1980 – The world’s first gigabyte-capacity disk drive, the IBM 3380, was the size of a refrigerator, weighed 550 pounds (about 250 kg), and had a price tag of $40,000( $113 thousand in present day terms ), 2.52 GB
1980 – ST-506 first 5¼ inch drive released with capacity of 5 megabytes, cost $1500
Last week, I was contacted by Finis Conner who is best known for his career in the disk industry (more later). I have known Finis since 1980 and we have communicated off an on during this vast expanse of time. We had a drink and talked about the early days of personal computers as well as the future of mobile computing. I have always liked Finis as a person. He has a warm and engaging personality. And he, like me, has dared to put his vision to the test time and time again.
The Story Begins in 1980
This story begins sometime in 1980. I was working at Digital Equipment Corp as a Senior Group Engineering Manager and I was responsible for the low end computer hardware group. I had joined the company in the summer of 1979 when I returned to the United States from Israel. For those that were around and can still remember, my group did the engineering on all Q-bus computers as well as Decmate (word processor).
The Central Engineering organization led (maybe that is too strong a word), by Gordon Bell, had a yearly strategy meeting off site. In those days, such meetings where called “woods meeting” because they happened in either Massachusetts or New Hampshire at some hotel located in a forest area. The senior managers of the Central Engineering Organization would nominate the individuals they thought were the brightest and had the most promise to participate. I was invited to attend. It was a two day event, I think. Outside speakers/presenters were invited.
At this meeting, the legionary Al Shugart spoke. Al had an amazing career spanning 45 years and did more than anyone to create mass storage for personal computer. His career began at IBM (1951) working on Winchester Drive Technology for Mainframes. Later he worked for Memorex, a mainframe disk manufacture. In 1973, he started Shugart Associates which pioneered the floppy disk, first with 8-inch drives and then introducing a 5 1/4 inch in 1976 (the capacity was 87,5k). Al was forced out of his own company in 1974. He moved to Santa Cruz and opened a bar with a friend and bought a fishing boat. Sometime around 1979 Finis got the idea for the 5 1.4 inch drive. His concept was that if you could create a drive in the same form factor as a 5 1/4 inch floppy drive, you could replace one of the two drives that personal computers had with a hard disk. In 1979, he approached Al Shugart with the idea. Together they founded Seagate Technology with only $1,500,000 in venture capital. Seagate went on to become the largest independent manufacture of disk drives.
Back to the Central Engineering Woods Meeting
At that time, I really did not know any thing about disk technology. Digital Equipment had its own disk engineering group run by Grant Saviers and a manufacturing group run by Bob Peffer. I remember that there as a disk called the RL02 which was sold with PDP-11s at that time. It was 10MB. It think it was a 14 inch disk but I am not sure. The disk itself was actually removable which was important. If you had at least two of these units you could back up your disk quickly. Otherwise you had to use DecTape which had a capacity of 258k per tape. Digital, like all computer companies at that time, was pretty much vertically oriented in the sense that it designed, manufactured and sold all the major components including software that made up their computer lines. That meant the company had a memory group, a printer group, a disk group, a terminal group, and an operating system software group. Little did we know at this Woods Meeting, that the structure of the computer industry was about to be turned on its side and Digital would slowly slide into the abyss. I was 35 years old at this time and Al was about 50.
I become one of the first and largest customers for a short time
So Al gets up to talk and educates most of us about disk technology. Then he pulls out a 5-1/4 inch Winchester drive with a capacity of 5 mega bytes that his company was developing. I was blown away with the concept and since I was involved in making small computer systems, my brain lit up with the possibilities. After his talk, almost no one went up to Al. I was surprised at the lack of interest. I went up to him and introduced myself and we chatted for a few minutes. We exchanged business cards and he said he would follow up with me. I think, soon after, I was contacted by someone in his sales organization. A month or so later, I was selected by Ken Olsen to lead the development of personal computers (we did not call them that). Knowing about the 5 1/4 inch disk effected my thinking greatly. Ken wanted me to build a product that only used floppy disks. This was driven by cost factors and because Ken had little understanding of software. I did develop such a product called the Professional 320 but my heart was in a hard disked based product called the Professional 350 which would feature a built in 5 1/4 inch Winchester drive. Both products failed in the market place for many reasons which I have already written about. And in some ways, the IBM PC, which was release a year after I started the CT program (the code name for our effort), proved Ken right but not for long. In 1983, IBM launched the IBM XT which featured a Seagate drive. Apple also used the Seagate drive (but not the electronics) as an option for the Apple III and later the Lisa.
