16. Database Lords: Larry Ellison and his Lost Co-Founders (1977-2010)

by Arun Rao

THESE ARE EXCERPTS FROM THE BOOK A History of Silicon Valley

 

Selling Databases to the CIA

In 1977, three software engineers, Larry Ellison, Bob Miner, and Ed Oates, founded Software Development Laboratories. They completed their first product, Oracle Version 1, in less than one year. Their customer was the Central Intelligence Agency (CIA). The engineers had significant experience designing customized database programs for government agencies. Miner and Ellison had persuaded the CIA to let them work on a lapsed $50,000 contract to build a relational database program, after they did some consulting work for a company called Omex.

A relational database allows business users to match data by using common characteristics (a set of relations). Relational databases, as implemented in relational database management systems (RDMBS), have become the main place to store digital information used for financial records, manufacturing and logistical information, personnel data, all Internet records, and much more. They have become the guts of the global technology infrastructure. For example, a data set containing all the real-estate transactions in a town can be grouped in many ways:  by the year the transaction occurred; by the sale price of the transaction; by the buyer’s last name; and so on. An RDBMS allows an organization to store massive amounts of data forever.

Oracle databases were the guts of the electronic world. To put these databases in context, 98% of the Fortune 100 companies depended on Oracle software to manage their information in 2001. Every time someone would use a credit card, buy a plane ticket, reserve a hotel room, order from any catalogue, surf the Internet, search Google or Yahoo, get cash from an ATM, settle phone bill, or so on, odds were that the person interacted with an Oracle database.

The idea for the product came from IBM research. Ellison and Miner had come up with RDBMS idea after reading about it in the IBM Journal of Research and Development, realizing that no one had commercialized it. The key insight Ellison and Minor had was that IBM was interested in RDBMS, which many believed would allow computer users to retrieve corporate data from almost any form. This came from an IBM innovation called the Structured Query Language (SQL), a computer language that would tell a relational database what to retrieve and how to display it. Ellison and Miner had a hunch that IBM would incorporate the new relational database and SQL into future computers, mostly mainframes. So they set out to provide a similar program for digital minicomputers and other types of machines, when conventional wisdom was that it wouldn’t work and would be too slow. It would be the first commercial, relational database. The founders renamed the company RSI.

To start the company, Ellison and Miner pooled $1,500 in savings to rent office space in Belmont, California. Ellison became President and CEO and took charge of sales and marketing for the new company, while Miner supervised software development. They convinced the venture capitalist Donald Lucas to become chairman of the board, after he stumbled upon their company working late into the night in its early offices on 3000 Sand Hill Road. While the first version of the program was never officially released, Version 2 came out in 1979, and was the first commercial SQL relational database management system (RDBMS), running on a PDP-11, a popular computer at that point. The product attracted customers who use it for simple business functions and came out two years before the IBM version.

In the start-up days of Oracle, Bob Miner was the lead engineer, programming the majority of Oracle Version 3 by himself. As head of engineering, Miner’s management style differed from Larry Ellison, who ran Oracle’s hard-driving sales team. While Miner expected his engineers to produce, he did not agree with the demands laid upon them by Ellison. He thought it was wrong for people to work extremely late hours and he wanted them to see their families.

Ellison just wanted results. Bruce Scott, an early Oracle database engineer, felt Oracle was successful mainly because of Ellison’s charisma, vision, and determination. One example Scott gave was when the engineers in the startup had space allocated to them and needed to get their terminals strung to the computer room next door. They didn’t have anywhere to string the wiring. Larry walked in, picked up a hammer, and slammed a hole in the middle of the wall. He said, “There you go,” and then walked away.[28]

In 1981 RSI began developing simple reporting tools after recognizing that customers wanted to write applications to enter and format data into usable reports. By 1982, RSI was profitable with 24 employees, 75 customers in the mainframe and minicomputer space, and reported annual revenues of nearly $2.5 million. Ellison was hiring salesman to aggressively increase revenues while Miner was more circumspect. Ellison hit the road and did demos of the product across the intelligence community to the CIA, NSA, Air Force Intelligence, and so on. Ellison was working 14 hours or more a day, and even had to cut his own salary at one point. His second wife left him as he focused on making products.

