It seems likely that in coming months many businesses will embrace Software As A Service (SaaS). This is where an application (or applications) are accessed via Internet connections and their usage is rented instead of owned.
Many of us already do this with services like Google’s Gmail or Salesforce CRM. We don’t own the software or install it on our computer. Instead we rent the usage and access the program from any Internet connected computer.
Yes, this indeed does look a lot like a return to the old days of IBM Mainframes – except in the old days most of us didn’t put mainframe terminals in our homes so we could work or access the mainframe via an iPhone.
Once businesses start using SaaS in far greater numbers there will be another issue they have to consider.
That issue is the claim that SaaS providers make about their reliability or uptime.
Most providers can be fairly confident in promising 100% — or nearly perfect — uptime.
This is because in reality some of the fine print in the hosing hosting agreements are tilted in the software company’s favor.
Three tricks for artificially making uptime look better than it might be:
- Not counting individual downtime — only networkwide. So if your site isn’t reachable that doesn’t as downtime unless the other users on the system also can’t access the network.
- Not counting scheduled maintenance
- Not counting small amounts of service interruptions