Saturday, February 15, 2014

Winning the Application Platform


Note: This appeared in the IMTS Insider dated Feb 13, 2014
By: Dave Edstrom

James Carville, American political commentator, is known for coining the phrase, “It’s the economy, stupid,” when it was used in 1992 to drive home Bill Clinton’s primary attack point positioning him against President George H.W. Bush.

In the computer industry, there is a similar phrase to focus the big picture, “It’s the platform, stupid.” The Holy Grail is not about selling a lot of software; the Holy Grail is about selling lots of software by winning the platform, specifically the application platform.

There is a famous 2006 video of Steve Ballmer of Microsoft jumping around the stage screaming, “developers, developers, developers.” What Mr. Ballmer was really saying was that Microsoft couldn’t continue its hold on the WINTEL (Windows and Intel) platform without taking care of developers.

What are the key components in winning the application platform? The first key component is to win the architecture. This is best done by having an open architecture. The second key component is to have an outstanding developer platform. The third key component is applications: software applications and hardware expansion solutions.

The first question to answer is, “what is a platform?” Prior to Java, the programming language and computing platform created by Sun Microsystems, a platform was typically defined as the combination of hardware and software that developers would use to build their applications. The WINTEL (Windows and Intel) system was the most well-known platform. “Windows” referred both to the operating system (OS), the software development kit (SDK) – which are the libraries developers’ use – and the integrated development environment (IDE), which developers use to build their applications. During the 2000s, a popular platform was LAMP. LAMP stood for Linux, Apache HTTP Server (aka a web server), MySQL and PHP, Perl or Python. The LAMP stack was very popular because it included the necessary OS, database and programming tools to create a website that was open source and free to use. The application platform includes the set of software tools necessary to build the application and, many times, this includes a database. Android is the open source platform that was created by Google. iOS SDK is Apple’s software platform for their mobile devices.

Why is owning the platform important? Let’s turn on the time machine and go back to a Sun Microsystems conference room on July 1987 when the first ever SPARC (Scalable Processor ARChitecture) system and SPARC microprocessor was introduced. In May of 1987, I started at Sun Microsystems as a Systems Engineer (SE) working in a pre-sales environment. Basically, an SE was the technical arm for the sales rep that explained the technology, did presentations, ran benchmarks and basically did anything on the technical side to drive sales. The average sales rep was in their late 20s, making $350,000 in the Washington DC Sun office. Sun had the best development environment and the best bang for the buck for the workstation market. Sun introduced the Sun 4/260 and announced it would license the technology for both systems and microprocessors. Many Sun employees questioned the logic of this strategy. The overall question could be summarized, “We are killing it in the market. We introduced a system that is literally 2 ½ times faster than any other system that is out there. Why would we be possibly want to license this technology as opposed to keeping it for ourselves?”

The question would be answered by Scott G. McNealy, President, CEO and co-founder of Sun Microsystems, when he visited in person shortly after the global announcement of the new SPARC system. The question was asked of Scott, “Why license SPARC as opposed to just keeping it ourselves?” In the next half hour, we all received a Ph.D. in the economics of technology from the Harvard and Stanford graduate. Scott asked the question, “What were the two most successful computer architectures of all time?” He looked around the room for a few seconds and then answered his own question. Scott went on to explain that the two most successful computer architectures were the IBM mainframe and the IBM PC.

The IBM mainframe was so dominant in the computer industry that many companies felt the only way to make money was not to compete with the industry giant, but to create boards that would be compatible or plug into IBM’s mainframe computers. The term Plug-Compatible Mainframe (PC-M) was created. Scott went on to explain the IBM mainframe market share was above 80 percent. IBM’s mainframe architecture was clearly dominant.

At the other end of the spectrum was the IBM PC. IBM felt pressure in the late 1970s and early 1980s to respond to the quickly growing market that was led by companies like Apple, with the Apple II, Radio Shack, with the TRS-80, and the Commodore PET. Feeling the time pressure, IBM used off-the-shelf parts for the hardware and leased DOS from Microsoft. This combination of off-the-shelf hardware and a leased operating system allowed for the creation of the PC compatible market. These PC compatible systems were called PC clones or IBM clones. The IBM PC architecture was the dominant architecture.

Scott then brought the conversation back to economics and asked the question, “How expensive would it be for Wang, Data General or any of the other proprietary architecture companies to completely change their architecture?” The rhetorical question drove home the point that changing your architecture would be an extremely expensive proposition and could put a company completely out of business. The reasons it is so expensive to completely change the architecture is that customers have applications that they have purchased and written. What do they do with those applications? This could be, and typically is, a nightmare for companies.

By licensing SPARC at both a system and microprocessor level, Sun was attempting to establish an architecture that would be the new platform for high-end workstations and servers. Sun had a world-class developer program. By building the systems on an industry standard VME bus, Sun made it easy for hardware developers to use the new SPARC systems.

Scott McNealy was able to educate all of us by first explaining the logic from an economic perspective. The technology perspective was important, but it was second to understanding the economics of a platform. It was a lesson that none of us would ever forget. At your company, are you simply creating products or a platform? Even if you are not creating a platform, this is a conversation that you should be having, because your competition is not sitting still. Remember, it’s all about winning the platform.