On-demand is a cult. It is seen by many as the only true path for success in the software industry and is pursued with a fanaticism and intolerance that allows no variation from fundamental tenets that are whispered in awe, but rarely understood.
So what exactly is it? At best it’s a marketing catch phrase. At worst it’s a cynical attempt to separate eager VCs from their cash. Interestingly, Wikipedia starts its article on “On-Demand” with the phrase “This article is in need of attention” – I couldn’t agree more!
The problem with ill-defined marketing catch phrases is that they are open to continuous re-interpretation. Take “open” for example. In the late 1990s, most of the Unix vendors spoke of their software as open with the clear implication that anything that wasn’t Unix, especially Windows, was closed and proprietary. Microsoft could have reacted by selling the benefits of their closed system but they took a simpler approach – they simply hijacked the phrase ‘open’ and made it their own and within a couple of years the world was talking about industry standard, open software from Microsoft and closed proprietary Unix systems.
One well worn approach when you realize that your special phrase has been stolen is simply to find a new one. Hence we see the emergence of Software as a Service, or SaaS.
What does all this mean? Why are customers interested in SaaS or on-demand software? Do they care about multi-tenancy? Is this really the end of enterprise software as we know it?
Customers’ attraction to SaaS, or on-demand, is understandable. It can give them the ability to get up and running quickly with a standard application that meets 80% of their needs. It releases them from dependence on IT organizations burdened down with the maintenance of legacy systems. It converts up-front capital costs into monthly operating expenses. It can reduce TCO by using shared infrastructure and simplified operations. All of these are powerful drivers for new ways of delivering software.
The cult would have us believe that only applications that are developed from the ground up with a view to this new world can be successful. But customers don’t really care – they want the end result and aren’t interested in the technicalities, providing it meets their requirements for cost, functionality, reliability and security.
In fact, many of the current SaaS providers are finding themselves in a technology trap – running complex infrastructure to support many users at extreme scale is extremely difficult; and like classic Greek tragedy, the strengths of youth rapidly turn into the weaknesses of maturity.
It’s even harder if you are committed to a multi-tenant, shared infrastructure where ensuring the integrity of data and the security of the network connection leads to ever more complex infrastructure solutions. A different approach is an on-demand solution that leaves the infrastructure to others. For instance, a software that is grid enabled allows users to have their own, separate, database.
Complexity can be a vicious circle and a grid architecture can keep it simple. It can exploit the increasing commoditization of hardware and infrastructure software to reduce costs and allow for cost benefits promised by the on-demand vendors without having to compromise or ‘dumb down’ existing software.
This could be the beginning of the third generation of on-demand computing. But if we called it that, then we would too be guilty of trying to start a cult.
Robert Youngjohns is the CEO of Callidus Software. Prior to joining Callidus, he held positions with Sun Microsystems and IBM.