“We Didn’t Invent Disk Arrays But RAID”

We got yesterday the following email from Randy H. Katz, with cc to Garth A. Gibson and David Patterson, the three authors who presented 30 years ago the paper A Case for Redundant Arrays of Inexpensive Disks (RAID) at the SIGMOD conference, following the opinion of Tom Gardner, treasurer and webmaster at IEEE Silicon Valley History Committee who affirms that they didn’t invent RAID (see sources below).

I was amused to wake up and read Tom Gardner’s piece in your newsletter this morning.

We are discussing whether or not to respond.

Note we never claimed to have invented disk arrays, as our paper clearly indicates, but the concept of the fundamental trade-offs between redundancy, reliability, and performance as embodied by the RAID taxonomy we introduced.

Almost 4,000 Google Scholar citations to the original RAID paper (as well as three test of time awards from three different professional societies) suggests it has had some impact in shaping the evolution of the disk storage industry.

And not just the academic world, but the industry, found this to be a foundational contribution.

This is what we are really celebrating after 30 years!

It is definitely fun to revisit a very exciting time in our career from many years ago, but fortunately for each of us, it was not the only thing we have done in our careers!

We asked Katz further if the University Berkeley’s inventors got royalties from their invention? The answer:

As an academic group, our goal is primarily to have impact, through influencing the thinking of others.

Many in the industry have told us that they read the RAID paper and immediately understood the future of the storage industry.

An engineer at Compaq told us that our paper influenced the first EISA controller, which they designed, to be a RAID controller.

Also Dave Hiltz told us that NetApp was formed in large measure to exploit the technology trends described in our paper.

So to answer your original question, there are NO Berkeley RAID patents, no royalties, and we shared our ideas freely to all, including Dave Gordon at ArrayTech, IBM, StorageTek, DEC, Tandem, and a host of other companies.

By the way, I thought your original article was a fine summary of the RAID paper. Conceptual breakthroughs can be more influential than engineering artifacts. I should point out that we built RAID-II as a 144 RAID disk network attached storage controller that became fully operational in 1992, and it was quite possibly the largest such system when counting spindles at the time.

Patterson wrote on his side:

The answer is no.

I also didn’t get any royalties from the RISC vendors.

Not sure why this question is interesting.

Read also:
RAID Is 30-Year Old
Invented by David Patterson, Garth A. Gibson and Randy Katz at University of California, Berkeley in 1987
by Jean-Jacques Maleval | 2017.11.20 | News
Patterson, Gibson and Katz, University of California, Berkeley, Didn’t Invent RAID !?
Emailed Tom Gardner, treasurer and webmaster at IEEE Silicon Valley History Committee
by Jean-Jacques Maleval | 2017.11.28 | News