Inside Onedrive

Ever since Microsoft gave away storage for a product I would buy anyways, I have been working to move ~1TB of storage to the cloud and have encountered many of the limitations of the service and learned a little bit of the technical underpinnings.

I suspect when the folks who created OneDrive imagined the service they thought pictures. And pictures have a reasonable size (1-2MB) and a small number of them. They did not imagine using one-drive to backup entire multi-100GB file systems.

And that impedance mismatch has been tough.

OneDrive uploads a small number of files at at time (2-4). OneDrive scans the entire FileSystem to find new files.

All of this will get fixed over time, I am sure although things are now kind of rough.

One of the more interesting things is that I have gotten some insight into how OneDrive is built.

OneDrive has four core elements

1. Skydrive.exe that is the sync engine that actually copies the data to the cloud
2. WSearch the search engine for windows doubles as a way for onedrive to keep track of the files on the file system.
3. The use-of-stubs to manage offline files and provide the illusion of a single file system. Possibly the only good use of hierarchical storage management in the history of the snake oil known as HSM.
4. A pretty UI

All of these technologies have been around for many many years and OneDrive is really a repackaging of all of them.

OneDrive, as a product, has the property of something that was cobbled together over time without any of the architectural integrity of competing products like DropBox or Google drive. This probably reflects the ambivalence Microsoft had towards cloud services. I am encouraged to read for Windows 10 Microsoft is working to improve Onedrive significantly.

One of the challenges for a company like Microsoft is that building a product that has the feature set of DropBox is easy, but building a competing product is a completely different can of worms. A competing product requires a deeper level of engineering than the cobbling of re-purposed technologies that the current Onedrive product is.

Microsoft’s decision to embrace DropBox may reflect that reality.