 About a quarter of Oracle’s 1982 revenues were poured back into research and development. This led to the 1983 Oracle innovation of the first commercially available portable RDBMS. The portable RDBMS enabled companies to run their DBMS on a range of hardware and operating systems, including mainframes, minicomputers, workstations, and personal computers. Oracle doubled its revenues to over $5 million in 1983. That same year, Miner and Oates rewrote the database code in the C programming language. After that, their RDBMS would no longer bound by any single platform and could be easily modified for many types of computers. RSI became Oracle Corporation.          

 The next few years brought more product innovations and growing revenue. In 1985 Oracle released versions 5.0 and 5.1 to operate in client/server mode so multiple desktop business applications could access a single database on a server. Oracle also began to explore clustering, an early move toward flexible, scalable software. That year, the company brought in more than $23 million in revenues. This more than doubled to a record $55.4 million in 1986.

Two important events happened in 1986. First, the database industry decided to make SQL as the industry’s standard language for relational database management systems. This in turn led to increased market acceptance of Oracle‘s SQL-compatible RDBMS.

Second, on March 15, 1986, a decade after the founding of the company, Oracle had an initial public stock offering of 2.1 million shares on the NASDAQ exchange. The company had a market value of $270 million and Ellison owned 39% of the stock. At the time, the company had 450 employees and was the fastest-growing software company in the world. It had recorded 100% or better growth in revenues in eight of its first nine years. Much of this growth came from Oracle’s targeted end users: multinational companies with a variety of previously incompatible computer systems. By 1986 Oracle’s customer base had grown to include 2,000 mainframe and minicomputer users. These customer firms operated in such fields as the aerospace, automotive, pharmaceutical, and computer manufacturing industries, not to mention government organizations.

The IPO would reward investors well. Twenty years later, Oracle had a global workforce of 65,000 and annual revenue topping $15 billion. A $10,000 investment in the IPO of Oracle back in 1986 would by October 2006 be worth over $4 million. Revenues would go from $20 million in 1986 to $11 billion in 2001, when the companies would have operating margins of 35% and a cash pile of $6-8 billion dollars.

In 1986 Oracle expanded its RDBMS product line and released a distributed DBMS based on the company’s SQL*Star software. Under the distributed system, computer users could access data stored on a network of computers in the same way and with the same ease as if all of a network’s information were stored on one computer.

     

Getting into the Applications Market

By 1987 Oracle was fairly successful. It had topped $100 million in sales and had become the world’s largest database management software company, with more than 4,500 end users in 55 countries. But Ellison was restless, and he wanted to branch out from databases into applications (computer programs) that used the information in databases for business purposes. Oracle created an applications division and began building its own business-management software, integrated closely with its database.          

 Ellison had decided by this point that he was a product guy and didn’t like most of the CEO duties. So he concentrated on that and delegated all the rest, something Ellison called “closer to abdication than delegation.”  Ellison was also a brilliant recruiter of programming talent because he knew the product so well. While the sales and executive team could have a revolving door, there was a “kernel group” who created the core product and stayed, accumulating knowledge and experience to continuously make the software better.

Oracle a year later introduced a line of accounting programs for corporate bookkeeping, including a database for personal computers to work in conjunction with the Lotus Development Corporation’s popular Lotus 1-2-3 spreadsheet program. The company also introduced its Oracle Transaction Process Subsystem (TPS), a software package designed to speed processing of financial transactions. Oracle’s TPS opened a new market niche for the company, targeting customers such as banks needing to process large numbers of financial transactions in a short period of time.

Meanwhile, hot backup allowed employees to continue working in the system while administrators duplicated and archived data (so it reduced overhead costs). The technology behind this, PL/SQL, generally allowed users to process data while it remained in the database.

