Modernize the architecture of legacy RPG applications

Modernize?  Modernize What?

There are in fact different flavors of modernization.  One that has been popular for sometime is the type that is facilitated by tools that improve the user experience, typically aiming to end the 5250 interface and replace it with a GUI.  This obviously gives a facelift to legacy applications.

That sort of modernization, however, does little rein in IT costs, extend the useful life of applications, increase responsiveness to business needs and so on.

To meet those goals it is necessary to modernize the back-end, the server side, the core of the application, the RPG or COBOL code.

This sort of modernization offers considerable benefits to organizations and is what we call “value driven modernization.”

What is value driven modernization for AS/400, iSeries and IBM i systems?

Successful modernization projects are driven by the business in pursuit of such business goals as improved bottomline, higher customer satisfaction and retention, quicker time to market, reduced operating costs and so on.

Remaining lifetime TCO for legacy applications can be evaluated as part of Application Portfolio Management.  With 80% of typical lifetime software costs going towards maintenance and enhancements it is advisable to periodically project remaining lifetime costs for applications and determine if there are strategies to reduce those costs.  In many cases modernization can reduced the overall remaining TCO for applications by reducing maintenance costs.

What options are there for modernization of legacy RPG and COBOL systems?

Databorough’s modules for re-engineering and design recovery facilitate several options for modernization.  In most cases modernization is fundamentally about re-architecting the application to make it more agile, most commonly following the Model-View-Controller design pattern.

Once the decision is made to implement an MVC architecture there are different options for languages and tools at each new layer of the application.

For the model and controller layers there are several Databorough modules that aid in data model extraction, business rule extraction, design recovery and reengineering.  The new components can be redeveloped in Java or RPG using a combination of automated and manual processes.  These components may then be accessed via stored procedures or web services from most types of modern client layers.

How is an iSeries modernization strategy developed?

While it may be tempting to technicians to launch right into engineering projects, experience has taught that the success of IT projects is heavily dependent on business involvement and thorough planning.  Key questions to ask are:

• What are the business objectives of the modernization project?
• How will the objectives be measured?
• What is the 5-10 year technology vision and how does this project fit into that?
• What staff skills can be leveraged, what needs to be developed and what will be contracted?
• What is the optimal participation of business team members?
• How do we construct a proof of concept project that will really tell us if we’re on track?
• How can we decompose the project into small iterations with frequent deliveries?
• How will we test the results to be sure we have not introduced errors?
• What needs to be communicated to the user community and what, if any, training will be necessary?

What does an IBM i modernization project look like?

The 50,000 foot view of most projects looks something like this, and a modernization project is no different, other than the details, of course.

Inputs – this of course is the current system, in whatever state is currently exists, with the knowledge of it that currently exists.  Typically this knowledge needs considerable improvement, which is where several modules from Databorough play a key role.

Technical Strategy – Most modern architectures follow some form of MVC, but there are many fine points to decide in how to implement it.  What technology should be used for the client layer for the primary users?  External users?  How will server functions be exposed?  What other systems and organizations impact technology decisions?  What skills do we have and plan for in the future?  What are the performance impacts?  How does cloud computing fit into the picture?

Tools – As just stated, a number of modules from Databorough greatly improve the success rate of modernization projects, but additional tools may also be useful for coding, testing, data preparation, management, etc.

Processes – What is the optimal sequence of tasks through a modernization project?  POC?  Multiple POC’s?  Is there any data work to be done?  Data design work?  Code preparation?  Start with massive documentation?  Who does what?  How can this be agile and iterative?  How will the system be tested, given that typically a full library of regression test does not exist?  There is no cookie-cutter formula for these projects, after all, they only occur every decade or two.  vLegaci specializes in this work and can accelerate your project and optimize your results.

People – While modernization may often seem like an internal IT project, history has shown the chance of success is heavily related to business and executive involvement.  That is a starting point.  How will you involve your own staff?  What will you outsource?  Will you need subject matter experts?  How will your people be positioned to support the system after it is modernized?

Outputs – As defined by the technical strategy and proven by testing.

Planning, measurements, management – Project management is the single most critical task in projects of this type. Planning for flexibility is key – there will be surprises.  Frequent delivery of software, as in agile methods, is desirable, if possible.   Businesses expect a schedule and budget to be met even though they are set with incomplete information.  Vision, creativity, openness and leadership are crucial for the success of complex projects such as these.

Contact vLegaci for a free initial discussion on how to shape your modernization project.