Automating the Deployment of Applications

Following my papers on the future of software and the most recent one on the acquisition of SUN by Oracle, I continue to see signs that point in the direction of reducing the costs of application deployment. The first company that I talked to recently is Phurnace Software. Phurnace is specialized in the deployment of Java EE applications. The solution includes a discovery of the target environment, including the typical configuration settings. Then it does a validity check of all the settings, provide a "what if" sandbox to see the impact of setting changes, deploys automatically the application and finally provide a complete report on what was done. This is already very close to the complete "life cycle automation" concept that I think is the future of software applications. Then I talked to rPath: This is a very similar solution, but the end result is a run-time version of the application that can be deployed as an image on different platforms, from bare metal to virtual environments. rPath has already a number of followers in the ISV world, and is now looking at the enterprise one. Finally I also had a briefing with XebiaLabs in Holland, which appears as a direct competitor to Phurnace.

All these solutions are extremely interesting and they are a path to the future of software in the sense that they are directly usable today, and will provide significant savings (between 50 and 90% faster deployment according to the vendors, but also an error reduction of 50%), but also because they are a glimpse into the future. Combine this with cloud-in-a-box platforms, SOA and packaged application modules and the future is almost there, or a least a significant cost reduction in application life cycle can be gained.

What is also interesting is that they are several of these solutions that have surfaced within a year, and solution proliferation is the first step in innovation adoption. As I wrote in a paper on innovation adoption 4 years ago, the sequence for adopting a new technology is the following:

1. Innovators enter the market on faith. Although IT innovators believe that they offer a better mousetrap, there is little hard evidence to support the claim.

2. New vendors join the market and legitimize the innovation and its market potential. Several innovators come to the same conclusion and bring a similar or an improved solution to the market. (This is where we are today)

3. The number of vendors builds up. As the supply side increases, the technology improves and starts pulling the demand side. Vendors compete essentially on technological advantages, and competition improves the product.

4. Larger vendors enter the market, and some consolidation occurs. Technology improvement becomes process improvement — in technology manufacturing, compatibility with vested interests, or in rapid complement development. Smaller vendors lack the means to structure the complementary technologies around their innovation.

5. Process improvement leads to cost decrease and quality increase. Larger vendors build market infrastructure and solution awareness. Complete solutions are now available. Large vendors have the distribution network and the support structure necessary to support widespread adoption.

6. Sales take off. The upward segment of the adoption curve begins.

I really believe that this is where we are. Automation is the key to IT"s future.

Regards,

JP Garbani