In 1989 Oracle was booming. The company was added to the S&P 500 index of stocks. Oracle relocated from Belmont to a new, larger office complex in nearby Redwood Shores, California. Seeking to break into new markets, Oracle formed a wholly owned subsidiary, Oracle Data Publishing, in December 1989 to develop and sell reference material and other information in electronic form. Oracle closed its books on the 1980s posting annual revenues of $584 million, netting $82 million in profit.

Times were good but the growth bought trouble. In March 1990 Oracle‘s revenues jumped 54% but net earnings rose only by 1%. Oracle’s first flat earnings quarter, attributed to an accounting glitch, shook Wall Street out of its long love affair. Oracle had been booking revenues too aggressively by discounting product and shipping incomplete or buggy sofwtare the day after the earnings announcement the company’s stock plummeted $7.88 to $17.50 in record one-day volume with nearly 21 million of the company’s 129 million shares changing hands.

In April 1990 a dozen shareholders brought suit against Oracle, charging the company had made false and misleading earnings forecasts. On the heels of this lawsuit, Oracle announced it would conduct an internal audit and immediately restructure its management team with Lawrence Ellison assuming the additional post of chairman, while Lucas remained a director. Oracle also formed a separate domestic operating subsidiary, Oracle US, aimed at addressing its domestic management and financial problems, which the company attributed to poor earnings.

Part of the problem was Larry Ellison’s management style. He was a sprinter who would work hard, rest, and then sprint again. Ellison could get bored of the company and then take weeks off to travel the world or spend time on his yachts. He could listen to his executives with intensity or completely ignore them. Ellison felt good when everyone said he was nuts because it was a sign that Oracle was trying to do something innovative. But Ellison also paradoxically cautioned, “when people say you’re nuts, you might be nuts. You’ve got to constantly guard against that possibility. You don’t want people saying you’re nuts too often.”[29] 

For the fiscal year ending May 31, 1990, Oracle initially posted record sales of $970.8 million and profits of $117.4 million; but these results were below Oracle’s own estimates. The company’s stock price fell from a high of $28.38 to $19.88 then plunged to $11.62 in August after an internal audit forced the company to restate earnings for three of its four fiscal quarters. Jeff Walker, the CFO, had messed up in receivables and cash management, and Ellison had not been a good supervisor. At one point, Walker sought a cash infusion from the outside, but Ellison held back as he didn’t want to dilute his personal equity stake. As a result, Oracle negotiated a $250 million revolving line of credit from a bank syndicate. A few weeks later the company reported its first-ever quarterly loss of nearly $36 million with expenses outpacing revenues by 20%. The corporate bank account had only $3 million at its low point. The stock tumbled once again, hitting a low of $4.88.

Ellison had to make major changes, which started with changing his management team. Oracle also: moved to reduce its annual growth rate goals from 50% to 25%; laid off 10% of its domestic workforce of 4,000; consolidated Oracle US’s financial and administrative operations; and folded various international units into a single division. A lot of top talent left at that point, including people like Tom Siebel. It was a low point for Ellison, as he had borrowed on his Oracle stock and so received margin calls. Meanwhile, his third wife Barbara left him and Miner, his co-founder, wanted out by selling. The company’s board wanted to kick Ellison out, but Don Lucas stood on his side. Ellison had to start paying attention to accounting and legal, not to mention the sales teams. Ellison started fighting back and said:  “I’ve always been more motivated by fear of failure than greed. And I hate losing.”

For 1991 Oracle topped the $1 billion sales mark for the first time in history and at the same time posted its first annual loss of $12.4 million. In October the company secured a new $100 million revolving line of credit from another bank syndicate. Oracle negotiated an agreement for $80 million in financing from Nippon Steel Corporation, which also agreed to sell Oracle products in Japan. In return, Nippon was given rights to purchase as much as 25% of Oracle’s marketing subsidiary in Japan, duly named Oracle Japan.  

