also another example of CDP.  But for most applications and for most people, a scheduled regular and periodic backup is sufficient.

Now we are going to talk about the following types of backup jobs: Full, Incremental, and Differential.  Each of these terms can be used to describe a local or a remote job.

A Full backup is exactly that, entire files are backed up.  A Full backup can stand alone and represents everything you have selected for the job.

Differential jobs build off of completed full backup jobs.  Differentials compare the current state of a file to the state of the same file in the full job.  If there is a difference, the file is backed up again.  Every file that has been changed since the full is backed up.  Over time, a differential job can swell to be very close to the size of a full backup…as more and more files have been changed, they must all be backed up.  Restores are also complicated by the requirement for the original full plus the differential.  An incremental is similar, but only the differences between the current file and the previous backup are recorded. 

Incremental jobs also build off the initial full.  The files are compared, usually at the block level, to the original file in the full backup.  Any differences are recorded…but just the portions that are different, not the entire file.  The next time the incremental is run, the files are compared to the state at the last run time, and again, any differences are recorded.  So, if a job runs every 12 hours, only the changes in the last 12 hours are recorded. Restores can become complicated, because the software may need the original full plus several incrementals to reconstruct a file.  The big benefit to incrementals is that any given job can be pretty small…the full is full size, but all subsequent jobs just represent recent activity, so they are pretty small in most cases.

Let’s start by suggesting that there are 2 major types of backups, local (onsite) and remote (offsite).

Local backups are great for protection from lost or deleted files and rolling back to previous versions.  If you are backing up to a device other than your primary hard drive, it can protect you from hardware failure.  Local backups tend to be fast and convenient.  restores from a local backup can also be quick.  Local backups may not protect you from theft, loss, or catastrophic incidents.

Remote backups protect your files from all 5 of the situations you may encounter, deletion, versions, loss or theft, hardware failure, and ideally from catastrophe.  They operate at the speed of your internet connection (so, much slower than a local backup routine) and tend to be more complicated to setup and in terms of restoring files. 

A blended approach, using both local and remote backups is often the most flexible solution.  It can cost more than either alone, but gives you the benefits of both.  At The Backup Plan, we recommend that all companies do remote backups, and if their demands justify it, then also retain a local backup.

Next time, we’ll discuss some additional concepts and terms.  Until then, save early and often.

There are several reasons to back up files.  Each demands a different approach.

1) Protect against file deletion.

You will delete files.  Someone else will delete a file that is important to you.  Your IT guy might empty your trash absent-mindedly while fixing your printing issue.  Your cat might jump onto your keyboard and “accidentally” hold shift and delete at the same time.  I think we can all agree that at some point files will be deleted.

2) Preserve versions of files.

Sometimes you accidentally save over a file.  Sometimes, you do a lot of work and think it looks great and purposely save the file, then you show it to your collegue who you are in love with who will never notice the way you dreamily gaze at his/her exquisite smile, and that person suggests it looked better yesterday…hope you have a backup. 

3) Loss or Theft

At some point, sleep deprivation catches up to even the most seasoned traveler, and a laptop is left behind on a plane, or in a cab, or at a massage parlor.  You may have parked the rental car in a public lot to stop and check out the surf, but the guy who waited until you walked away and popped your window just scored a lightly used MacBook Air.

4) Hardware failure

The tape companies love telling you that it isn’t a matter of if your hard drive will fail, it is a question of when.  And, assuming you use your computer long enough, they are right.  Tapes fail, too, btw. Every component in your machine can fail, and some of them take out others.  Sometimes you were hitting on a hot doctor from cardio and thrilled about the super clever jokes you were firing earlier over coffee, and you turn on your computer to post a Facebook status update, but nothing happens, and you realize that the hug you were about to virtually hgih-five yourself about meant that your laptop spent 30 seconds pressed up against the artificial-heart-stopping magnet in her labcoat pocket.  Shoot.

5) Catastrophe

Natural or man-made, disasters destroy computers.  Floods, fires, and wolverine attacks all destroy computers. 

I am willing to bet that 2 (or more) of these 5 have happened to everyone reading this. Can you think of other reasons to backup files?  Can you think of a reason not to?

If you don’t backup your files, you are either dumb or lazy.  That is right, you are:

a) dumb.  You are a moron.  Big words make you blink blankly; they boggle your brain.

b) lazy.  You know you should back up all those files, but, maybe you will get around to it.  After this next episode of LOST (which is what your files will be if you never back them up).

c) An optimist.  see a)

d) an employee – and you assume your work takes care of this for you.  See b)

So, yeah, I am sure there are other explanations.  However, there isn’t really an excuse.  When I was a consultant, and somebody’s laptop picked a pope, most of these panicked people asked me right away about their pictures, then music.  Oh, yeah, then the work files.   Imagine losing all of your digital pictures.  Your wife (or husband) would be so mad!  It is just like your house burning down, except without the maimed cat and your wife isn’t hitting on any firemen.  So, yes, ok, you are right, it is slightly better than your house burning down, but you still lost all your stuff.

I agree that most employers should make this an automatic thing, individual employees shouldn’t have to waste their time backing up files.  It should just be done.  For that reason, I am a big fan of the managed service provider model.  In this model, there is a managed service, it is a tested, functional, easy to spin up service, and an IT consultant or shop or department uses this to provide backup services to their own customers and users.  This scenario puts the management and oversight in the hands of a trusted and professional IT person.  The actual back-end is managed by an experienced backup service provider.  Sounds good!

So, what am a getting at, what is my point?  Basically, I started The Backup Plan in 2005 to address this need.  The Backup Plan has grown to be a managed service provider, with resellers using our services, taking advantage of our infrastructure and experience, to provide backup services to their own customers.  The Backup Plan has chosen the best of the vendors out there, combined their offerings with some best practices, documentation, and rules, and created a solution that addresses the needs of quite a few companies as well as end users.  But that is not the point of the blog.  The blog is here to explore and share ideas about backups.  What are good things to do?  What are the new players?  What technologies could change the game?  Please, check me a bit if I self-promote too heavily.