A tale from the IT crypt, thanks to another thread on re-inventing grep…

Back in the pre-Y2K era of the NT Affinity boondoggle, a development team was tasked with creating the “DIGITAL OpenVMS Disk Services for Windows NT” (NTDS) remote storage system for Windows NT clients.

Imagine creating an NFS server, and creating NFS clients for Windows NT, all allowing access to a hardware storage pool available on some bigger not-NT server.

This product was being newly created too, so a fairly extensive new-development project split across the server-related development work, and the NT client work.

In one meeting early on, I’d pointed out this was all quite literally re-inventing NFS, and so using an existing (or modified) NFS server could probably speed up the whole product release.

Was told no, the server only accessed arbitrary blocks in one (partition) file on the host, so definitely not at all like NFS.

Okay, but accessing arbitrary blocks in one file is a proper subset of accessing arbitrary blocks in multiple files, right?

Was told no; totally different. But never did get any explanation why NTDS was all so assuredly and totally different from NFS.

🙄

The NTDS project wasn’t a particular success with customers though (and a similar outcome with NT Affinity more generally), so shortening the schedule and saving some development wasn’t a factor in that (un)success.

#retrocomputing #TalesFromTheITCrypt #ReinventingTheWheel #openvms