Going outside for a Disk
Soon after Al’s talk, I learned that Digital’s Disk group was developing a competitor to the Seagate drive. It would never had occurred to that group to go outside for a complete drive. I felt that the group would never deliver on time since they started much later than Seagate. So in order to use the Seagate drive, I went to them and pleaded with them to go directly to a 10mega byte drive, committing that I would switch over to their drive as soon as they got it done. I told them that they would leap frog Seagate that way. Later, when Seagate announced their 10 mega byte drive, I persuaded the Desk Group to move directly to 20 mega bytes. You probably have the idea by now that I did not really want to reply on Digital as a supplier. I had to do the same thing with the printer group but that is another story.
I think it is important to stop for a second and pause on the size of this disk. 5MB is less then the memory used for the average photo we take today with a digital camera. The Seagate drive probably cost end users about $2,000 in todays dollars. Today we can buy a terabyte drive which is 200,000 times as large for less than a $100. In other words, it would have cost more than a million dollars to have a mega byte of storage in 1980
Now it cost about ten cents.
Sell, Manufacture and then Design
The drive was not without its problems. While most product development started with design, then went to manufacturing, and then was marketed and sold, Al had reverse this by selling the product, then manufacturing and finally getting around to designing it 🙂 Eventually the product worked as specified.
Seagate Goes Public with my help
Sometime in Aug. of 1981,I got a call from Finis, Al Shugart’s partner asking if I could place an order for something like 1,000 units which was valued at a million dollars. Al wanted the order on the books because the company was going public in about a month. I said I would do my best and was able to get the order processed in time. Of course, such a conversation these days would not be permitted by the SEC. Finish, tells me that he cut a deal with the buyer from Digital at the Crow’s Nest restaurant in Santa Cruz. He wrote out his offer on a napkin and agreed to sell Dec the 5MB drives over three years starting at $525, then $500 and finally $475 in the third year.
The only major company to actually ship the Seagate 5MB drive was Apple. They bought the drive but not the electronics from Seagate and sold about 20,000 units at a list price of $3,499 (they only paid $400 for the drive). Seagate only manufactured 25,000 units that year. Finis deal with Steve Jobs on the Apple product (Steve was not the CEO at that time-Michael Scott was).
Later IBM would become the biggest customer with the introduction of the IBM PC XT. Finis signed the deal with Don Estridge who unfortunitly died in the crash of a Delta flight in Dallas in 1985.
The Pro 350 did not ship until early 1983 much later than originally planed. Just after that, I left Digital to become President of Franklin Computer, an Apple II clone company that did 80 million dollars in sales in its first year of operations before being crushed by Apple in a lawsuit. By April, I was out of Franklin, pushed out but happy to go. We did have a clone of the IBM XT underdevelopment but never released. I saw Finis from time to time at industry meetings. In 1985, he started Conner Peripherals.
Finis goes on to start Conner Peripherals and pioneers the 3 1/2 inch drive
Finis left Seagate and started Conner Peripherals in 1985. Soon after, that company merged with CoData which developed the first 3 1/4 inch disk. The company did very well and had revenues of more than 1.3 billion dollars by 1990 and was eventually acquired by Seagate in 1996. After than Finis started two more companies, Conner Technology which manufacture 3 1/4 inch disk in China but was not successful, and BluStor where Finis is still CEO. Al Shugart died in 2006 at the age of 76.
iPod Changes everything with a 1.8″ drive (2001)
I am not going to go into this now but clearly the introduction of the iPod with a 10 gigabyte hard drive (20 years after Seagate introduced the 5 1/4 inch 5MB drive) was a major discontinuity in the evolution of computing and broke the hold of desktop/notebook computers.
The Future of Disk Capacity and the end of the consumer disk industry
There is no reason to think that disk capacity will not continue to follow Moore’s law and grow exponentially for, lets say, at least the next ten years. I just bought a new Mac which has the new Fusion Drive which combines a one terabyte hard drive with a 256 GB flash memory. In ten years, we could expect to see two orders of magnitude increase. That means I would have a 100 terabyte hard drive and a 25 terabyte flash memory. But what would I do with such a capacity especially since my internet connection will probably be between 100 and 1000 Mbits? Movies are the content that uses the most capacity. An HD movie can be easily compressed into 7 GBs on average. Lets make it easy on ourselves and say that the most we will need for the average movie will be 10 gigabytes. That means we can store a 100 movies in one tera byte or 10,000 movies on the hard disk that I could theoretically have in 2013. But that makes no sense when I will be able to stream any movie from the internet with the same resolution. I would never buy 10,000 movies and even if I was so inclined, I can’t imagine ripping or illegally down loading 10,000 movies (although I have one friend who would try to do this). I certainly could not use up the storage with photos. So what is going to happen? I think consumers will slowly move to all solid state (flash) memory. We will see the consumer hard disk go the way of the tape drive, floppy disk, and optical disk (DVD). There is no longer a disk in any IOS device like the iPhone or iPad. My Macbook Air does not have a hard drive. I am not sure what will happen with the disks the will be used to store things on the cloud.
Hi Avram. My version of the history re DEC’s early involvement with Shugart/Seagate is a little different than you recall. I met with Al and Finis in CA shortly after they formed the company. (Finis took me for a ride in his Cobra – very cool!). I saw that the 5.25 hard drive was so important to the future of the industry, that I convinced Gordon Bell that I should invite them to the Stratton Mt engineering fall retreat to show the not yet engineered ST506 drive. No outsider had ever participated. Al & I had a relationship that went back to his days at Memorex and he and I discussed the formation of Shugart Associates and its first product (floppy) and then the Shugart/Seagate 5.25 drive.
The disk buyout group that worked for me sourced the drive for the PRO and Rainbow (John Rose). I thought Rainbow shipped first, but could be convinced otherwise. At least Rainbow was the first x86 system with a hard drive.
Rick Rosa was the purchasing manager that negotiated the contract. At that time I ran Disk Engineering and Bob Puffer was responsible for disk manufacturing and Rick was in his organization. After the Stratton meeting, I recall Al and Finis coming to Maynard in the middle of a blizzard to meet at Maynard’s La Petite Auberge for dinner. When Al sat down at the bar after being two hours late, he ordered a double bourbon, chugged it and had another. Bob, being a blue blood conservative New Englander, said to Al, “I thought we were having a business meeting” and Al replied, “my doctor said I should have two of these every day.”
It is true that at a later meeting Finis, Rick and I finalized the deal and it was written out on a cocktail napkin. When Finis was CEO of Conner, he had this framed and posted it on the back of his coat closet door. Since Conner and Seagate were fierce competitors at the time he didn’t want Conner customers to see it. Of course, it was your and John Rose’s commitment to build desktop systems with the drives that drove the deal.
There was no internal development to compete with the first Seagate drives. Much later we did build larger capacity 5.25s, and in the meantime bought many many 5.25s from several suppliers. The Disk Buyouts Group was managed by by Ed Barron, and Ed East did all of the performance and reliability testing when the various suppliers offered up next generations.
Grant Saviers, DEC 1968-92
Grant, thank you so much for your comments. I really appreciate them. It is amazing how we can all have different memories of the same “facts”. I try my best to remember things accurately. I checked everything with Finis. I guess I should have checked with you as well. I can certainly imagine it was you that brought Al to Stratton Mt. (I forgot that name). I remember John Rose. I don’t think the Rainbow shipped first and certainly not with a hard drive but that I can check with Barry Folsom. I can only imagine the meeting between Al and Bob. I also remember Rick Rosa’s name. Did not know you were at that meeting at the Crows Nest. Vah Erdikian will recall some of this. I will check with him. Now that part that I will dispute is about the internal 5 1/4 inch development in your group. You should know but I recall many meeting about this and having to get agreement with people in your organization to get the go ahead. I think there was someone named Phil Goldman involved. I will check with Vah. If this is just a dream, I will admit it of course but I don’t think so. Once we agreed to go outside, we had the full cooperation of your organization. Thanks for that.
My only interest her is to document the past correctly but with a bit of humor. We lived through a remarkable time. Thanks again for you comments.
In 1968 when i first became a security analyst covering the electronics industry. I had almost no background with computers expect I wasa deciple oF Gen George Deroit, who thought American Resarch and evelpment co founded DEC and owned 50% of the company.
DEC could and be said to have destroyed American Research and development (the first Public Venture capital firm) because he total market Value of American Research and Development should ar a discount to it DEC’ss Holdings. All other of American Research and Devopment activity was ignored by wall street because correctly even if a new investment was spectacularly successful it could not mount up much compared to DEC which was gowing 50% a year (without cooking the books as SDS, Control Data, Honeywell and RCA were doing.
I started studying IBM and the Seven Dwarfs ( DEC Was not even a dwarfs) Xerox had to get into the computer industry so it purchased Scientific Data Systems for over 1 billion , SDS claimed to have a 100 million in Revenues (about 50 Million with honest accounting) and it was selling at over 100 time earning (actually they were loses not earnings), Once SDS disappeared from the Dwaf list DEC was promoted to a dwaf with many of SDS’s investors taking Their money from SDS and putting it in DEC.
