This is a drum that bears beating as often as possible.
Backup your data.
I was reminded of this point again recently, when a friends granddaughter lost all, or at least most, of her collage school work and pictures when the drive in laptop died. And with no backups, I don’t even want to think about what that must have been like for her.
In my experience when I’ve done IT consulting work in the past, backups are one of the first things that get overlooked, ignored, or forgotten about. This is even more true, when the backup requires some level of user interaction.
Proving my point, when I mentioned the above incident to my parents, my mom realized that she wasn’t backing up her computer at work. When it was setup, it was setup so that it could be backed up, but the act of plugging in a USB stick and running the backup command… Well, lets just say it got lost in the cracks.
At the same time, backups need to be taken offline to be truly safe. If you’re backing up to say USB disk, and it’s still connected to the computer, the chance is always there that you might accidentally delete it. Or that one of the crypto ransom viruses, could encrypt it rendering it useless. Moreover, if the backup disk is still attached to power, there’s always the possibility that a power surge or lightning strike could destroy it.
In short, if the disk is plugged in; it’s not a good backup since it could be erased or killed by a power surge. But if the disk is left unplugged, then the act of having to plug it in becomes a hurdle to getting people to actually back things up.
Ultimately, I have 3 points in writing this.
- Remind people to backup their data
- Remind people that are backing up their data to make sure their backups are actually working
- For those that aren’t backing up their data offer up some options that they might find useful to start backing up their data.
With a cloud based service, you install a program on your computer that allows you to upload copies of your data to the cloud service’s servers at scheduled intervals. The service provider then is, in theory at least, responsible for insuring the integrity of the data.
For most people, cloud based backups are probably the way to go.
Once, the software is installed and setup, it does it’s thing on the set schedule and you don’t have to think about it again, until you need to recover your data. That solves the hurdle problem of backing up and to some extent the offline protection problem.
Actually you get a further benefit in that the cloud provider’s data center is probably not in the same city or even state as you are, so if say a major hurricane destroy’s your town, your data is still safe.
On the other hand, cloud backups have some drawbacks, potentially major ones, and especially if you’re a photographer that shoots a lot.
All of these cloud services use your internet connection to upload the data. For most users, this isn’t much of an issue, as most people don’t generate a lot of data. 500 emails of 250 words a piece, is only 610 KB. A 100 JPEGs from a smart phone is only about 400MB. Lets just call this 500 MB per week of new data.
The average internet upload speed — the part that matters — in the US is only 6.3 Mbps. That 500MB of new data created by the average user only takes 10 minutes to upload. That’s completely reasonable.
But I’m a photographer, this is a photography site, and of course I’m writing about the problem of cloud backups as they relate to a photographer.
When I got back from 2 weeks in Alaska, I had 250GB of raw images. At the average 6.3 Mbps, it would take more than 3.6 days to backup. And that is allowing the backup to utilize 100% of my internet connection’s upload capabilities. In practice, doing that means I’m going to severally degrade my ability to use the internet for anything else for the same time.
Even a weekend morning shooting can be 32GB or more of data to deal with. And don’t forget, camera resolutions continue to increase. Also this completely ignores video, which consumes much more space than stills generally do.
Beyond time, there’s a second aspect, data caps.
Pretty much all of the ISPs in the US currently have some kind of data cap in place. AT&T, my current provider, charges me extra if I go over 300 GB — by the end of this month it will be 1000 GB so there’s at least some relief. Even then, you still have to account for normal internet usage, including streaming video content and so forth. But there are a lot of places where internet usage is billed by the byte, where uploading a lot of data is going to be costly.
And this also completely ignores the fact that your data is now in someone else’s datacenter, and more importantly on someone else’s computer in their datacenter. Now I’m not a tin-foil-hater or paranoid that the government is out to get me, but having you data in someone else’s hands does have a privacy and security aspect that should be at least considered.
