Oracle 8i, Java, and the Mustang Index
By Ken North
Today's DBMSs should be a model for a Harvard Business School case study about software that results from years of intense competition. Will Web developers continue to benefit from increased DBMS capabilities that let us deliver new classes of applications? To plan for the future, one can sometimes look at historical trends and project them forward. Should you plan for arrays of inexpensive servers and ubiquitous computing? Yes. Commodity-priced hardware and software will enable us to proliferate databases containing all types of structured and unstructured information.
This "prediction" doesn't require a crystal ball because it is evident from long-established industry trends. Despite recent whining about the state of the computer industry, buyers of computers and software receive excellent value -- consistently better performance at decreasing prices. Those who disagree with me should consider the Mustang Index. In 1964, Ford introduced the Mustang at a price of $2368. At the same time, a typical, medium-sized computer leased for $15-17,000 per month. Today you can lease a business-class server for about $60 per month. That's less than 1 percent (.0035) of the 1964 lease payment. In 1964, your computer lease payment was the equivalent of buying seven Mustangs per month, but today your Mustang payment is equivalent to leasing five computers per month.
The declining price of computers introduced strategies such as, "Can't we solve this by simply throwing more hardware at the problem?" Software hasn't matched the hardware price/performance ratio since 1964, but software economics are nonetheless impressive.