I ment with General Deriot and spent a day discussing the industry. He was very interested in my views of my because he was interested in fresh views.
I told him that I thought that the computer industry really had almost nothing to do with computing but was really just a data processing industry that was able to count how many widget were sold at a faster pace and who was was behind in the payments for the widgets.
So if we are counting widgets what we needed was vast storage capacity to track where these widgets where they were going. n . So in this EDP industry Peripherals were much more important than processors..
I said that the computer portion of EDP was very advanced and pheriferals were all bassed on spining whells and Drums.
I then noticed that the computer companies like IBM actually made very little money on computers and enormous amounts on peripherals. We both thought that in a few years all of these wiring things would be replaced by a completely different kind of device.
The trouble with genius is you you see too far ahead, Now spinnig things are beginning to disappear and the elctronics industry has become pretty much of a Mass consumer market. Very few companies will be able to keep up how fast the consumer moves and how low the margins are for seeling them hardware just like cars.
Avram great job in committing this to paper (cloud storage). thank you!
Avram, thanks for capturing this important early ‘personal computing’ history!
I frankly don’t remember when we added a hard disc to the Rainbow – John Rose may have a better recollection.
What I do remember is the Friday night in Seattle when I told Bill Gates that DEC was going to do a PC-compatible Rainbow. Bill doubling over in ‘pain’ as it struck him that his OS wasn’t the chokepoint he thought it was. Bill always fought to control key chokepoints and this information was clearly a set back for him. Not that it really mattered in the end.
A few pieces of PC history trivia:
* DEC’s keyboard is the defacto PC keyboard standard still to this day
* Bill Gates sent IBM to Digital Research to get their CP/M OS. Gary Kildall went flying that day. IBM went back to Microsoft and Bill found a OS for $50K that became DOS. Marilyn Darling, former Digital Research person, Bill and I shared these stores at a dinner in Boston
* Windows keyboard shortcut’s were a contractual commitment between DEC and Microsoft
* A meeting with Bill Gates at DEC resulted in a decision that the mouse would be two buttons instead of one button (Apple) or three buttons (workstations)
* The Rainbow was one of the first in the industry to have a universal power supply.- frown upon by other engineering managers at DEC at the time due to the extra $6s in product cost
* The Pro lead the way for making various interconnects to a PC easy to add/remove
Now back to rotating disc drives. In 10 years time, a 100 terabyte hard drive and a 25 terabyte flash memory won’t – as Avram indicates – have ANY relevance to devices consumers/users have with them, BUT it will matter big time to all the Internet Data/Cloud Centers running Big Data applications. What consumers will have is a 0.5 terabyte flash drive, which will be in their iPad 10, along with 25 4K/QD movies that can quickly be ‘Airplay’ed to any – and I mean – any nearby display. Think of the iPad 10 truly being your DVR with the first 10 minutes of 100s of shows cached on the iPad making it easy to entice you to watch an episode or see a 4K/QD commercial relevant to your in-the-moment intent-to-purchase query.
There is a new enterprise 2.5 UX paridigm coming as well, but that is off topic. For another time.
Barry, Universal Power supplies in the Rainbow PC – well almost – I was in Australia, doing regional new product launch support at the time. Our first prototype Rainbow arrived, and the Product Manager grabbed it, and went to plug it in, it had a 110V only Power Supply, and our 240V supply “let the smoke out”. As I recall, we reverse engineered the Power Supply (a metal box type), building a circuit diagram and working out what voltages/currents it was delivering from that. We shopped around, and found that there was a PDP8A power supply (which was a flat circuit board), which had the right voltages – It fitted in by being braced with cardboard guides diagonally where the Box PS would sit.
It worked, and I put a sticker on the back, saying that as soon as it had finished at the Product Launch, it was to go to the Repair Centre, who by that time would have real PC100 Power Supplies.
Thought no more of it, until about 18 months later, in another city, pop the top off a Rainbow back in from a Customer Loan, to see the jury -rigged Power Supply.
I recall a discussion (over a BBQ at his place in Colorado Springs) with Paul Massiglia, who Grant will know well, and who later was on the Raid Council amongst other things. It would have been about 85 or 86 – We spoke of the concept of having storage subsystems that we internally intelligent, and would adapt based on usage. i.e. a big black box, with some solid state storage, some spinning disks, and even some high speed tape cartridges (I have always been a proponent of Hierarchical Storage Management). The concept was that the “system” would look at the files, and determine say, that the payroll files were only needed on a thursday night, and could be staged, even back out to tape until needed, but actually brought back in ahead of requirement, where as say, the authorization files would stay in SSD.
