Nerds 2.0.1: A Brief History of the Internet. Vol. 2 - Serving the Suits (1998)
by Robert X. Cringely
Videos in this documentary
Nerds 2.0.1: A Brief History of the Internet (1998) is a three-hour documentary film written and hosted by Mark Stephens under the pseudonym Robert X. Cringely and produced by Oregon Public Broadcasting for PBS. This three-part sequel to the well-received Triumph of the Nerds, recaptures the same trendy style so effectively used by Robert Cringely. A sequel to Triumph of the Nerds, Nerds 2.0.1 documents the development of ARPANET, the Internet, the World Wide Web and the dot-com bubble of the mid- and late-1990s. It was broadcast two years prior to the collapse of the dot-com bubble.
Vol. 2 - Serving the Suits continues the journey through the web with a look at some of the key figures who helped develop the Internet. Nerds 2.0.1: A Brief History of the Internet, Vol. 2 - Serving the Suits traces the growth of networking from academic use to the business world. The first "nerd" to make a million was Bob Metcalfe, who mastered the art of connecting computers in a network. Then, companies such as Sun, Novell, Cisco, Microsoft, and 3Com began their rapid rise. Cringely talks to some of the principals involved.
In the 1980's, personal computers became a common fixture in homes and offices. Supplying business with computers and software grew into one of the biggest industries in less than a decade. Soon, networking became a profitable business for engineers previously restricted to networking mainframes.
At the turn of the decade in 1970, big business was taking another look at the impact of computers. During the sixties, computers were mostly restricted to the air-conditioned back rooms of major research centers, but smaller and more affordable computers were starting to show up in offices. Xerox was afraid that their products, mostly copiers and typewriters, would disappear from a paper-less office. Xerox realized that they couldn't be surprised by the office of the future if they were the ones to build it.
Bob Taylor Bob Taylor was hired to help build a new research center for Xerox, where the best minds would do nothing but forge the future of computers and technology. Taylor wanted the new center near a major university where new ideas were already being created, so he broke ground near Stanford University in Palo Alto for Xerox's Palo Alto Research Center (PARC). He recruited some of the researchers he funded before, including Jerry Elkind and Severo Ornstein from BBN.
The atmosphere at PARC was electric, where some of the best technologists and scientists worked on their own dreams of the future. Their research was funded from the growing profits of Xerox, which was quickly approaching the size of IBM, therefore project budgets were nearly unlimited.
Two of the researchers at PARC, Butler Lampson and Chuck Thacker, worked on a project to put a computer on every person's desk. The computer, called an Alto, was really the first small personal computer that could be used in a business environment. They didn't really know how useful Alto would be, but it was their job to discover the future. Several of the Altos were built and given to PARC's staff to see what they would do with the computers.
Computer Communication Compatibility
At about the same time the Alto was appearing on office desks at PARC, Bob Metcalfe finished his Ph.D. thesis at Harvard and was hired by Bob Taylor in 1973. Metcalfe's thesis work was an efficiency study of the ARPAnet and the ALOHAnet, but he had never seen the ALOHAnet work firsthand. So, one of the stipulations of his accepting the job was a three-month visit to Hawaii before he reported to work. In Hawaii, he studied the ALOHAnet in more detail and came back to the mainland with a notebook full of data and ideas for improving the network.
Norm Abramson Norm Abramson, the builder of ALOHAnet, created a protocol looser than that used on ARPAnet, something similar to a party line on a telephone. Because there were no isolated network lines, everyone had to share the same pathway. Sometimes two or more network packets would be sent at the same time and none of them would arrive at their target locations uncorrupted. To correct for this problem, the receiving computer would check for errors in each packet and would only send back an acknowledgment if it received a good packet. If the sending computer didn't get an acknowledgment, it would wait a random amount of time and then send it again.
Bob Metcalfe had some ideas for a better protocol. Instead of just sending out a packet, the transmitter would first listen for any traffic and only send it out if no other computers were sending packets. It's similar to the way polite people talk in a group. Instead of interrupting other people, they wait until there's an pause in the conversation. The protocol was called CSMA/CD, which stands for Carrier Sense, Multiple Access with Collision Detection. That just means multiple computers could connect to a single cable and would try to have only polite conversations.
