Custom Search
Entrepreneurial Capitalism and Innovation:
A History of Computer Communications 1968-1988
By James Pelkey

Entrepreneurial Capitalism & Innovation:
History of Computer Communications
1968 -1988
By James Pelkey

This history is organized by three co-evolving market sectors and also standards making.
An overview of the schema is presented in the Introduction.

Ch. 1: Emergence
Ch. 3: Competition
Ch. 5: Market Order
Ch. 11: Adaptation

Ch. 2: Vision
Ch. 4: Arpanet
Ch. 6: Diffusion
Ch. 7: Emergence
Ch 8: Completion
Ch. 10: Market Order

Ch. 9: Creation

Ch. 12: Emergence



Chapter 12
Internetworking: LANs and WANs 1985-1988
Local Area Networks and Wide Area Networks


12.29    Retix

Like Vitalink, the founders of Retix would partner with a multi-billion company. In the spring of 1984, Andrew (Andi) De Mari learned that Ing. C. Olivetti SpA of Italy (Olivetti), his former employer, had signed a mega-contract with AT&T, one calling for AT&T to purchase 22% of the ownership of Olivetti, and, among other terms, for Olivetti to supply AT&T with a personal computer AT&T would resell as its own. [1] De Mari’s entrepreneurial instincts quickly envisioned an opportunity that met his search pattern. Logically, if Olivetti intended to supply a competitive personal computer to AT&T, they would want a source of StarLAN adapter cards. (The “StarLAN” version of Ethernet was a standard AT&T had championed in standards committees as an Ethernet standard for twisted pair wiring, a compelling advantage for the phone company.) Di Mari knew, after contacting former associates within Olivetti, that Olivetti management had no plans to engineer and manufacture a StarLAN card, and would be open to a proposal from Di Mari. Furthermore, De Mari had received unmistakable signals that Olivetti’s venture capital team would be interested in investing in a start-up he headed.

StarLAN might be a source of sales and venture capital to launch a company, but De Mari doubted it would serve as the basis of an eventual IPO company. In discussions with those he hoped would join him, the idea of focusing on OSI software seemed a wiser strategy. [2] De Mari remembers a meeting of ten of so engineers in late 1984: [3]

“We looked at the networking market, and decided that if we developed products like network bridges and routers that connect networks that didn’t exist at that time, we would surely go out of business. We sensed a growing interest in OSI (Open Systems Interconnection), so we decided to gamble our entire company on the future of OSI.”

In March 1985, De Mari and the founding team incorporated Retix with an initial investment of $700,000. De Mari knew that he would need to raise significantly more money if Retix was to become an IPO-company, possibly $10 million or more over the next few years. The first step had to be to close Olivetti as a corporate partner, for Olivetti was known as a venture capital investor and, if they decided not to invest in Retix, it signaled the wrong assessment of both Retix’s future business prospects and Olivetti’s confidence in the management of Retix. Meanwhile, thanks to sales of StarLAN adapter cards to Olivetti, sales for 1985 totaled nearly $722,000. If AT&T was successful in selling personal computers, the future of Retix seemed promising.

The years 1985 and 1986 were consumed with: meeting the needs of Olivetti, i.e. AT&T; developing an OSI business; and building out the company’s infrastructure. While selling StarLAN adapter cards to Olivetti hopefully meant more than selling to AT&T, it was clear that success would be measured by making AT&T successful. The challenge would be to form strong relationships with AT&T personnel without upsetting Olivetti or creating new problems. Developing an OSI business required a continued presence in standard-making bodies; selecting, coding and productizing those standards judged to be the best market opportunities; and creating profitable relationships with the leading companies staking a claim in the emerging OSI market. Most importantly, Retix had to prove it was the OSI leader by providing as much OSI software as possible to the vendors participating in the Enterprise Networking Event (ENE) scheduled for June 1988. Building out the infrastructure included hiring the right people as quickly as possible, growing sales and profits to exceed minimum goals and not just letting sales and profits happen as may, making sure financial resources were always available when needed, and initiating efforts to enter the Internetworking market.

In August 1986, Retix closed a $1 million round of Preferred Stock with Olivetti. All was going as hoped, even planned. Then in November 1986, AT&T announced a major shake-up in its flagging computer operations. An untenable loss estimated to be as much as $800 million for 1986 precipitated management and strategy changes. [4] The news for Retix could be interpreted positively -- since ties between AT&T and Olivetti were strengthened when Vittorio Cassoni, an Olivetti senior executive, was named President of AT&T’s newly formed Data Systems Division, i.e. its computer businesses.

Retix’s sales soared to $5.2 million ($3.7 million) in 1986. [5] While sales of $1 million to Olivetti reflected AT&T’s problems selling computers, sales of licenses of OSI software more than made up any weakness in StarLAN sales. (Furthermore, sales of MAP licenses represented approximately half of OSI sales.) [6] On the other hand, by the end of 1986, the project to develop an Ethernet and StarLAN bridge had fallen an estimated six-months behind plan. The reasons for the delay, difficulties hiring experienced talent in the United States, prompted De Mari to accelerate discussions of international product development and even the possibility of making an acquisition. Conversations ranged from securing a grant from the Irish Development Authority to assessing companies as acquisitions. The focus quickly settled on the U.K. company TSL Communications. The crush of activities also elevated the priority of raising new investment capital from venture capitalists. As warned, venture capitalists quickly voiced a closing condition that a new President with IPO experience be hired.

