SAAS is no longer the only CRM / ERP Cloud Computing Model

SAAS, software as a sevice, has been the dominant delivery model for on-demand applications for some time. It replaced the initial ASP (application service provider) model that emerged at the beginning of the decade. The vast majority of applications that we use online are delivered using this model - this blog, for example, and all of the Web 2.0 applications that we obsess over.

But a new model is emerging which takes the best of both worlds - cloud infrastructure.

So what is it? Who's using it? And why should you care?

The emergence of cloud infrastructure options is in some respects back to the future to the pre-2000 world of collocation and managed hosting. But in many ways, it is radically different. Cloud Infrastructure providers such as Amazon, and to a certain extent Zoho and Force.com to name some household names, provide access to virtualized servers (ie large server farms that function like lots of little servers) within a professionally managed hosting environments. You buy just the processing that you need. This democratizes the application hosting game.

The emergence of these virtual servers is a potential game-changer for the CRM and ERP markets. Until now, the main advantages proclaimed by SAAS market leaders were that they were always on and available anywhere with no servers or software to maintain. Building these facilities was expensive and complicated and presented a barrier to entry that even large entrants have stumbled across. Another issue with these large shared facility customers is load balancing over shared resources (ie if someone sharing your server uses a lot of processing power, your application may feel it).

Now, smaller entrants such as SugarCRM and OpenBravo ERP are offering preconfigured bundles that can be rapidly installed in Amazon's Cloud. This approach may present the best of both worlds for some customers. This is not to say that the products emerging are functionally fully competitive with NetSuite for example but in time this will start to happen. Here are the advantages and disadvantages that I see:

1. Still in the Cloud: The applications are fully web-based and are indistinguishable from their true SAAS comparators in this respect. Customers don't need any local infrastructure other than an Internet connection and a browser.

2. Less Expensive: The applications that are leading this charge often come from the OpenSource world (which will be the subject of an upcoming post) and are often dramatically less expensive to own and operate (or even free except for the Cloud platform costs). On the other hand, these applications are less mature in most cases.

3. No lock-in: It's your "virtual" server, it's your data. Migrating from one system to another is never easy but at least you can access you data directly.

4. Extensible: Many of these systems offer the ability to integrate with other cloud reporting tools securely within the server environment. For example, data for an ERP application can be stored in MySQL and be made accessible to a reporting application in the same cloud at near LAN speeds or another cloud via web services.

5. More competition: SAAS Market leaders such as have complex and expensive infrastructures to deliver their applications and have shown impressive uptime. Now, however, many small providers can enter the space without building the same level of expertise or infrastructure. It's not the same as the dedicated engineering teams that the large players have but that may not be necessary.

6. Beware the small software company: One thing you definitely get with NetSuite and SalesForce is SIZE and reliability. They aren't going away anytime soon. This is just as critical to the new companies' customers - after all, we are talking about systems that run your whole business, not the latest photo site.

So, my overall take. SAAS is becoming less a category of software than a delivery model for web-based software. The emergence of cheap cloud infrastructure will create new opportunities, new business models and ultimately more choice.