At PARC, Metcalfe worked with Thacker, Lampson (the two inventors of the Alto), and David Boggs on building a computer card for the Altos. Bob Metcalfe Their two test machines were named Michelson and Morley, after the two early 20th century physicists who disproved the popular belief that an invisible "ether" filled the universe.
After several months of building and testing, Metcalfe's team successfully networked the two Altos together. They needed a name for the network protocol, so they continued the ether joke and named it "ethernet." Soon more Altos were connected to the ethernet and then different computers were added, creating a local internet inside PARC.
Although Metcalfe had a useful product, at that time PARC didn't have an easy mechanism to move it to production and to market. He decided that he would produce and market ethernet himself. He was not a businessman, but he was a researcher, so he decided to research the process of building a successful business and then do it. Using the Directory of Western Venture Capitalists he scheduled breakfast, lunch, and dinner meetings with anyone willing to talk to him. To start, he wasn't looking for venture capital, but advice and information.
After nearly a year, while perfecting his product, he found backers and opened his new company. For a name he combined three words that described his business: Computer Communication Compatibility, or 3Com for short.
While growing up in India, Vinod Khosla, dreamed of starting his own high-tech company and becoming rich like the founders of Hewlett-Packard and Intel. Vinod Khosla When he came to this country to study business at Stanford Business School, he got a chance to help draft a business plan for a technology company named Daisy Systems. The business succeeded, and Khosla cashed-in and walked away rich at the age of 27.
He still wanted to start his own company, and in 1982 he found a product he knew he could sell, a workstation. A graduate student named Andy Bechtolsheim studying computer science at Stanford had already built a name for himself on campus because of the workstation he designed. A workstation is a computer more powerful than a personal computer but small enough to sit on a desktop. It had built-in networking, because Bechtolsheim knew researchers needed it, and it used the Unix operating system, a nonproprietary operating system developed at Bell Laboratories.
Bechtolsheim designed the workstation to fill a void in the computer science department. He was frustrated with the aging time-sharing system used at Stanford, and he thought everyone could get a lot more work done if they had immediate access to a computer and still be able to exchange data. Andy Bechtolsheim His idea was to license the new technology to companies that could build it and then get the computers from them. He had already licensed the workstation idea to over seven companies before Kolshod approached him, so he thought Kolshod just wanted another license.
However, Kolshod told him "I don't want to do that. I want the goose that lays the golden egg. I don't want the golden egg." He wanted Bechtolsheim to join a partnership with him to build the workstations for sale. It was difficult to convince Bechtolsheim because he wanted to stay at Stanford and complete his Ph.D. work, but he eventually agreed.
To help with the business side, Kolshod recruited a fellow Stanford Business School graduate, Scott McNealy. McNealy had no experience starting a company, but he was excited about the prospect. Now they had two business people and a hardware expert. All they needed was a software expert to cover all facets of the product. The choice was easy, Bill Joy.
Bill Joy Bill Joy was a young computer science professor across the bay at Berkeley. Joy had quickly gained the reputation as one of the best computer scientists in the country at the age of 27, the same age as the other three. Having worked on implementing parts of the ARPAnet, Joy had plenty of networking experience. He had also rewritten the Unix operating system to incorporate TCP/IP and released it as Berkeley Unix (BSD).
Khosla found more venture capital for his new company (initially called "VLSI Systems") from John Doerr of Kleiner, Perkins, Caufield and Byers (silicon valley venture capitalists). Doerr had known the initial trio from Stanford, and he understood how big a Unix workstation would be in the research and business market.
With a few million dollars in the bank, a prototype motherboard, and a networking operating system, the company began building their workstations under the new name of Sun (in homage to the Stanford University Network). By 1988, they passed $1 billion in sales and were the second fastest growing computer company in history. It was a success story rarely matched since.