The year 1992 brought much change. Ellison bought in Ray Lane, a senior partner from the consulting firm Booz Allen Hamilton, to help turn around the sales team, while his partner Robert Shaw came to build up Oracle‘s consulting arm. They would be two great hires and they helped turn Oracle around. They built their own executive teams and forced Ellison to listen and respond to criticism and unpleasant facts. In 1992, Oracle7 also came out after four years of research and development and two more years of customer testing. It supported a larger number of users than previous versions, handled more transactions, allowed data to be shared between multiple computers across a network, and improved application development features. It won industry praise and emboldened Ellison to talk up database technology for the Internet. Bob Miner also left in 1992 as he had trouble managing a large team of engineers and soon found out he had cancer.

By the end of its 1992 fiscal year, Oracle‘s balance sheet had improved as sales inched modestly upward and earnings rebounded, with the company reaching $1.18 billion in sales and $61.5 million in profits. Oracle entered 1993 with no bank debt, solid long-term financing in place, and in an improved financial position controlled by a revamped management team.           

Ellison and Lane came up with a three-part strategy. First, they wanted to build Oracle 7’s database market share at whatever cost. Second, they had to make sure their database was the best on the market by beating Sybase and Informix decisively. Third, Ellison wanted to branch out of databases into other applications. He was interested in video-on-demand (which would be a dead end).

By 1993, Lane was doing so well in US sales that he moved to global sales and was eventually heading to become the company’s President. Ellison also made mistake firing the entrepreneurial Geoff Squire, who would go on to build Veritas into a strong software company.

By mid-1994 Oracle‘s sales had reached $2 billion, its consulting services accounted for 20% of sales. Consulting became an important part of Oracle’s model. While its competitor, SAP, had a nice relationship with Anderson Consulting, the world’ largest IT systems integrator, Oracle was pressing down hard.

The standard in the enterprise software business was for a large company to buy dozens of “best of breed” applications from a range of vendors. Then the company would hire an IT consulting company, like Accenture, to connect all these bits of software, with their multiple databases and systems. Data would become fragmented, duplicative, and conflicting across the databases behind the multiple programs. The analogy would be buying dozens of car parts from different places and then putting together your own car (without the different parts-makers having a common standard tying them together).

Since so many of Oracle‘s applications were so bad, its consultants pushed a “best of breed” strategy and then worked to integrate outside applications. Meanwhile, Ron Wohl would take over application development at Oracle and try to improve the internal products. The battles between Wohl and the Lane/Shaw duo would be rough, as the latter didn’t like selling the buggy software coming out late from Wohl’s team. Meanwhile Oracle also started buying and partnering with other, smaller software companies.

                 

Fighting Microsoft over Internet Strategy

By 1995 Larry Ellison had “found Jesus” and had a specific vision and strategy for the Internet. During a keynote presentation at a conference in Paris, Ellison introduced his vision of the network computer, a small, inexpensive device that would run applications via the Internet. The keynote speech took the technology world by storm and would pit Ellison against Bill Gates of Microsoft, who still believed in the PC/server model of computing.

With Oracle‘s revenues topping $4 billion, in May 1996 Ellison took on the “Wintel” (Microsoft Windows software plus Intel‘s processing hardware) monolith by unveiling the “Network Computer” (NC). It was a kind of stripped-down PC with no hard drive and therefore no applications. Joining with such partners as SUN Microsystems and Netscape, Ellison offered to free corporations from the costly upgrades Intel and Microsoft forced on them with every new release of Windows and the x86 family of processors. Using Ellison‘s $500 NC, data and applications could be stored and accessed as needed via the World Wide Web or remote server computers, equipped, naturally, with Oracle’s databases. Since corporations would no longer have to buy storage and applications for each computer, they could save millions with no loss in functionality, and Oracle would have a vast new market for its database products. By late 1996 this strategy had evolved into the “Network Computing Architecture,” a complicated new three-tier world for corporate computing consisting of a client computer (the computer accessed by the user), an applications (such as word processing software) server, and a database server.

