The New New Thing
In my last post, I talked about how Salesforce.com had bought back the venerable timesharing model by applying the new architectural model of multi-tenancy. Multi-tenancy made timesharing viable for the Internet era.
But as in all things tech, multi-tenancy is now being by-passed by the next “New New Thing” (NNT for short). As is often the case where new technology is concerned, the NNT doesn’t have a catchy name yet, but it is a clearly superior solution to an existing problem.
Most of the entrenched legacy vendors (Oracle, SAP, and Microsoft, etc.) who tried to climb on the multi-tenancy bandwagon found it impossible to move their enterprise apps to the cloud and keep a reasonable level of features. Instead, they focused on creating a large amount of FUD around security and customization. In general, the fears they created as well as the high cost of switching was enough to keep most of their larger customers in the on-premise fold.
For next generation SaaS players, including PatternBuilders, multi-tenancy was less alluring—not because of the “tyranny of the installed base,” but because we now have access to architectural components that solve the problems multi-tenancy addressed without the complexity. Tools such as:
- Platform as a service (PaaS), which is available from a variety of vendors including Rackspace, Amazon, Microsoft and Google.
- Cheap open source software.
- Metadata driven architectures for easy maintenance and customization.
- Virtualization for easy deployment and fair allocation of resources.
In aggregate, these technologies solve the same problems that multi-tenancy architectures were created to address while offering better security, customization, and more reliable performance. Together, this makes the added architectural complexity of multi-tenancy hard to justify. In fact, our bet is that multi-tenancy as an architectural approach will soon be considered obsolete. I will dive into the details of how PatternBuilders uses these technologies for our flavor of SaaS in another post.
More importantly, these technologies allow PatternBuilders to support both traditional enterprise deployment (on premise) and the hosted/SaaS/cloud deployment model – pick your buzzword– with very little additional effort. We made the decision to support both models because it would provide a broader user base and was easy to do. The fact is the cloud, multi-tenancy or not, is just not appropriate (and/or acceptable) for some customers’ uses. The reasons vary, depending on industry, but we don’t see those barriers to adoption going away in the short term.
What’s your bet? Do you think multi-tenancy is on the way out? Do you think on-premise applications will soon be as extinct as the Dodo bird? Let us know what you think!