Saturday 14 September 2013

The Future of Information Technology

In the early days of information technology (IT) the path to new enterprise systems was clear-cut; a company would acquire a mainframe (or rent time on one); hire a legion of technicians; and crank out applications. The advantage to this approach was that each application was custom fit to a specific business problem. The downside was that the overall expense of the resulting systems (hardware, labor, and maintenance) was prohibitively expensive for all but the largest organizations. In addition, designing and implementing these systems was often error-prone and extremely time-consuming.
In the 1980’s companies found that they could leverage relational database technology to create applications on “open systems” (Unix-based mini-computers). In most cases the result was higher performance systems at a fraction of the cost. Database vendors such as Oracle, Sybase, Informix, and Ingres dominated this space; with the help of large-scale open systems pioneers such as Sequent, Pyramid, Stratus, and Tandem. As the market expanded in the 1990’s, the early systems vendors were eclipsed by commodity offerings from Sun Microsystems, Hewlett-Packard, and IBM; and later even PC-based servers.
In the 1990’s the industry experienced another software revolution with the adoption of packaged enterprise applications. The goal of packaged applications was to offer a general solution for the most common problems solved by IT organizations – accounting, finance, manufacturing, human resources, and others. Companies such as Oracle, SAP, Peoplesoft, JD Edwards, and a handful of others led the charge.
“Vanilla” packaged solutions are typically customized to meet the specific needs of individual organizations. By leveraging packaged applications organizations were able to deliver systems faster and reduce costs. Reduced costs are possible because research and development costs for applications are spread across a larger base of enterprise customers.
With that said, packaged applications are not an ideal fit for every organization. Packaged applications make the most sense for applications that are (a) non-strategic (e.g. systems that do not offer a competitive advantage), and (b) a close enough fit so that customization does not erase all the savings. Building on packaged applications is now well established.
Still, packaged applications are not without limitations.
  1. Customization can be very costly – it is not unusual for software customization to cost more than the package itself.
  2. The organization must still pay annual software maintenance fees to stay current and receive support. Maintenance fees are not as expensive as the original software license, but they still represent a significant ongoing expense.
  3. Periodic upgrades, which are also required to stay current, can take as much effort and expense to install as the original software.
  4. Most of these applications require specialized technical resources to maintain/manage.
The most recent advance down this evolutionary path is On-Demand applications. On-Demand applications are developed and hosted by a service provider – licensing is usually based on usage (e.g. $10/user/month). The early pioneer and best-known provider of On-Demand applications is Salesforce.com. The goal of On-Demand applications is to avoid the cost usually associated with purchasing and customizing packaged applications. Companies can have the full power of a packaged application and only pay for what they use.
These early On-Demand applications are historically utilities (e.g. CRM) that serve as an adjunct to core enterprise systems. However, there are signs that this is changing. New vendors have emerged in this space and many of the traditional packaged application vendors now offer On-Demand versions of their product; of course, they would still prefer to sell packaged software. The next wave of enterprise applications will take this concept one step further by providing “hosted” applications. Hosted applications serve as a substitute for core IT applications. The application provider essentially serves as an out-sourced provider of core applications. They provide the software, systems, and technical experts.
There is no need to own application servers, computing/network equipment, or specialized technical experts. The service provider provides all of these items for the enterprise. The result is even lower cost because the cost of hardware, software, and technical resources are spread across a larger base – rather than just R&D.
The migration to smaller faster computers using multi-core technology and virtualization create even more value for hosted applications. Advances on these fronts give service providers the ability to leverage faster/cheaper hardware across an even larger base that includes multiple enterprises running multiple applications.
Are you ready for the next wave of even lower cost high performance hosted computing solutions? Can you afford not to be?

No comments:

Post a Comment