Future applications

NETWORK WORLD NEWSLETTER: MARK GIBBS ON WEB APPLICATIONS 02/22/2006 Open source Web apps stomping over large-scale enterprise software By Mark Gibbs

It is intriguing to see the shift that is happening in the world of enterprise software. When these large-scale systems started appearing on top-end hardware it was thought that they must be fabulously well engineered and therefore worth a lot of money because they were supposedly capable of serving thousands of users.

What we found was that these enterprise solutions were sometimes, but not always, well engineered. In terms of functionality, they were often good but frequently not outstanding and while they could mostly handle huge numbers of users, they were very complex.

Were these enterprise solutions worth it? Roughly 50% of organizations got their enterprise solutions such ERP and CRM successfully running. Among that group, I suspect that many would argue that the value was not what they hoped. We can be pretty certain that the other 50% with their horror stories of grand and expensive failures would be definitive about the lack of value.

Those failures and consequent financial losses are becoming less common but not because the traditional enterprise applications are getting better. Nope, what has changed is that enterprise software is no longer seen as a priority and the reason for the change is primarily due, I believe, to the impact of Web applications and open source software (OSS).

What Web applications have enabled is a new way to provide enterprise application functionality built on service platforms that are inherently manageable and scalable; to wit, Web servers.

But what about the application code that drives the application functionality? The explosion of server-side scripting languages, the CGI interface, and the inherently granular nature of Web communications brought new architectures into the market that changed the rules and methods of development.

Now it is true that these architectures and their implementations presented their own challenges but compared to traditional enterprise development models they also disposed of some of the most restrictive elements of trying to develop large scale applications. One of the key issues was removing the need for high-end hardware.

Today it is the likes of LAMP (Linux, Apache, MySQL, and Perl/Python/PHP), Java (although Java appears to be somewhat waning), Zope , and the hottest platform at present, Ruby on Rails , that are redefining what enterprise-scale applications look like.

But all of these languages and services have one thing in common that makes them special and very different from traditional enterprise applications: They are open source.

OSS along with Web application technology are the biggest forces in breaking the log-jam of functionality, implementation, and performance that limited the success of enterprise applications.

If you look through back issues of this newsletter , you’ll see that OSS has slowly but surely become a dominant issue in the products and services I have covered.

Now while I dislike making grand predictions because now and then … oh, maybe occasionally … well, maybe quite often I wind up with egg on my face, I feel comfortable predicting that this year will unleash a flood of OSS-based enterprise-scale software based on Web technologies deployed both in-house and on a software-as-a-service (SaaS) basis.