Everything Old Is New Again
I had a discussion recently with a very talented but young product manager on the merits of SaaS (Software as a Service) and multi-tenancy versus traditional Enterprise Software. The discussion got a little easier when I gave him some history about how we got here. Yeah, I realize that knowing this history makes me ancient in programmer years.
SaaS could be defined as the combination of a new architectural model known as multi-tenancy with a payment model which has been around since the mainframe era — timesharing. In this post I will focus on the history of timesharing. In the next post I will discuss multi-tenancy.
SaaS re-introduced the hardware/software rental model known as timesharing to the world. For those of you who are young enough to think that Paul McCartney was always a solo act it may seem shocking, but yes people where renting software and data centers decades before Salesforce.com was founded.
Timesharing was one of the primary software delivery mechanisms during the mainframe era. At that time, very few companies could afford the huge cost of their own mainframe to run their enterprise applications. So it became quite common in the early days of enterprise software for a hardware company to write an application for a large company and then run it for them in the hardware company’s data center. In fact, this is still a large business for IBM. Even after the advent of packaged software and mini computers, timesharing was a common model. For example, ERP pioneer ASK Computers and HP had a large and profitable timeshare business together with ManMan ERP (ASK’s ERP solution) and the HP 3000. In fact helping convert ManMan from HP/VAX to UNIX and client-server was my very first software engineering job.
The advent of cheap computers, followed by the introduction of the PC and low cost Wintel and UNIX servers, as well as new and different programming languages and techniques, transformed a robust business into a niche. Application timeshare providers discovered that they were in a very expensive business as less expensive and more advanced solutions came on the scene.
Companies now had a choice. Go with a timeshare provider with higher price points due to the infrastructure required to support data centers, lots of hardware, software, and trained personnel to keep unique versions of an ERP package running for lots of different companies. Or go it alone by building your own data center, buying your own hardware and software, hiring your own IT staff, and then paying Andersen Consulting aka Accenture to make sure everything works together. The price points that timeshare vendors charged individual companies to be profitable made it seem cheaper for companies to bring it all in-house.
Companies soon learned that running a data center and maintaining 3rd party developed applications was hard, error prone, and expensive, but it gave them more control and often, was still cheaper than timesharing. But lower hardware prices, the Internet, and mid-size companies desire for more sophisticated software created an opportunity for a revamped version of timesharing to be successful. That opportunity eventually led to SaaS, Salesforce.com, and laid the foundation for the cloud computing platforms that are becoming a ubiquitous part of the technology landscape. How that opportunity evolved, how and why the PatternBuilders model differs from a traditional SaaS vendors will be the subject of my next post.