My brain's been immersed lately in thoughts about Software as a Service ("SaaS") and how it plays into the Enterprise Apps/ERP world, especially for Oracle Applications customers…my mind is swimming in SaaS soup. Some of this thinking has been admittedly prompted by the news and discussion rooted in this week's Dreamforce 2009 conference (19,000 attendees at this year's conference? I think Salesforce.com has some traction in the marketplace…), as well as Workday's upcoming Webinar on "The Future of Software". Some of my thinking was also prompted by an item from a PowerPoint slide near the end of Larry Ellison's keynote at Oracle OpenWorld, indicating that Fusion Apps will be available as an SaaS offering.
In my thinking, which includes exchanging lots of ideas with other folks in the Oracle Apps user community, I'm finding that there's not much depth of knowledge about the the SaaS model or it's benefits. You can find a nifty primer on the SaaS model here. So far as benefits, three leap immediately to mind:
- No software installation or maintenance: SaaS eliminates the need for customers to memorize a gazillion-page set of implementation and user guides; patch application and rollout is handled by the SaaS provider - the work of doing so is ideally transparent to users.
- Shorter deployments: your SaaS provider can implement in days or weeks rather than months.
- Emphasis on the core of the business: many of us are currently in situations where we spend significant money and time simply keeping our in-house installations running. In the SaaS model, those resources are freed up to focus on the core of the business - you can focus on the business output rather than the IT tools.
The SaaS model makes sense to me in general when I look at Fusion Middleware and Fusion Applications. The technical stack for both includes quite a few "moving parts" working together…it's a higher level of technical complexity than we've seen in Oracle products up to this point (that's part of the trade-off for apps and tools that offer more flexibility and increased business value). Moving to a SaaS model frees apps customers from the burden of managing that complexity.
So let's be a little more specific. Take the example of my own shop which, when you get right down to it, is not much different from many of the medium-size to large-size enterprises out there in terms of our Oracle footprint. We maintain our own infrastructure for our business systems: hardware servers, storage, network, software environments…the works. We have an Oracle E-Business environment, version 18.104.22.168 (with a significant number of customizations) running on iAS (with components of Fusion Middleware like BI Publisher) and the 11g R1 database. We also have a custom applications environment, running on an 11g R1 database and 10g AS. In that custom environment, we leverage OC4J fairly heavily. However, now we'll have to make the transition to WebLogic for our custom environment - OC4J has faded into the sunset. That transition is not a small matter. The high-level Oracle guide on the migration is 32 pages long (insert nervous coughing here). To make matters even more interesting, the E-Business Suite will continue to utilize 10g AS/OC4J. And did I mention that we're about to start up a project for upgrading to R12.x.x? We should be able to knock it out in around 18 months or so, in parallel with migrating the custom environment to WebLogic. So now my shop will need skills in maintaining both 10g AS and WebLogic. Can you visualize the big, black kettle boiling up with some really bad juju…more skills required, more complexity in maintenance, greater consumption of resources with keeping the two environments cooperating (yes, we do have some integration between the two). Yeah, let your mind run down the track with this scenario for a minute. It's my future. That is, of course, unless we shield ourselves from some of this cost and complexity by moving one or both environments to an SaaS model, letting the SaaS provider deal with part or all of this. And we could avoid any additional hardware costs while we're at it.
So it seems to me that my own shop makes a pretty good use case for SaaS. Whether we'll do it or not remains to be seen…the decision is not in my hands. Still, it doesn't take much to build a case worth investigating.
Now I've hit the SaaS concept pretty lightly here. And I'm admittedly really late to the SaaS party. But I'd like to hear what you think. Does it make sense for your organization? Let's keep this idea going, for better or for worse, in the comments!