Sync, Not Backup

You've either had or will have a data loss disaster. Mine came in 1999 when I moved from Phoenix to Boston. Booting my primary PC into Windows Setup, I chose what I thought was the correct partition to which to install the OS, tapped Enter on the keyboard and ... Gone. I wiped out years of email and some other important documents. Although I tried to recover the data using whatever primitive forensic tools I had at the time, they didn't do the job.

Even the most skeptical among us tends to find religion when disaster strikes. I certainly did. My data loss episode triggered a never-ending, compulsive need to back up everything regularly, and the piles of backup media I still have resembles a computer-backup history walkthrough. There's tape, ZIP disk, CD, DVD, DVD-DL, and various internal and external hard drive types, including a crazy set of 1TB FireWire 800-based drives that cost $999 each at the time and were swapped out on my Windows Server box on a weekly basis.

But time moves on and as cloud services become more accessible, affordable, and capable, the mind turns naturally to a hosted solution. After all, why should storage and backup be any different than hosted productivity servers (e.g., email, calendaring, communications), management solutions (e.g., System Center, InTune), or any of the other offerings Microsoft and its competitors are now racing to put up in the cloud?

Actually, as it turns out, storage/backup is quite a bit different. Sure, companies like Amazon with its S3 service and a host of smaller, consumer-oriented solutions (Carbonite, Mozy, and so on) provide companies and individuals with a variety of options for online backup. But as I examine my own usage patterns and consider the needs of business users, it occurs to me that there may be a better way. That is, most of us don't necessarily need a file store in the cloud. What we need is for our data to be wherever we are. We need sync.

A recent development in Microsoft's consumer-oriented cloud services triggered this change in perspective. For years, I've used Microsoft Live Mesh, an always-in-beta incubation project, to synchronize my important data—documents related to work, favorite photos, and so on—between my various PCs and a cloud-based Live Mesh "web desktop." This web desktop provided up to 5GB of storage and essentially acted as offsite backup. Because it was available via any web browser, I could theoretically get at these important files from anywhere in the world if I had to.

Live Mesh was fast, reliable, and efficient. I used it over the course of my previous three books, including Windows 7 Secrets, during which time my co-author and I shared book-related documents over Live Mesh. It worked wonderfully.

This month Microsoft is beginning the transition from Live Mesh to Live Sync. Live Sync is based on Live Mesh but with some important differences. The company is connecting the service to Windows Live SkyDrive, which provides an ample 25GB of storage space, but is paradoxically lowering the amount of content you can sync to the cloud from 5GB to 2GB and not raising the limit as one might have logically expected.

My initial reaction to this was negative. But as I moved my content from Live Mesh to Live Sync and thought about how I used this service, it suddenly occurred to me that I had never once even used the web desktop. And since Live Sync can sync a virtually unlimited amount of content between PCs in peer-to-peer mode, the 2GB limit isn't really an issue. It applies only to what gets synced to the cloud.

Playing devil's advocate, it's pretty clear that Microsoft isn't yet interested in hosting cloud-based backup. And in a blog post describing this transition from Live Mesh to Live Sync, the company noted that the costs associated with such a service would be daunting—indeed, untenable—when scaled to meet the needs of the hundreds of millions of customers it has. Furthermore, Microsoft found that almost no customers were even using an appreciable part of that 5GB of Mesh cloud-based storage anyway. So this isn't really a problem for most of its existing customers, just as it wasn't for me personally.

The value of peer-to-peer document sync can't be overstated. Many times I've started working on a document on a PC in one room, gotten up, walked into another room, sat down at a second PC, opened the same document, and gotten right back to work. All of the edits and changes I just made at the first PC are present, and no work is lost. Peer-to-peer sync provides a way for me to work on documents on the road and have the changes synced back to my desktop PC at home in real time. In other words, it's what many people are really talking about when they think of backup. It's really about replication, about storing important documents and files in two or more physical places so that if one machine is lost or stolen, or otherwise inaccessible, the data lives on.

Now, sync doesn't replace traditional backup. There will always be needs, including regulatory needs, for data to be backed up and protected in more traditional ways. If you think about how Microsoft's latest version of System Center Data Protection Manager (DPM) works, it's hard not to imagine a more pervasive cloud piece to this tiered backup solution (beyond the already-present single vendor solution through Iron Mountain). Some data needs to be backed up locally and easily accessible. Some needs to be semi-permanently archived to tape. Some needs to be in the cloud or elsewhere offsite but accessible if needed. And a working set of data needs to be synced, something that can occur side-by-side with whatever backup solution you're using.

In the cloud era, there will be concerns about security, data breaches, and performance. But as with any cloud concerns, these issues will be addressed over time as the economics of not maintaining terabytes of data locally will become more and more of a selling point. Just as IT departments will need to wrestle with shifting job duties as more services are hosted offsite, we will need to formalize how and where our important data is hosted and made available.

Looking around the Microsoft ecosystem, you see many examples of replication. There's the Drive Duplication technology in Windows Home Server that I suspect will make its way to future Windows Server and Windows client versions. There's Active Directory (AD) replication between domain controllers (DCs), folder replication in managed environments, and so on. These and other similar technologies all point the way to how we can and should manage our data going forward—and that's true for both individuals and companies of all sizes.

How we get there will depend largely on the technologies we adopt. Generally speaking, I think that cloud backup won't be as important as people may now believe. More critical is that our data be made available when and where we want it. And from what I can see, that's not backup. That's sync.

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish