Question: “Is it advisable to migrate Windows user data when refreshing or replacing their computer?” This was asked in the context of two scenarios:
- Migrating an existing Windows 7 machine to Windows 10
- Refreshing a hosed-up Windows 10 machine to a known state
For scenario 2, I recommend the “Reset this PC” feature in Windows 10 1607.
For scenario 1, here’s my long-winded response: Avoid allowing your environment to get into this scenario in the first place.
I compare this with how I treat my phone. My photos, contacts, emails, bookmarks, apps list and settings are all stored in the cloud. The only items I’d lose if my phone were lost or stolen would be SMS threads. Not a concern. My phone, to me, is 100% disposable. I don’t sweat losing anything valuable if my phone were crushed under a truck wheel, dropped in a bathtub, or doused in gasoline and ignited with a blow torch (although, that would be kind of cool, hmmm).
Desktops and laptops should be no different. Now, I’m not saying “the cloud” is the primary solution here. ANY centralized repository will suffice as an “off-device” storage point.
Because of the following potential faults:
- Migrating data and settings is costly. It takes time to design, build, test and maintain for an organization. The more variations in the environment, the more work is often required on the back end. Even when a process/system has been in place for a long time, that doesn’t justify it’s existence (i.e. “because that’s the way we’ve always done it” = space shuttle O-ring maintenance processes too)
- Having data stored locally is a liability. Period. Ask any attorney.
- Having data stored locally is a security risk. Ask your InfoSec person (drink additional coffee first)
- Having data stored locally adds need for disk encryption, RMS and DLP controls (more time and cost).
- Having data stored locally incurs additional storage capacity and performance requirements on each device (depending upon the nature of data formats)
If a device is lost, destroyed, or stolen, you will need to address…
- the potential lost of a vital copy/version of each document on the device.
- the potential leak of private/sensitive information on the device.
- the additional downtime incurred from replacing the user’s device as compared with providing them a “vanilla” device which immediately restores their access to all documents they need to do their job.
Is it always possible to implement this client-centric data model? No. There are a variety of reasons, one classic example is local processing requirements, such as CAD, media authoring, music production, etc. (high-bandwidth or high CPU demand, particularly when extensive graphic performance are a concern).
However, over the last ten (10) years of consulting and FTE work where I’ve been involved with deployment processes, I’ve only seen a few that really needed it (mentioned above). The majority don’t really need to support local storage of business data.
The reason I bring this up is that I often see a lot of hand-wringing, project investment, and focused effort on applying a “solution” to the wrong problem. This is just one example. I have plenty of others to share if you’d like me to.