While Ellison battled Microsoft, he was still practical. In 1996 Oracle ported all of its development tools, object technology, and modeling and analysis tools to NT. Recognizing that Microsoft‘s Windows NT operating system was becoming increasingly popular with small businesses, Oracle delivered a multi-node scalable database for Windows NT clusters.           

In 1997 Ellison unveiled Oracle8, based on his vision of the Internet and network computing. Meanwhile, the “best of breed” consulting strategy was falling apart, as it was difficult and expensive to link many different application programs. That same year, Ray Lane wanted to leave for the Novell CEO post, but Ellison bribed him to stay with $2.5 million in options. Many Oracle applications weren’t good enough, so Ellison had to fire Ron Wohl and take over the applications development group himself. He would have to start making applications for things like order management, tables, and accounting; all were things Ellison had found boring before, but he now had to master them. Ellison also started to feel that Lane wanted him to fail.

                 

Selling its Own E-Business Application Suite

Oracle prepared to release version 11i of its E-Business Suite in 2000. It would provide the most substantial integration of CRM and ERP applications to date. It was intended as an entire ecosystem of enterprise computing and the IT consulting industry, as Oracle claimed customers could get all they wanted from one place. There was no need to find a “best of breed” supplier and then pay expensive consultants for years to patch together systems. Ellison was changing the company’s strategy to have a single, global database support a range of business applications, a suite from marketing, sales, supply chain, manufacturing, customer service, accounting, and so on

It was an attack against Microsoft‘s client/server model of computing again. The difficulty with that was anytime new software or hardware was updated, it had to be installed across hundreds of computers or more. That required costly IT labor hours.

Ellison also started in 1999 to get more involved in sales force compensation, an important issue. He took this responsibility away from Lane and created a better comp plan that was transparent, had stretch targets, and allowed the best salesmen to make more money and the worst to clear out. He also corrected dysfunctional incentives, such as this one:  if a salesman sold a million dollars of Oracle product directly, he got $100,000; if he sold Oracle product through a partner and made the company $600,000, he got $120,000. So the sales force pushed the less profitable deals for large bonuses.

Another big event in 1999 was the arrival of Safra Catz, a former DLJ investment banker, who became Ellison’s chief of staff. As a former lawyer and banker, she had a forensic approach to digging out facts and then analyzing them. Ellison appreciated her methodical approach, saying: “In an argument when nobody has any facts… the person with the strongest personality wins. But when one person has the facts and the other doesn’t, the one with the facts always wins. When both people have facts, there’s no argument.”  After Catz arrived, Ellison started stripping Lane of more responsibility.

Oracle also launched its rebuilt application server, Oracle9i Application Server. It included Web caching technology that dramatically increased Web site performance and scalability and cached dynamically generated as well as static pages.  

Oracle shipped Oracle E-Business Suite Release 11i, the first Internet-enabled suite of business applications built on a single data model for seamless, real-time business intelligence. It was a big deal because Ellison was turning against the client/server model of computing for an Internet “cloud” model. The hardest part was convincing Oracle’s own engineers that this was the right direction and to get them to support the new product strategy. Ray Lane fought Ellison‘s technology decision and was aghast at making the Internet the core of the company’s platform.

Ellison had a brilliant strategy for selling 11i. He first implemented the entire software internally at Oracle, and showed that it saved Oracle $1 billion annually. Oracle then spent $300 million marketing this message:  “By using our own E-Business Suite, Oracle saved $1 billion in one year.”  The message was mostly correct; the new suite required that companies adapt their business processes to it, but it was cheaper and more powerful, as an independent Economist Intelligence Unit study showed.