Personally, the biggest killer of cloud backups for me is simply the internet connection aspect. I don’t even have average in the US upload speeds — which really sucks if you’re trying to upload video content to youtube too. The time it would take for me to back stuff up, especially at a rate that doesn’t degrade my internet access, would be simply impartial for me.
If you can’t ship your data off to the cloud, the only other option is to back it up locally somehow.
There are a number of options when it comes to backing up to a local device. These range from something as simple as using a USB hard drive and any one of the various backup programs to back your computer up to it. To more technically complex options like a NAS (Network Attached Storage) device. To a layered approach that involves backing up to a NAS then backing that up to something like a USB disk, or even up to a cloud provider.
When it comes to local backups, there are two primary disadvantages.
First, it’s local. This means if your home or office burns down and the backup is there, you lose your backup too.
Secondly, because it’s local you’re responsible for setting it up, maintaining it, and operating what parts need regular interaction.
With that said, for me, as a photographer, the volume of data I end up dealing has made local backups pretty much the only economical solution for me. It’s just much easier and much more efficient for me to push a couple hundred gig around over my network and then to a fast USB3 disk than it would be for me to push it up to a cloud provider — maybe when gigabit fiber ever becomes wide spread things will be better.
So what options exist?
The most inexpensive of the local storage options is simple external USB disks. I’m partial to Western Digital’s MyBook drives, but Seagate offers similar options as well. In truth, the brand doesn’t really matter, both options are equally reliable, at least on the large scale, and cost around the same price.
That said, there are a couple of things to remember or consider when using USB drives.
First, actual hard drives that are in these enclosures are typically sourced form the lowest tiers of the manufacture’s product lines. There are certainly executions to this, but that’s generally the case.
Drives in these tiers tend to favor power efficiency over performance. As such they may do things like aggressively spin down to save power — the cycles of spinning up and down and up again can eventually lead to a drive failure.
Moreover, since the margins on these drives are tight, they aren’t always the top tier quality drives to start with. That said, in my experience, reliability is far more important in drives that are used daily or all the time, not ones that are connected once a week for an hour or two to back up to.
And I don’t want to make it sound like I’m trying to fear monger here. I use USB external disks, WD MyBooks specifically as part of my backup routine. Mine have been in use for a couple of years now, and I have a friend who backs up the same way. Out of our combined 6 or 7 drives, I can only think of 1 that failed, and it was dead on arrival.
The way I mitigate the potential for having a backup disk die, and therefore leaving me without a backup, is to have more than one backup disk.
Specifically, I use 2 disks that I rotate on a weekly basis. Yes, this strategy means if I lose a backup disk for some reason I lose the latest week of my data, but I’ll still have the data from 2 weeks ago, which is definitely better than nothing.
Two disks is my default recommendation for most people backing up to USB disks.
For people with only a single computer and who need a local solution, USB external disks are my recommendation. They’re relatively inexpensive, there’s no need to bother with a whole lot of extra complexity, and since you’re only dealing with one computer they’re pretty straight forward to get working right.
The one problem with USB disks is that you don’t want to leave them plugged in all the time. The biggest side effect of this is that it’s much harder to have an automated tasks that runs to back you up. To make automation easy, you need a system that is online yet separate from the computer it’s backing up.
The solution to this problem is to put your backup target on your network.
There are a huge number of ways to accomplish this, but the one that’s easiest to get started with is a Network Attached Storage device, typically called a NAS.
These are basically compact integrated computers that have integrated easy to use interfaces for setting up network file shares, and managing the hard drives you plug into them.
Most of the mid and higher tier models use standard SATA drives like you’d put in a computer. And since they don’t usually ship with drives you can tailor the capacity of the NAS to your needs and budget.
With a NAS, you set up your computer, or computers, to back up to a shared drive on your network. Since the NAS stays powered on all the time, this can be set as a scheduled task to run in the background when your computer is on — and even being used.
