Historical Perspectives on the Computer Utility and Software for Lease on a Network

Software as a service (SAS) involves charging users a monthly subscription for the use of software instead of a fixed price paid in advance. The author intends to look at historical roots in the software industry that lead to this evolution.

Corporate Mainframes, Service Bureaus, and Software Products

At the onset, computer software was largely a complementary good - even a complimentary one, in that IBM would give away software, free of charge, to customers who purchased its mainframe computer systems.

Software was developed by service companies, as the task of developing software was a one-time intensive effort and few companies had sufficient needs for new program development to justify an in-house staff.

These service bureaus served similar customers, and began to standardize their products (payroll software written for one client could serve another, as their needs were largely similar).

Standardize products, sold in packages, were the model that most software companies followed for many years, and it continues to be the predominate model for software sales.

The Computer Time-Sharing Industry

A slightly different model is based on computer time-sharing, in which several companies who could benefit from the use of a mainframe at times, but not constantly, will purchase the use of a mainframe computer when they need it, and the same computer will be rented to other firms.

As computing advanced, it became possible for multiple clients to use the same mainframe simultaneously, running the same processes on different sets of data. The terminal-mainframe setup is similar in many ways to the present client-server model of computing.

In terms of a business model, this closely follows software-as-service model: users were paying for time on a computer system, which was a bundle of hardware and software to perform certain data-processing tasks.

This model was in place for a few decades before the personal computing revolution, which pushed software onto a client machine for local execution. Arguably, if there were a network in place through which client terminals could access mainframe computers over large distances, software as boxed product may not have evolved at all.

The Driving Forces

The desire for computing in business was driven by economic forces. Fundamentally, a clerk with an adding machine can perform 30 or so calculations per minute with varying degrees of accuracy. Meanwhile, a mainframe computing system can perform millions of calculations in a minute, with perfect accuracy.

Hence the value of computing to business is efficiency and accuracy versus cost: where the need to do a large amount of processing in a short amount of time with a high degree of accuracy justifies the cost (specifically, the cost of buying a mainframe was less than that of hiring an army of clerks to attend the same tasks), businesses sought to automate their operations.

Mainframe computing slowly faded away when personal computing, at a cost of $1500 to $3000 per computer, became an cost-effective alternative to mainframe computing.

In some instances, it also made a difference that personal computers were acquired as capital assets that could be depreciated over the course of a few years, whereas the cost of mainframe computing was an operational expense.

Whatever the case, the mainframe model was largely dead my the mid 1980's for most companies, largely replaced by client-server computing models.

Application Service Providers (ASPs)

There was a resurgence of the SAS model in the later 1990's as Application Service Providers (ASPs), provided service to companies that included a bundle of software, training, and support for a monthly per-user fee. These were classified as "horizontal" services (which provided generic applications such as payroll to a wide range of customers) and "vertical" ones (which provided specialized solutions for specific industries).

By this model, companies did not own the software (possibly not the hardware or personnel, either), but paid a fee for a service provided to deliver a packaged solution.

The emergence of the internet lowered barriers to entry into this market. Small firms could compete with giants like CSC and EDS, as it was no longer necessary to maintain huge mainframes or an extensive private network: your data center could begin as a single server and grow over time as clients were added.

As a result, ASP computing was a hot trend, and small companies sprang up0 like mushrooms, hoping to grow large over time. Naturally, many died in the dot-com crash, but some truly huge firms, such as PeopleSoft, survived the crash and remain to this day.

Various other facts and historical tidbits are provided, but germane to the subject of this article: the ASP model is similar to software-as-subscription to commercial customers.

Software on Demand and Web-Native Applications

In 2001, news broke that the "next big thing" would be web-based computing, in which customers would purchase a license to use software on a remote server rather than storing programs locally (and purchasing them as "packages").

The key difference, conceptually, is that software-on-demand is a service to the individual user, rather than to the company, The example given is personal e-mail and calendar services, such as Hotmail, which require no client software other than a Web browser: all computing is done, and all data stored, on the Web server.

This also extended to "collaborative workspaces", which provided a common workspace for various individuals. Such solutions provided standard productivity software (word processing and spreadsheets) in which the software and data resided on a remote server.

One might also suggest that online multiplayer games (WoW) follow a distributed computing model, though part of the software runs as a client-side application, much of the functionality and data reside on central game servers, without which the client-side component is unusable.

To date, this hasn't gone much further. There is some speculation that it might (companies like Google are doing significant R&D in this space), but examples in which this approach has replaced traditional client-side software on a widespread basis are presently nonexistent.

It is also worth noting that, in many cases, this is simply a matter of labeling, especially when companies buy "private" systems to do collaborative tasks: old ASPs were now passing themselves off as "on-demand software" companies, but made no change to their fundamental business operations or the technology product they sell their customers.

During the same period, the Web also became a medium for the provisioning of client-side software: users would purchase a license for a program and download the software over the Internet. The same could be done with upgrades and patches. It was no different from the "disk in a box" approach, except for electronic delivery,

The Impact of SAS on Independent Software Vendors

Independent software vendors often view SAS as a "disruptive technology" - in that established vendors with strong market solutions are being threatened by newcomers with a different product paradigm.

These newer products are tighter, more flexible, and often more functional than the juggernaut solutions ASP companies provide, though they have the weakness of not being refined over time and proven in the industry. Still, all technologies begin as "disruptive" and may mature to replace the established technology.

Comments on specific old-school companies:


As of the time the article was written, a paradigm shift toward SAS is conside4red "pure speculation" and SAS solutions have a very small market share in very few niches of the industry.

However, from the content provided in this article, it has been shown that this is not something entirely new, and is in fact more of a change back to a previous way of doing business in the software industry.