Oracle finished fiscal 2000 with revenues of $10.2 billion and earnings at an all-time high of $6.3 billion due to an extra $4 billion from selling shares in Oracle Japan. By the following year, Oracle prospered like its former self of the 1980s with soaring sales, new product releases, and a myriad of new ventures both in the United States and abroad. The company finished the year with sales close to $11 billion and $2.6 billion in earnings.

In June 2000, Ellison fired Ray Lane, the company’s President and Ellison‘s heir apparent, saying:  “It’s like a marriage that went bad. I don’t know [what went wrong].”  Lane had cleaned up the US sales force and helped the consulting business grow, yet Ellison didn’t see him as his replacement anymore. Lane never had a real social relationship with Ellison, who stated it this way:  “Ray’s a duck hunter. I raise mallards every spring. We couldn’t be more different in personality and pastimes.”

 In 2001 Oracle released Oracle9i Database, with technology supporting software as a service. Oracle also redesigned its business applications to run on wireless and mobile devices. Oracle9i Database added Oracle Real Application Clusters, giving customers the option to run their IT on connected, low-cost servers—expanding performance, scalability, and availability of the database.

In 2002 Oracle Database 10g was introduced. An early database for grid computing, it allowed groups of low-cost servers to be connected by Oracle software and runs applications faster than the fastest mainframe, in addition to offering self-management capabilities.

Meanwhile, Ellison started re-tooling the Oracle sales culture to get creativity out of the process and make it more “engineered.”  Ellison wanted a new process that involved first identifying a customer’s decision maker and documenting it in their sales system. Then, the salesman could send a set of key customer references and case studies showing how other customers got better performance on a lower cost system, with case studies. Third, the salesman could send a proposal quantifying cost savings in hardware, software, and labor by implementing a specific Oracle product at that customer’s company. Finally, the salesman could send a contract with standardized terms and a price quote.

By 2004 Oracle began to offer easy-to-implement, low-risk, affordable solutions for small and medium-sized businesses with Oracle E-Business Suite Special Edition and Oracle Database Standard Edition One. In 2005 Oracle OpenWorld was the biggest event in Oracle’s history, opening its doors to more than 28,000 attendees, and offering more than 800 sessions and activities.    Oracle was the king of American IT technology conferences; only his friend Steve Jobs could do better.

     

Growth through Targeted Acquisitions

From 2003 onward, Oracle‘s growth strategy shifted towards acquiring other companies making business software. The motivations behind Oracle’s largest acquisitions were to increase market share in large business software markets, to expand profitability by consolidating high-margin, customer support revenue while cutting labor costs, and to offer a complete technology “stack” of software applications and hardware.

All of Oracle‘s large deals over the next six years met these requirements. Peoplesoft, Siebel, and Hyperion all strengthened Oracle’s market share position in the applications market while contributing captive customer bases that pay highly profitable support fees. BEA Systems was important for its stake in the middleware market, which helped Oracle integrate so many applications it sold. Finally, SUN Microsystems brought recurring support revenue. It was also interesting for two other reasons:  first, it demonstrated Oracle’s willingness to move into servers and storage (including hardware); second, Oracle took control of Java, a key programming language for Web and Internet development. In 2010, Ellison succinctly stated: “our strategy is in creating and acquiring intellectual property.”

In mid-2003 Oracle initiated a hostile takeover of PeopleSoft Inc. for $5.1 billion. The Pleasanton, California-based PeopleSoft, was in the process of acquiring J.D. Edwards & Company and was not amused by Oracle’s takeover bid, no matter how attractive the offer. For its part, Oracle raised its offer several times in the succeeding months, to as high as $9.4 billion, only to be met by a storm of controversy. Few people, save Ellison, were in favor of the takeover. Shareholders of both firms were unhappy. The US Department of Justice got involved over antitrust issues. By the end of 2003, Ellison was determined to win the battle, whatever the cost. Oracle’s year-end revenues fell for the second year in a row to $9.5 billion.