Success from Failure
For every successful high-tech company there are hundreds of failures. Novell is a story of both failure and success. In 1982, Novell Data Systems was a small computer company located in Orem, Utah near the Wasatch Mountains. It had found some success building computer systems for local businesses, but the good times quickly took a downturn. They were soon finding it hard to make payroll, and they had to auction off office furniture to pay their employees. Some of the company's investors called in Ray Noorda to see if he could turn the company around.
Drew MajorThe same year as their demise, four contractors were hired to write software to network CP/M computers to a common "disk-server." The disk-server split up it's hard-drive into virtual drives, one for each of the computers networked to it. Drew Major and the other programmers started working on a new idea for a "file-server." Instead of a private section of the disk, each networked computer would have access to all the files on the server. Then people could share data across the network. There were some problems to work out, like two people simultaneously accessing a file, but Major and the others recognized a big seller.
While they were working at Novell, they saw a demo of the first IBM PCs available in Utah and bought one right away. They saw great potential in the IBM computers, and decided to try out their new ideas on it. They were contract workers at Novell, so they weren't too worried about the eminent layoffs - they could just find other contract work while they worked on their file-server.
Ray Noorda When Noorda toured the business to see if he could save it, he visited with Major and the other contractors and liked what he found . He immediately saw the potential of the file-server. Noorda decided there was something to save, but the business would have to change and drop everything except for Major's file-server.
During the next year, the four contractors (still not full-time employees) worked on their idea and eventually built a network operating system. They called it Novell Netware and shipped it as soon as it was ready. There were lots of bugs in the code, but it was on the market first.
Netware quickly became the dominant networking software for PCs, competing against a weaker operating system packaged with 3Com's ethernet cards. Noorda came up with a novel marketing plan -- essentially flipping 3Com's strategy. Novel began selling ethernet cards at near cost as long as the customer bought Netware to run on their networks. 3Com thought they were being ripped off, but Novel became their biggest customer for ethernet cards. It's a strange business.
Riding the Bear
Microsoft probably benefited the most from the growing IBM personal computer market. Because IBM didn't have the skill-set to write an operating system for their desktop computers, they chose to hire Microsoft, still a small Seattle company, to do the work for them. After the first MS DOS was released, IBM decided to keep the business relationship and let Microsoft maintain the operating system. Microsoft was pleased to serve, because they knew IBM ruled over the computer business and they would go wherever IBM went. Microsoft called it "riding the bear."
Bill Gates In the mid 1980s, IBM was working on a new personal computer, and they wanted a revamped and more powerful operating system to match the capabilities of the new computer. They contracted Microsoft to work with them to design and write the second generation operating system, OS/2. This new operating system added several new features, such as a graphical interface with windows and a mouse. However, they didn't want to add any networking for OS/2. IBM was still in the mainframe and minicomputer business, and some of the management didn't want to the personal computer division competing with the rest of IBM.
Bill Gates and the rest of Microsoft didn't see it that way, but no matter how much they argued and pleaded IBM would not add networking to OS/2. Microsoft just watched as Novell took over the networking market:
"Around '83 and '84 and certainly ... by '85 Netware was reaching critical mass. And so Microsoft felt really like there was a huge missed opportunity. In fact, I remember some memos Bill wrote, in sort of '84, '85, '86, where he said, you know one of the biggest disasters for the company is that ... is that we have no assets in networking or very weak assets in networking."
In late 1989, Bill Gates made his first attempt at buying the competition, Novell. He called David Bradford of Novell and offered to band together and compete against IBM, but Bradford refused the offer. So, Microsoft went to 3Com instead and offered to band with them and compete against Novell. Metcalfe was already frustrated with Novell, so he agreed with the arrangement. He believed that with Microsoft's relationship with IBM, they had an opening to the largest market, IBM PCs.
"What Microsoft failed to tell us was that their relationship with IBM was falling apart at that moment. Which came as a big surprise about three days after we signed the deal."
founder of 3Com
Later in 1989, Microsoft announced OS/2 LAN Manager and the headlines proclaimed that it would control most of the networking market by 1991. But that forecast fell short and Novell still controlled the majority of the market in 1991.
