CIMdata consults with many end-user clients on their PLM and simulation strategies which, of course, turn out to be all about data management and process. In spite of clients having spent millions of dollars on solutions, it is incredible that when we bore down with the question about “How does collaboration and data sharing really happen?” the clients invariably get this sheepish look, stare down at their shoes, and mumble “shared drive.”
It is amazing how much organizations rely on local network shared drives, and how out of control they are. I am reminded of “The Tragedy of the Commons,” which is about how shared resources can be exploited and destroyed for perceived individual benefit at the expense of the common good1.
I am surprised that there seems to be no available solution to what I call “The Shared Drive Problem.” A decade ago (at GM) I wrote my requirements for a Simulation Data Management (SDM) solution. On reflection, I named it “Windows File Explorer on Steroids.” 10 years later, I have still not seen a viable solution.
People need to easily store and share data, and to collaborate. Companies deploy shared drives, which quickly become an unmanaged, out-of-control fiasco. Shared drives are the worst possible solution, except for all the others.
What’s the problem? It’s pretty simple: shared drives must be manually managed, and administration requires privileged (root, sysadmin) IT access. While some groups have disciplined processes around naming conventions and projects, it’s still manual. Unless you have the decoder for the naming convention, the data is not searchable or understandable. After a year or two, even the original authors of the data have no idea what is there.
What to do? What is “File Explorer on Steroids?” Here are my thoughts.
- Only files need to be managed.
- Data and user administration should not require system-level privileges.
- Administration is by groups, roles, and projects, not predominantly by named user.
- Metadata fields are easily defined and include free text and drop-down selections.
- Versions are tracked and automatically generated.
- Configurations and assemblies are allowed and managed.
- Data status is allowed and coupled with visibility. (Preliminary, proposed, released, archived …)
- The user view is configurable. In other words, columns in the file explorer view can easily be selected by the user.
- Simple change control (user privileges) and change approval workflows are available.
- Data can be immediately sorted by metadata fields.
- Data can be searched by text on contents.
What’s so hard? Let me know what you think!
I am also indebted to my CIMdata colleague Tom Gill for pointing out that others have called this a “digital landfill.”