Oracle announced the acquisition of PeopleSoft at the end of 2004 and completed the transaction in January 2005, adding PeopleSoft Enterprise, JD Edwards EnterpriseOne, and JD Edwards World applications to its product lines. This was followed by the acquisitions of Siebel, Retek, Oblix, and other strategic companies.           

In 2006 Oracle released Oracle Database 10g Express Edition, its first free database edition for developers and learning DBAs. Oracle acquired several companies including Sleepycat Software, the makers of the world’s most popular open-source database, Berkeley DB, and released Oracle Secure Enterprise Search, a new standalone product that enabled secure, high-quality, easy-to-use search across all enterprise information assets.

In July 2007 Oracle bought Hyperion Solutions Corporation, a global provider of performance-management software solutions, through a cash tender offer for $52.00 per share, or approximately $3.3 billion. Earlier that year, Oracle filed a court case in the Californian courts against a major competitor, SAP AG, for malpractice and unfair competition. In October 2007 Oracle announced a bid to buy BEA Systems for a price of $17 per share, an offer rejected by the BEA board, which felt that it undervalued their company. In January 2008 Oracle bought BEA Systems for $19.375 per share in cash for a total of “$7.2 billion net of cash.”

In September 2008 Oracle started marketing servers and storage in a co-developed and co-branded data warehouse appliance named the HP Oracle Database Machine.

In April 2009 Oracle announced its intention to acquire SUN Microsystems for $7.4 billion ($9.50 per share). After a bidding battle with IBM, in January 2010 Oracle acquired SUN. SUN Chairman Scott McNealy had to leave the company he co-founded 28 years earlier. He wrote in a bittersweet memo, “my hat is off to one of the greatest capitalists I have ever met, Larry Ellison.” McNealy preferred that SUN would be the great and surviving consolidator, but he was happy with the sale and his payout.[30]

Sun’s technology gave Oracle a place in the server, storage, and processor domains. Oracle became a direct competitor to more companies, even hardware customers to whom Oracle sold its database and other software for use on servers sold by those competitors. Now IBM, Hewlett-Packard, Cisco Systems, and EMC would be direct competitors. SUN made most of its revenue from selling computers but Oracle executives said they didn’t regard SUN as a hardware company. As Oracle President Safra Catz stated, hardware would mean factory ownership and large capital investments. However, SUN outsourced nearly all the manufacturing, assembly, and servicing of its hardware. It was more of a hardware design company.

Oracle’s sales pitch was one of integrated products: hardware and software built to work together so customers don’t have to do the integration work themselves or pay an expensive third-party consulting firm, like Accenture, to do it. Also it would reduce software development costs and bugs, and improve security. Finally, Sun’s computer designers could tailor hardware to the combined company’s software, promising further gains in efficiency (something that Oracle was very good at advertising).

Beyond hardware, Ellison said that Sun’s Java programming language and its Solaris operating system were the main attractions, calling the highly popular programming language Java “the single most important software asset we have ever acquired.”  Oracle could offer a more complete set of corporate software, ranging from Sun’s hardware, operating system, and programming tools to Oracle’s existing database and business applications. The end goal would be to help companies automate operations like finance and customer relations management.

Oracle was always a hard-charging company, and it got into trouble again in late 2010. In July 2010 Oracle was indicted for fraud by the US Department of Justice, after an employee tip-off and investigation starting in 2007. The government accused Oracle of defrauding the US General Services Administration (USGA), which negotiated contracts for the government, on a software contract running from 1998 to 2006, involving more than $1 billion in sales. According to the filings, Oracle agreed to give federal buyers discounts of up to 40%, which Oracle said was steeper than the discounts it gave similarly sized corporate customers. In reality, Oracle’s sales force was allegedly authorized to give similar-sized customer discounts ranging from 40% to 70%, where 90% of other corporate deals contained discounts larger than what Oracle gave to the government. The government also claimed Oracle failed to inform it that Oracle was giving other customers better deals and. Oracle allegedly went out of its way to manipulate deals so that they wouldn’t have to report them to the government.