In June 1987, Retix finally entered the Internetworking market with its RetixGate Model 2244: a two-port local bridge interconnecting Ethernet and StarLAN LANs. Even so, Retix was late to market and faced intense competition.

A few months later, in August, Retix closed a $5.5 million financing based on a $16 million pre-money valuation. While a new President had not yet been hired, encouraging interviews had been had, and optimism prevailed. Before the end of the year, terms would be finalized with Stephen (Steve) Frankel, formerly with Micom and recently President of Emulex. For mutual reasons, Frankel would consult with Retix for up to a year before becoming an employee. (He formally joined Retix in September 1988.) Helping to close the financing was the prominent role Retix had secured in the upcoming and important ENE show.

Fiscal year 1987 was another impressive year with revenue growth of 126% to $13.6 million ($10.0 million) and with net income of $560 thousand. The budget for 1988 positioned the company to become a public company. Retix was now recognized as the market leader in the fast-growing OSI- software market with a 50% market share. The OSI market was forecasted to grow from $15 million in 1987 to $164 million in 1991.[7] So far, competitors were other small start-ups: notably Touch Communications, founded by “angel” investor Charles (Charlie) Bass, founder of Ungermann-Bass; and The Wollongong Group of Palo Alto, a leader in the TCP/IP software market, and of comparable size to Retix. International competitors played a more prominent role than usual with: Sydney Development Corporation of Vancouver, B.C., Marben of Paris and Logica of London, all financed with private capital.

1988 would be an important year for Retix, as it was for most computer communication companies.

In January, Retix announced its second local bridge: the RetixGate Model 2255, a two-port bridge adding support for Thin Ethernet (the 10BASE2 standard).

Then in April, the trade press made much of the fact that AT&T had declined to increase its stake in Olivetti, a conversation that had been initiated by Carlo de Benedetti for personal reasons. [8] Days later, another press release confirmed Cassoni was leaving AT&T and returning to Olivetti as managing directory. As for their 10-year agreement, nothing was said. When Retix management learned that AT&T had cut its purchase orders by 75% is not clear, but likely that spring. [9] No longer could Retix assume a close and favorable relationship with AT&T, which in turn cast new importance on mounting an aggressive Internetworking strategy.

In June came ENE, considered the preeminent event for the Computer Communication industry for 1988. In planning for over two years, if the OSI standards, and especially the MAP/TOP and GOSIP Profiles, were to gain the needed economic momentum to displace the existing Computer Communication standards, such as TCP/IP, SNA and DECNET, companies had to leave ENE convinced to convert to OSI --immediately. For Retix, it was the opportunity to demonstrate their OSI leadership, in particular with X.400 and Directory Services (email), and to introduce key new products (X.400 and FTAM end user products and a MAP 3.0 compliant MAC Level bridge) to bolster OSI sales. [10] Retix would fare better than did the Show itself, for the Corporation of Open Standards (COS) failed to have their conformance testing software ready, leaving critics to conclude: “user acceptance of OSI, is still a season away.” That assessment notwithstanding, Retix’s sales of OSI technology and products received a big boost.

In December, in an aggressive step to boost its position in the Internetworking market, Retix acquired all the outstanding common stock of TSL Communications Ltd. In a transaction accounted for as a pooling of interests, Retix now had a source of its remote bridge products, engineering personnel and the European market presence that would have taken years to accomplish otherwise.

Revenues for 1988 grew 99% to $27.0 million ($21.3 million) with net income of $822 thousand. While management had successfully grown two new product lines in exciting emerging markets - OSI Technology and Products as well as Local and Remote Bridges – Retix’s future remained tied to the outcome of the Olivetti and AT&T relationship and of AT&T’s ability to sell personal computers. (See Exhibit 12.17 Retix Revenue Mix 1988)

Exhibit 12.17 -- Retix Revenue Mix 1988

Product Line

Revenues ($M)

Percent of Total

LAN Controllers (Olivetti)



OSI Technology and Products



Internetworking (Bridges)



Support and Custom Engineering



Total Revenue



The uncertainty, and thus challenge, in the 1989 financial plan was forecasting a high revenue growth rate and profitability consistent with the last three years in order to qualify for an IPO, and to do so without knowing what AT&T and Olivetti were going to do, or if OSI was really going to take off.


[1] The ownership percent is reported as high as 25% in: “Telecommunications in Europe” by Eli Noam, Oxford University Press. P. 248. Another term of the arrangement gave AT&T the option to increase eits ownership to 40% in 1987, a term that will cause future complications. Revenues of Olivetti were $3.82 billion in 1985 compared to DEC’s $6.7 billion.

[2] The logic mirrors that of Concord Data Systems: one technology to launch a company – dial-up modems or StarLAN adapter cards and another technology to build an IPO company – token bus LANs or OSI software.

[3] “Retix cashing in on OSI,” Network World, Sept. 22, 1986, pg.25

[4] “AT&T Computer Shake-up Rattles Out a Warning,” Chicago Tribune, Nov. 9, 1986, pgs. 1-2

[5] I remind the reader that for companies that acquired companies in subsequent years, accounting conventions could require a restatement of prior year financials to reflect the results of both companies. I will report the restated revenue and include Retix’s sales in parenthesis.

[6] See Footnote 3 above

[7] TBD

[8] “Trouble in AT&T – Olivetti Paradise,” Computer Business Review, April 17, 1988, pps.

[9] TBD

[10] Enterprise ’88 Show Directory