Case Studies:

Platform Independent Content Management

Our client has a video content management and delivery system that is used to control content across dedicated networks to dedicated servers. They had contracted with and accepted software implementing this management system from another vendor.

However, the management sub-system failed to meet it's own basic requirements, didn't integrate well into the balance of the system, and the functionality that did exist operated unreliably and very inefficiently.

In addition, a large number of proprietary products and proprietary extensions to open standards were used in the implementation, so expensive software licenses were required, along with their on-going maintenance costs. Coupled with the inefficient architecture, the true costs of the system skyrocketed.

Our client attempted to patch things together, performing basic triage on the sub-system to get a reduced set of capabilities operating reliably, and undertaking a program to remedy some of the performance problems. Some basic functions of the management system typically took over 16 hours to run, if they completed at all.

At a certain point, our Client chose to pull the plug on that implementation: the cost of rework was higher than the cost of redevelopment. The supplier of the system, who could never get it to operate correctly, walked away.

Project Recovery

Our client, having endured enough, made the courageous and painful decision to terminate the existing effort, and to start fresh. For this fresh implementation, they chose SilkSpeed. We had previously helped them with some of the triage and performance improvements. They needed the system completed on schedule and estimated cost. They wanted confidence that they wouldn't be faced with a repeat experience.

The Rewrite

SilkSpeed engaged its normal process of dealing with the business and technical requirements, of performing factoring and data modeling on them. We developed the new architecture of the system with our characteristic methods. We eliminated proprietary technologies where we could, and aggressively adopted open standards.

We finished with a system with some very desirable characteristics. The new implementation cost slightly more than they had spent on only the attempts to rework the previous sub-system. It could run on any common platform, such as Solaris, Linux or Windows NT/2000, on any common database product, Oracle and Postgres in this case. It had no proprietary standards used in its implementation.

And it was fast. How fast? That previously mentioned operation that typically took 16+ hours today takes less than a minute. This is without tricks or using proprietary database extensions, just good design practices.

Open Standards, Platform Independence

The mantra of Open Standards and Platform Independence are themes here at SilkSpeed. Open Standards gives product independence, protecting your investment in your systems, provides a competitive marketplace for products, producing high quality, competitively priced products. Platform Independence gives you the freedom to select your platform not on whether it supports your software, but on more important considerations: cost, reliability, scalability.

The open standards used for this project include CORBA, Java and Java Beans, SQL92, and various Java API and server technologies, such as JBoss, J2EE/EJB, Log4J, and Tomcat. It works equally well with multiple databases, on multiple platforms.

This project is a clear example to us of the value of open standards and platform independence for the development of scalable distributed processing systems.