But how does desktop and laptop data backup differ from enterprise data backup? What types of products are available to use for desktop and laptop backups? And are cloud backup services a good option? W. Curtis Preston, executive editor at TechTarget and independent backup expert, answers these questions and more in this Q&A. His answers are also available as an MP3 below.
Table of contents
>> Why should people care about backing up their laptop and desktop data?
>> How does laptop and desktop data backup differ from enterprise data backup?
>> What are the best practices for backing up your laptop and desktop data?
>> What types of products are available for desktop and laptop backups?
>> Should you use cloud backup services for desktop and laptop backups?
>> What are some of the common problems people run into with desktop and laptop data backup?
I concur with what you said in the opening about laptop and desktop data being just as vulnerable -- it's actually more vulnerable. We put server storage systems some level of RAID. So if a single disk drive goes bad, it doesn't take our data. Whereas usually with a laptop or desktop, you have one disk drive, and if that fails then the entire system fails. That's the first reason people should care about laptop data backup and desktop data backup. Second, the data at the edge, which is laptop and desktop data, is the most recent data of a lot of companies that have the policy of "we don't back up laptop and desktop data." They just expect the employees to copy it up to a server on a regular basis. Those companies are fooling themselves because it's typically not going to happen -- it might happen sometimes, or it might happen on a semi-regular basis, but not all the time.
And so when I listen to people and they tell me that they have that type of policy, I truly think that they don't care about the data in their company. Because data on the edge seems to be the most recent data, it also tends to be the more valuable data. So to not back it up at all is simply crazy. And to rely on manual processes by the people at the edge to back it up is also foolish. You need an automated process for desktops and laptops, just like you need for the rest of the data center.
There are a couple of main differences between laptop data backup and enterprise data backup. First, with laptops you can't always assume that it will have the same IP address. So you can't reach out to it the same way you can with typical backup software that says "we're going to back up 'Elvis' now." With typical backup software, it goes out to get "Elvis," does a DNS lookup, finds its IP address, and talks to "Elvis." It doesn't quite work that way with laptops. In other words, you've got to turn things on their ear and have the laptop talk to the backup server when it can. And it has a network connection, but you cannot assume the network connection is there. You also cannot assume that the IP address is the same -- it could literally be all over the place. That's the biggest difference between laptop and enterprise data backup.
The other thing is that it's hard to back up a laptop while it's being accessed. Also, if a user notices the backup process is going on, they're probably going to shut it off while they're working. However, unlike with servers, you can't assume that laptops and desktops will be on at night. It's very common for people to shut off their laptops, power-off their systems or hibernate their PCs as they go home. So again, you can't count on that. Hopefully you can educate your users so that they stop doing that, or you have software that deals with that issue, where it actually backs up throughout the day. So there are definitely some challenges there.
The final thing that I can think of, especially in the case of laptops, is that the amount of bandwidth that's available to transfer backups is often very small. I use broadband wireless when I'm on the road, and sometimes that doesn't work and I have to use the hotel Internet, and sometimes even that doesn't work -- but that same Internet is what you need to use to backup your data. So you've got to deal with all of those different issues that you don't have to deal with in server backup.
You first have to be aware of all of those issues, and not simply take a server backup product and start backing up laptops. The other is to do the same thing you do in every other area of IT -- you need to evaluate a couple of different products to see what they're like. Do some proof of concept (POC) with a small group of users and get some feedback from them. Because with regular data backup software it's more about you -- you're the backup administrator, and you need to be happy with how it performs and how it works. Laptop and desktop data backup is more touchy-feely and has a lot more people involved. With laptop and desktop data backup, you'll want to know things like when the backup kicks off -- did you notice? Did it bring your system to a halt? And some of the systems will queue up their backups to another part of the disk drive, and then as soon as they get an Internet connection it sends all this data over. So their disk is busy and their Internet connection is full. Then the users aren't happy, so they'll shut down the backup process.
Make sure you try data backup products from a number of different companies that you're thinking about using. Try it with a number of different users and get feedback from them on how it works. Also, you should do the usual testing of backups and restores. It would also be interesting to do some control tests to see if one product is faster than the other. User feedback is very important because you're about to go out and touch all their desktops and laptops. Think about antivirus software for example; we've all grown used to it, but we hate it. But when the antivirus software kicks in, users know they shouldn't shut it off because they could catch a virus if they do. But with backups they might not feel that way. So if it's really annoying they might just shut it off.
The best thing you can do is multiple POCs with multiple products and see how well it works.
Data backup products for laptop data backup and desktop data backup fits in a couple of different categories. From the software perspective, we tend to be talking about continuous data protection (CDP), near-CDP, source deduplication software, or delta-differential software. CDP continuously replicates changes from that laptop. If it doesn't have an Internet connection; it will cache up those changes and send them once it has one. Near-CDP is like CDP in that it's continuously sending changes, but it also has to take snapshots in order for the system to work. So in the case of Windows it's going to take a VSS snapshot, and again, you're going to be continuously sending changes throughout the day.
The other two products are more like traditional backup products, with one exception -- they're never again going to do a full backup. Once you get the first full backup, then the only thing they're going to send each day are deltas -- the new, unique blocks. In the case of delta-differential software, it's only going to send the blocks that are unique for that day and that laptop. A true source deduplication product is going to send blocks that are unique to the environment, meaning that if this user with this laptop downloads the same PowerPoint file that another user downloaded to his/her laptop, and then one of them backups up that PowerPoint file using the source deduplication software, the next person that runs the backup is not going to transfer that file across the network. So if 20 people downloaded the same PowerPoint file, that file will be backed up 20 times. However, only the block-level changes to those files will be backed up each day with both types of products.
So all four of these products have a couple of things in common. The first is that they are incremental forever. You get that first full backup, and then you never again do another one. The other is that the incrementals that are sent either throughout the day or multiple times each day are only the new bytes. It's not full-file incremental. Again, you can't take a regular backup product and just turn it on for the laptop because of how huge some of the files are. For example, if you look at just PowerPoint and Word files, they're really space inefficient and end up creating giant files. If you change one word in a multi-megabyte file, it's going to be backed up again. If you do that with a couple of files, you'll end up filling up the whole pipe. You have to have this sort of block-level incremental concept, and these are the four different ways to accomplish that.
All four of the software products listed above are available from cloud backup services. The cloud provider manages the service and the backup server. They are also responsible for dealing with issues that come up. For example, if they notice a user has been shutting off the client and it's been a number of days since the last backup, they will then notify someone who can do something about that. It's their job to deal with that along with product updates. All of these things are important, so the idea of using a cloud backup service for desktop and laptop data backups is a good one.
Just because you've gone to cloud backup doesn't mean you can ignore that different pieces of software behave differently. You still need to look at different types of software so that when you look at cloud backup services, you can ask them what they use. Some companies such as EMC Corp. Mozy do it all. They're the service provider and they sell directly to the customer. Then there are a bunch of companies that basically use another company's software product, and then they just manage it. So with those types of companies you need to ask them what type of product they're using. And then again, make sure you're testing multiple types of products, not just all of the same product from multiple cloud service providers.
The most common problem is purchasing a data backup product that changes the user experience too much. If you can choose the right backup product, then you can get this job done with relatively little impact, especially if you're willing to train your users to leave their computers on at night. If you can do that, then your laptop data backups and desktop data backups will be a lot easier. But if you can't do that, don't want to do that, or don't think you can do that successfully, then the problem really becomes doing backups without anyone noticing, and that is a significant challenge. Too many people have chosen this route only to have the users revolt. Then the next thing you know it's been a month and nobody's data has been backed up.
This was first published in August 2010