Beyond a NAS
As I noted, NASes have one major shortcoming, they’re always online. Whereas a good backup solution should really be offline when it’s not being used to make it much harder to be accidentally deleted or damaged by power surge.
The solution to this is to add a layer to the overall backup solution.
Most modern NASes have the ability to back themselves up to either a USB drive or to one of the myriad of cloud storage providers.
This means you can backup your laptop and desktops to the NAS, automatically when they’re on, and then use the NAS to backup everything else to a more secure medium.
Of course, this system potentially requires some more user intervention; as you have to plug a USB disk into the NAS. But it limits the user intervention to dealing with a single point, the NAS, as opposed to having to deal with each computer individually.
My Personal Setup
I used a tiered backup solution, similar to what I described in the NAS section. Only instead of an integrated NAS, my network backup target is a full Intel E3 Xeon based Linux server. I would point out, I didn’t build the server just for backups, those are only one of the many roles it serves.
The computers on my network then backup to the server automatically.
My MacBook Pro running Time Machine backs up on whatever schedule Time Machine determines. On my Windows desktop I use a couple of scheduled tasks, one that does a file level backup of just my data every evening, and the other that creates a full system image every Sunday afternoon.
The server itself, with all the backups of the workstations, is backed up to a pair of rotating USB disks every Sunday night. This was scheduled so I could plug the backup disk in around dinner time, and unplug it when I get up Monday morning and everything would be complete by then.
Admittedly, there are a great number of implementation details I’m leaving out here, but the overall idea is that I only have to mess with plugging in one drive into one computer, once a week. Everything else is happening automatically without me having to mess around with it — which insures that I don’t forget to back up important pictures or something I just downloaded.w
I would be remiss if I didn’t spend at least a few minutes talking about backup software. This, more than anything, is probably the biggest potential minefield in the whole process.
Broadly speaking backup software can be separated into two types; software that backs up files, and software that makes images of whole disks. Each of these methods have their pros and cons.
In my own systems I use both methods for different reasons.
File Level Backups
File level backups can be as easy as copying files in File Explorer or Finder, but most frequently file level backups are handed by some kind of file sync program. If you’re a Dropbox user, you have some idea of what you can expect. Files are synchronized (copied or deleted) with the backup target on an individual level. If you edit a file, the file gets copied to the backup target, but only the file you edited.
I’m not talking about JPEGs or raws here, what I’m talking about is a technique where the software makes a complete copy of the drive that it’s backing up. An image of it at the binary level.
Image backups have a number of benefits over file level backups, principally because they can record data that file level backups would miss. Things like the boot sectors or the hidden partitions of a disk. These hidden bits of data are often necessary parts of making your computer boot and otherwise work.
Moreover, image backups are typically very easy to use to recover a whole failed computer from. You don’t have to reinstall the OS and your programs and then copy back the data and hope everything works. Instead you can just “reimage” a new drive from the backup “image” file and your computer shoot boot right up.
The biggest downside to image level backups is actually in restoring single files form them. Not that it’s impossible, but because the image is typically a proprietary format, you can’t just copy the file back form the backup to your computer.
Software to Look Into
I’m not going to make any recommendations here, in no small part because what I may be comfortable with is not necessarily going to be ideal for someone else. That said, these are some of the solutions I use and some of the easier options I’ve seen.
Both Windows, since at least Windows 7, and Mac OS have built in backup options. Mac OS’s system is called Time Machine, and it’s pretty darn easy to setup.
The exact details for Window’s built in backup system is going to vary from version to version. The following link is Microsoft’s instructions for Windows 10.
The two file level backup programs that I find work the best are rsync on MacOS and Linux, and Robocopy on Windows. Both of these only copy what’s changed, not everything, and both of these are included with the OS so you don’t have to go install them.
I admit, this has been a very high level overview of backing up your data. In practice there are a ton of variables and options, far more than I can really effectively address. However, my main point here was to reiterate the importance of backing up your data.