(Regarding SSD (Solid State Disks) I recall a DEC OS Group, who were having problems getting sufficient compile/link build runs of the test versions of their OS done during the evening “offpeak” window – By switching to a mechanism for “Pre-fetching” sources, and dumping binaries to SSD and having other modules deal with them from there, they we able to go from 2 builds a night to 8 builds.
(The same basic concept also led to something I explored when designing an internet Data Centre for a Telco – predictive caching – We were exploring Caching accelerators, for internet traffic – The basic concept being that the first look up of say the Monday Morning files of the Wall Street Journal would keep a local copy in the network wide cache until the TTL for those pages was hit, or a cache cleaner process determined that they had been replaced. What we looked at was looking at the profiles of the demand and supply of the data, timewise. If the first person in the office (or network) to check the news, was at 7AM, and the newspaper “published” at 3AM – we would pick common ground, and prefetch the data at 5AM.
The concept of “Predictive caching” was also used by Bob Wyman (ex DEC, now at Google, and the godfather of publish/subscribe technologies (per his time s CTO at PubSub, a startup that specialized in this space) to optimize the google experience.
But, history aside, I think that these things can tell us a little more about where we need to go with future storage – Whether it is silicon or spinning, is neither hear nor there to the user. The real issue will be the management of Personal Terabytes of information (I remember my first day at DEC in 79, I was given 2.5MB RK05 pack, and told that I would be unlikely to find enough stuff to fill it…).
Self tagging data files (i.e. rather than storing a phonecall010112.wav – File xyz has been determined to be a voice conversation between the owner, and parties a,b and c (determined from a central voice print database server) – It is analyzed, transcribed, keyword data is, metadata-ed and indexed, so that ultimately one can ask one’s “file system” for a transcript or recall of the phone conversation with a,b,c and self where we talked about disks and cupcakes, will most likely be the way of the future. (One could argue that the Mumps Database/OS filesystem – code is data is code was their mantra – is part of a step down that path).
One of the biggest challenges that this will engender is the ownership and security of the information. I think that current security paradigms are inadequate (and potentially we need to revert to some of the compartmentalization models of Bell, Lapadula, Biba et al, who designed much of the initial DoD Rainbow book stuff, back in the 70s. Whether the individual, the other parties, or the employers retain “ownership” or access to, information will no doubt need to change. The Apple Itunes model of “we see that you have the signature of say, ‘John Lennon’s Imagine’, so we will automatically authorize you to retrieve it from our cloud, is a step in the right direction, but it also brings back into play the “longevity” of the requirement – renting vs buying music, and indeed, whether that right to access can then be on-sold or gifted..
I spoke with an EMC person a few years back, who was quite proud of a client that had storage requirements growing by half a petabyte a year – they of course, wanted to keep it all spinning – in practicality they were just too lazy to get rid of the stuff that they didn’t need. I suspect that they had also reached the point where they didn’t know what they had on disk, but kept it all just in case. equally, I have seen time and again, where the same piece of data if often kept in a dozen or more locations within an organization – a dangerous precedent for data integrity, but that’s another story for another time.
People couldn’t imagine what they’d do with a gigabyte at home. Then a terabyte. Believe me, by the time petabyte disks become available, you’ll need one. Nothing is more limited than our imaginations about what we’ll need in the future. 10GB for a movie? Not likely when they’ll be 3D, if not holographic. As for my iDevices, they’re all full. Give me a disk, please.
Aah. it is so not refreshing to see such a naive view of the world. Yes, software expands to meet available hardware (that was a lesson we learned a couple of decades back). But …the demand tends not to be exponential, because smart people (like those here) end up doing smart things in terms of compression, and other technologies to change the load characteristics of the application. While it gives the perception of adding another dimension and ergo an order of magnitude – 3D is really about bluffing the binocular vision of humans and is not particular data intensive in the delivery end of the process. Holographic – barely worth speculating about.
MisterQ, agree with you. Sorry Ira.
in fact my company and I introduced the first hard disc on MSDOS systems I had to sell everyone the idea that having a hard disc changed the way you used a dos computer. It worked. We went public but walked into the pC clone problem and a bank took us under.
We did get our next computer of the year award for doing a really portable hard drive system that let you take the drive home to a second computer that it plugged into.
But MR heads were happening and we sold the drive company to WD.
My reason for this note is i have done the thing that puts spark back in the PC and was trying to find conner to see if he was interested in getting involved. Maybe you would contact Finis and ask him to send me an email so we can get back in touch.
In return if you are interested i will show you the idea. It is storage related and tightly patented.
It also drops the price of SSD in half.
If you can help I would appreciate it.