JES is a new way to deploy and license the infrastructure of an Internet business. Sounds innovative, but I would say it just shows that there are finally some smart people at Sun thinking about doing things right.
JES is built on the idea that most portals are really built of a dozen or so applications like the portal application framework, Web server, application server, calendar, e-mail, instant messaging, LDAP, single sign on and many others. All of these pieces are usually put together one at a time and with a lot of work dedicated to getting them to actually work as a single unit. My past is littered with with the wasted and long hours or complete failures because of applications not mixing well. What Sun has done is create a clean integration of all of these tools that runs smoothly and is tested as a unit to discover integration problems. The end result is a lot of the core business software is up and ready to use in a few hours. Think about the Integration of all the software components that make up the infrastructure. They are like building an airplane from little pieces and a blueprint. Instead, JES is a fully assembled and tested Lear Jet.
Like I said, not really innovative, but it takes a lot of work and politics to integrate all your standalone products into a single install, with a single interface for management, and one application to license, and get maintenance agreements on. My hat is off to the guy that proposed such a departure in the Sun way of doing business.
But now to the second part of the innovation of JES that might give JBoss a run for their money. The JES set of components in the common install management tools is available with a cool pricing model. Think about having it all with a portal application framework, Web server, application server, calendar, e-mail, instant messaging, LDAP, single sign on and many others for just $100 a year per full time employee. This is just the business model for pricing and not some magic number that is part of the CPU count, the number of gigabytes of storage, e-mail accounts, your first born child, and a chariot that turns into a pumpkin at midnight. The key is that this is just a way to price the system, it is not the price for the number of actual users. This just buys support for the tool and your software updates.
This is more cool and obvious if we look at a company that has 100 employees for a cost of 1000*$100 (it is free up to 99 employees but there is no maintenance). Now even though the company bought the 1000 employee license, they can have a million customers and their 1000 employees use the system without any additional costs (except for the pesky hardware). A good example is Google - with 5,000 employees and 7 million customers, Google would only need to pay for the 5,000 employee license. The economies of scale look like open source with maintenance.
The tools are all based on open standards, not open source. Seeing and modifying code is not an option. But when you have a good maintenance agreement, the risk and cost of maintaining open source versus a maintenance contract is about the same or better since most of the components are usually used as is and not customized. Because they are tested as a suite, there is even less need for the IT department to be fiddling with code that should be considered infrastructure rather than custom applications. Of course this is still a modular system and you can swap out almost any piece with an open source or commercial alternative (like the application server).
