Brian de Haaff's Blog

Cloud computing, utility computing, SaaS, on-demand, grid computing, etc. The jargon is back. The acronyms, the vendor jockeying, the funny names, and yes, the confusion. But does it really matter any more than it did 10 years ago when you couldn’t grab a bite to eat at the local Silicon Valley deli without hearing those three special letters – ASP – that were at the forefront of our collective brain when we talked about applications running in the cloud. (They usually came right after the other three letters: IPO, but let’s not go there.)

I remember well because I was in the vortex of the last cloud revolution. In fact, for better or worse I coined the category Application Infrastructure Provider (AIP) in 1999 while working at Concentric Network, responsible for marketing for the hosting product line and data center services. We were responding to perceived customer demand and media attention for ASPs (Application Service Providers). We had bandwidth, we had space, and we had servers—put it together and you get Infrastructure and the name. We productized the service and Network World named it Product of the Year. It was not our best seller.

So what’s changed in 10 years? Will the outcome be any different when we talk about businesses running software (or storing data) offsite and accessing it via Web browsers? Will they actually do it en masse? The pundits say, “Yes!”, but will it really happen this time?

I do think we’ll succeed this time because things are different.

There’s still lots of hype—note a recent June Deutsche Bank report that states “We expect SaaS revenues to grow to $56.5B in 2015 (nearly 70% of the overall market in just 8 years time).” But, I think that things have changed for cloud-based on-demand services in five major ways that will alter our collective computing future:

1. ASPs (and AIPs) were single tenant driven, which means that you were responsible for piecing together your own infrastructure (whether you provisioned your own or rented it from the provider) and also for creating or buying the software that ran on it. This was akin to moving your own data center to a third party facility. You still had almost all of the same IT headaches. Today, cloud-based, on-demand services are multi-tenant. Vendors pull together infrastructure and resources, and automatically allocate and balance them so many users can simultaneously use them.

2. Because the platforms are multi-tenant, the vendors can cost-effectively take on more responsibility for the underlying technology. Businesses now take advantage of the scale and efficiency that comes through the vendor’s ability to serve multiple companies on the same computing platforms. This is a fundamental shift because businesses are no longer responsible for maintaining the infrastructure themselves, no matter what type of on-demand service they choose.

3. Salesforce.com. Enough said.

4. Ubiquitous broadband. Enough said.

5. Every established software company is making significant investments in moving their offerings to an on-demand, pay-as-you-go business model (not to mention the heavy hitters like Amazon and Google that are already offering their own on-demand services).

In addition, to the fundamental changes that have taken place, the benefits of on-demand services for businesses and their end users are clearer. The advantages include instant access to the software that businesses need to improve their operations, cost efficiency of pooling their software and computing needs with other companies, reduced risk of failed software implementations and tapping strained internal IT resources, and flexibility in terms of being able to change providers. This all comes with the tradeoff of relinquishing some control, but it appears to be well worth it to most companies.

So, it appears that on-demand services are really happening this time, so let’s try to agree on a framework for talking about the various offerings. And let’s start with a diagram to simplify the discussion:

When you boil it down, on-demand services are all cloud computing based. And there are really only three types of them:

1. Software-as-a-service (SaaS) is really where this wave started and it makes up the largest bucket. These vendors typically provide a single application or a cluster of small applications targeted at a general business task (e.g. CRM, billing, HR, etc.). Examples include Salesforce.com for front-office activity, Netsuite for back-office activity, and Paglo to manage IT and operations. They are “application specific” meaning that the business uses the functionality that the vendor provides rather than building their own. The vendor builds out and manages both the infrastructure and the software and the customer only needs to input or manage their data in the applications.

2. Platform-as-a-service (PaaS) offerings get slightly more complicated. Here the vendor provides the infrastructure, software stack, and typically a data source. This is an emerging cross-over area that successful SaaS and online vendors are now promoting. This is often where the SaaS provider opens up their platform to third party applications (whether built by a business to serve its own needs or developers to sell to the businesses that are already using the core SaaS offering). So while this expands the application functionality of the SaaS provider, the applications are still related to the core offering. For example, Salesforce.com’s Force platform is designed mostly for customer-relationship oriented applications that leverage the data in a Salesforce.com customer’s account and Facebook’s F8 platform is ideally suited for additional applications that tie to the user’s social network. We would also put Google’s App Engine into this category (despite the fact that they do not provide a data feed).

3. Pure cloud computing platforms provide the server infrastructure and the business is responsible for their own software stack and ultimately the application that runs on it. The business can run just about any software that it wants as they are “application agnostic.” The main force (and a good example) in this market is Amazon Web Services

Here’s one way to compare the application flexibility and level of responsibility that the business takes on with each option:

It’s true, ASPs are dead, but they set us on the path to get on-demand services right. Advancements in technology and service delivery have improved the value for businesses to adopt these services. Now, let’s just start being clearer about what’s happening, so businesses can best choose the services that will help them succeed.

We are well on our way – long live the cloud!

[This article appeared originally here and is republished in full with the kind permission of the author.]