IT Implementation Projects – It’s all about the data

Almost all software projects in organizations these days are driven by the business and rightly so. It will eventually get to the more technical people who will break the business requirements into more technical specifications. There is a tendency in such specifications to focus on what technology will be used at the user side to present results to the business with little thought put into the data and where it will come from. Eventually, a ‘tactical’ approach is used to get the data in the easiest way possible. This leads to various efforts to get at the data from proprietary integration efforts to the ubiquitous Enterprise Service Bus (ESB) technologies which collect the data in a form of ‘staging post’ prior to delivery to the consumer. At its very worst, it involves copying the data somewhere else so that it can be easily accessed by the new application required by the business. This leads to multiple issues surrounding this data:

• When proprietary integration is used
  
  
  
  • It can be difficult to quickly implement changes that are required.
  
  
  • A major effort must be put into monitoring and managing the environment.
  
  
  • Integration is usually only usable by one application.
  
  
• When an ESB or ESB like approach is used:
  
  
  
  • The data integration is then a multi tier process with the data being collected from different places. This then has multiple points of failure thus leading to stability problems.
  
  
  • If multiple trips are required to get at the data, the response times to users become delivery times.
  
  
  • While limited usage can sometimes work well, such solutions generally tend to have problems scaling to high volume usage.
  
  
  • How secure is the data? Each hop that it goes through is a potential security loophole. Also, who is accessing the data; these solutions make it very difficult to provide a user’s credentials all the way through the process which leads to problems with authorisations sometimes solved by using servers with ‘super user’ status.
  
  
• When data is copied:
  
  
  
  • It can quickly go out of date creating a bigger problem to then keep it synchronized with the original data.
  
  
  • If the new application also updates the data, a synchronization process will be required in both directions which can lead to update clashes.
  
  
  • If existing applications that access the original data are still running, there will be a time lag between when data entered on one system is seen on the other system.
  
  
  • What happens when another business requirement needs the same data or a subset of it?

What data that is required to satisfy a business requirement and how it will be accessed should be the first part of any technical investigation for a project.

Determine what data will be required as early in the project as possible as this will help in the following ways:

• It can clarify process issues to the business in relation to how the new system will interact with existing systems.


• It will identify if all the data that is required is actually available.

The SOA Gateway will enable you to access the data in minutes instead of weeks or months with custom built infrastructures.

• The SOA Gateway ensures that your data is only accessed by authorised personnel or processes.


• The SOA Gatewayhas monitoring features to help ensure that all is running well.


• The SOA Gateway will allow changes to be made easily and to enable access to additional data as required


• While there will be a cost to this, it must be compared with what it costs in time and money to implement and maintain a custom built infrastructure


• The SOA Gateway enables the creation of sensible ‘business views’ of the data (based on multiple sources) such that each represents a valid business object rather than simply representing the underlying structure of the data. Done properly, this sort of effort will pay dividends in the future when other business requirements require access to the same business view.

Having access to production or test data quickly presents the opportunity to quickly show prototypes to business users with actual data as against mock-ups. This can be extremely helpful to deliver the product that the business wants and helps to manage expectations.In summary, while the above efforts will cost a little more in time, money and effort at the start of a given project, the Return on Investment (RoI) will be easy to measure in terms of:

• Sharing the same data with existing applications so all parts of the business are working with the data ‘real time’.


• Delivering an application that satisfies the business requirement through early prototyping with real data.


• Delivering a system that only allows authorised users to access the data


• Delivering a system that can scale with the business.


• Delivering reusable business views that can be reused in future projects time and time again.