Amazon has clearly been leading the way in API-fication by providing a broad range of fundamental software services packaged as APIs. From the basic EC2 compute and S3 storage capabilities, they have expanded to now offer more than 30 services across infrastructure categories of compute, storage, networking, database, deployment/management and messaging. All of these components are incredibly valuable and important, but an application developer still has to construct higher level business processes from these fundamental building blocks. In addition, they have launched the AWS Marketplace, which is a catalog of hundreds of software packages that cover everything from application development to traditional business software. However, this marketplace has only taken the first step in making it easy to install and deploy software applications or stacks as machine images. They haven’t yet enabled third-party companies to provide application components packaged purely as APIs.

We are still in a time of transition. More and more technical organizations are realizing they really don’t want to install and manage software — even if it is running in someone else’s data center. The preferred model is to rely on software service providers who can (and must!) deliver a high quality services. Amazon has proven that developers are quite happy to outsource the data center, and Salesforce.com has proven that end users and IT organizations are content to simply consume a Web-based application — but what about all the layers in between?

Enter business process APIs

Enter the providers of business process APIs.  These APIs have three characteristics that distinguish them from infrastructure-level APIs:

-       They are truly plug-and-go, requiring minimal programming, and thereby approaching the promise of *legofication* that I alluded to in my last post;

-       They are usable by a broader range of developers, beyond the most technical ones, including HTML designers and higher-level coders;

-       By targeting business processes such as payment processing or expense management, they are directly linked to revenue generation.

The following table lists some of the new API services providers (APIsps) who provide packaged business process services. (Note: my company is an investor in Alfresco, Gigya, SmartRecruiters, Rubicon and Viralheat.)

In analyzing some of the data from ProgrammableWeb, it appears the infrastructure services are those getting the most reuse. Not surprisingly, the basic capabilities of mapping, messaging and search are the top three. However, many of these basic services are free (or very cheap), so they might not drive significant revenue. Those involving search and transactions (Amazon eCommerce and eBay would definitely provide more lucrative revenue opportunities. There is a long tail of services, which I simply aggregated under “Other,” but within that group are undoubtedly some high value business processes.

Moving on from packaged software to SaaS

After years of packaged software use and the transition to open source, applications seem to be well down the path of SaaS-ification, with the next wave involving the decomposition of the various application services into APIs. One of the comments (thanks Darren) on the previous article reminded me about the long path we have been on to properly package these APIs. It all started with portable DLLs and shared libs, and then we went through a bad phase with DCOM and then moved onto to XML and SOAP which were unfortunately too prescriptive and constraining. Thankfully, http and RESTful services emerged to provide a Web-style stateless approach. Each of these iterations has made it much easier for developers, but I am sure there are amazing ways to further improve the packaging and consumption of APIs. The LEGO (Lightweight Enterprise Gadget Orchestration) concept was an attempt to push the community to think about what comes next. And the NextStep Interface Builder idea that I mentioned last time is another area that a bunch of young companies seem to be exploring.

We are still early in the APIsp adoption phase. Entrepreneurs and developers should identify the top business services and work to create elegant and simple ways to drive these processes through code and beautiful end-user experiences.

Robin Vasan, managing director at Mayfield, invests in cloud, SaaS and mobile technologies. Some of his current investments include Alfresco, Couchbase, Marketo, Centrify and Webroot. Past successes include Akimbi, Trigo and webMethods. Mayfield has also been involved in such other leading companies as 3Com, 3PAR, Citrix, Concur, Legato, Nuance, Tibco and Vantive.

Image courtesy of Flickr user kennymatic.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Sector RoadMap: Content personalization in 2013
• How fourth-quarter 2012 will affect IT spending in 2013
• The fourth quarter of 2012 in cloud

A common question surrounds AWS and the Yankees: Are they evil empires, the best places to be, or both?

I recently did at a bit of remodeling at home and as part of the process got rid of about half of my book collection. I only kept those that I had a special attachment to or would be hard to find online. After I got a Kindle, physical books started looking “old” all of a sudden. Amazon has managed to make paper books pretty much obsolete in the span of a few years. This week’s announcement of the AWS Marketplace is going to make you feel the same way about running your own server hardware.

Why use a marketplace?

Traditionally, the cloud hype has been mostly around infinite storage capacity, scaling up, load balancing, elasticity, and capex vs.opex. I am going to let you in on a little secret: most companies out there are neither Zynga nor Netflix and their applications do not require scaling to hundreds or thousands of servers to accommodate peaks of demand. If capacity requirements grow over time, the pattern tends to be predictable.

However, nearly every company needs to run document management systems, CRMs, wikis, bug trackers, project management tools and other web-based software. Server applications tend to be tricky to setup and require a non-trivial amount of sys-admin knowledge to run and maintain. The AWS Marketplace encapsulates all that complexity and allows end users to discover, purchase and deploy complete server applications with one click. Though a CRM system is certainly more complex than a game that slings birds around, the marketplace concept does a good job of bringing the simplicity of the Apple App Store model to server applications running in the cloud. It has the potential to revolutionize how most businesses consume IT.

At BitNami.org, we package open source applications so they are easy to install in any environment, including the Amazon cloud. These packages are available free of charge, but no matter how much we try to simplify the experience, there is still significant friction in the process in terms of just making the decision to give an application a try. The AWS Marketplace removes this friction by making it easier to discover and deploy applications.

This significantly lowers the barrier of adoption of cloud computing at the departmental level, making it easier for business units to bypass traditional IT. Why wait weeks to have a server delivered and setup when you can get pretty much the same result by whipping out your credit card and paying $50 a month for a small instance running the app you need now?

It’s now or never to challenge Amazon

However, the initial version of the AWS Marketplace is not perfect and there are still issues that will need to get ironed out around licensing, payment and support, including how to handle upgrades or integrating existing activation schemes with the  one-click deployment model. But the first stake has been put in the ground.

There are other cloud providers, though, some of whom have their own libraries of applications. Predictably, after this announcement, the rest will look into following suit and launching their own marketplaces. But most of them lack both the huge customer base and momentum of Amazon, making it harder to convince ISVs to come onboard. Once the network effects fully kick in, there will be even less incentive for application vendors to support other platforms.

The thought of a single player dominating an industry is always uncomfortable. However, unless its competitors get their act together soon, the dominance of Amazon in the cloud market will continue to grow; it’s already starting to look eerily similar to that of Microsoft in the 90s.

Application vendors have limited resources, and if they only have the bandwidth to participate in one or a handful of app stores, they are going to choose the ones that offer the greatest revenue potential. Typically, that means the ones with the greatest number of users. Amazon is clearly in the lead with respect to user adoption, and is therefore going to be the most attractive marketplace for most vendors.

Daniel Lopez is the CTO of BitNami, a leading provider of application images for Amazon Web Services and other cloud and virtualization platforms. 

Image courtesy of Flickr user Eric Beato.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Examining open hybrid cloud options for the enterprise
• Infrastructure Q1: Cloud and big data woo enterprises
• Migrating media applications to the private cloud: best practices for businesses