You can push the Content-Addressable Storage timeline back much further
Indeed, to early database days, before "objects" were even talked about.
This is a terrific object storage history map from Silicon Valley object storage guy Philippe Nicolas*, who has put together a spreadsheet detailing the history of content-addressable storage (CAS**) – otherwise generally known as object storage. I have heard so many odd things about suppliers and technologies in this market …
Indeed, to early database days, before "objects" were even talked about.
Thanks for the chart and history. However I wonder what, in the meantime, has happened to its complement: Content Addressable Memory? Back in the 80s, I remember it being used in support of some CPU operations. There was also academic research into its use in bulk. What's the state of the art today? (Yes I've just had a squiz at Wikipedia, but I doubt that it represents the current status.)
Might just be the academic pedant in me nit picking at terms, but content addressable storage is generally different from object storage in the storage world I live in, to the point that I almost never hear the terms used interchangeably.
Certain implementations share flavors of one another, but the key is in the name: content. A content addressable storage system is one where the content is hashed (or a portion of it hashed) to for the purposes of producing an address that serves the dual purposes of implicit deduplication and naming of the data. Object storage, on the other hand, tends towards using the *name* of the object as a way of locating it, by hashing it and dumping it on a consistent hashing ring with object servers owning parts of the namespace.
Content addressable memories have been constructed in hardware. I believe several chips were produced i the early 1980's to do this. I worked on a project at GEC where we were contemplating using one to implement a switch (case) statement that would execute in a single cycle, rather than step through a table. We had a choice of a chip made by GEC or one made by a German company.
However, the project was scrapped in favour of using a VAX off the shelf - as it already had software, and we would have had to port Unix to our weird hardware (which also had a transputer array).
Surely Object Oriented Database Systems are relevant here? If so, then those date from the 1980s and I used one of them about 18 years ago. It was called ODBII and was an academic project that Fujitsu bought the rights to. It was then pushed quite a bit by ICL after they were bought by Fujitsu, which is how my employer ended up using it.
Ridiculously late to the party.... oops!
Even more entries in the mix now in 2021 with ONTAP and OneFS having S3 adaptors added to them, and in the world of software-defined I see Nutanix too has a s/w defined object solution on their platform (lots more features than I expected) - plus Azure's Blob is another public cloud object solution to look at (Azure mentioned in article but not Blob specifically)
Also found this which is interesting https://www.gartner.com/reviews/market/distributed-file-systems-and-object-storage