Bob MetcalfeAt 3Com, because their half of the partnership was falling short of their goals, the board of directors decided a change of management was in order. Metcalfe suggested that he should be CEO in 1990, but the board chose Eric Benhamou instead. Metcalfe quickly resigned from 3Com. He's now a content gentleman farmer in Maine and writes a weekly column in Info World.
After inventing Ethernet and founding 3Com he does have some advice to prospective entrepreneurs:
"It helps to have good parents, and then it helps to work really hard for a long period of tie and go to school forever, and then it works to drop quite by accident into the middle of Silicon Valley, where you're swept up into an inexorable process of entrepreneurship and wealth generation, and you pop out the other side with a farm in Maine. I hate to oversimplify."
Only one year after Sun was formed, another company destined to make a big splash in the networking world was also based on work done at Stanford. Cisco, founded by husband and wife Len Bosack and Sandy Lerner, started experimenting with connecting their two detached networks located in two different buildings on campus. With the help of two other Stanford staff members, Bosack and Lerner ran network cables between the buildings and connected them first with bridges and then routers.
Bosack's router was really an updated IMP, which would transmit only the traffic that was meant to get out and accept only the traffic that was meant to get in. Bosack and Lerner designed and built routers in their house and experimented using Stanford's network. When word got out about their routers, other universities and research centers asked to buy them. Bosack and Lerner went to Stanford with a proposition to start building and selling the routers, but the school refused. Bosack and Lerner knew they had something worthwhile, so they founded their own company and named it "Cisco," taken from the name of the city to the north.
Their house became company headquarters and every room was used for building, testing, manufacturing, or shipping. Sandy Lerner They had no capital, so they charged all the startup costs on their credit cards. However, even without a real sales staff, they started to make a profit very quickly. When they were pulling in hundreds of thousands of dollars every month, they decided it was time to act like a real business, and they needed help recreating their company.
They went to several venture capitalists and made their pitch, but all refused until the 77th:
"At that point I think we were -- Cisco was doing, I think, a quarter million, maybe 350,000 a month without a professional sales staff and without an official conventionally recognized marketing campaign. So it wasn't a bad business just right then. And so I think just for the novelty of it, the folks at Sequoia listened to us."
cofounder of Cisco
Don Valentine was an experienced venture capitalist with such companies as Apple and 3Com in his portfolio. He saw great potential in Cisco and he was impressed with Bosack's and Lerner's enthusiasm. However, he wanted to hire some experienced management to run the business - something venture capitalists call "adult supervision."
"The commitment we jointly made to each other is that we at Sequoia would do a number of things. We would provide the financing, we would find and recruit management, and we would help create a management process. None of which existed in the company when we arrived."
Valentine recruited John P. Morgridge as president and CEO to oversee the company and be Bosack's and Lerner's boss. There was tension between the three from the start, and Lerner claims that the first words from Morgridge's mouth when they met were "I hear that you're everything that's wrong with Cisco." He denies he ever said that, but Lerner never got along with Morgridge after that.
As Cisco grew to a billion dollar business, more management was added and Bosack and Lerner were feeling pushed out. Lerner was always concerned about customer relations, but she didn't get along with the upper management. In 1990, seven of the vice presidents came as a group to Valentine (with Morgridge's knowledge) and made an ultimatum, either Lerner left or they would leave. On August 28th Sandy was asked to leave the company, and Bosack left after he heard. They both immediately sold their two-thirds stake in Cisco, taking about 170 million dollars.
Sandy Lerner now owns a cosmetics line and funds several charities. To get away from California, she bought Ayershire Farms, the 40 room house in Virginia where she raises draft horses. She is still bitter about her experiences at Cisco. Although she had separated from her husband Len, they are still close friends and manage a trust that funds several charities. After leaving Cisco, Bosack started his own company, XKL,in Redmond, Washington and also funds charities including SETI (the Search for Extra-Terrestrial Intelligence).
Cisco has grown even more since 1990, now valued at over 6 billion dollars and controls over three quarters of the router business. Morgridge is still president and CEO, and has directed Cisco's growth through aggressive marketing and acquisitions including new ventures into ATM and gigabit Ethernet.