Friday, January 2, 2015

5 Reasons to Modernize Mainframe Computer Systems

by Brenda J. Christie


Modernize Legacy Systems

It is probably common knowledge, that humans do not embrace change, at least not initially.  It is easier to be comfortable with known factors than to face something new.  Fear of change extends to corporations as well.  After all, corporations are made up of humans, too.

However, as mainframes continue to consume a great chunk of the corporate IT budget, ignoring discussions about whether to modernize and taking action to modernize legacy systems is harder to do.  Typical questions concerning continuing to operate a mainframe, as opposed to outsource application to the Cloud, can easily revolve around ROI, what tangible product does it produce, how do we lower energy consumption running a mainframe, and a host of other questions which can be posed by financial and budget analysts.  These questions would most likely also include "Why Keep The Mainframe?"


#1 UI (User Interface)

Shirley S. Savage's article, "Updating Applications Makes Good Sense", which appeared in the November 2014 edition of IBM Systems Magazine, explores some of the discussion around mainframe modernization.  The #1 business driver of modernization, according to the article, is UI (user interface).  I think I have to agree with her.  For a variety of reasons, people today are more likely to conduct banking transactions on a smartphone or tablet than they are to walk into a bank and conduct business with a human being.  Similarly, they will use online travel applications, rather than go to a travel agency. They can make dinner reservations, hail a cab, get multiple car insurance or even life insurance quotes in the time it takes to complete a single bank transaction with a teller.  So, 'Yes,' the user interface is #1 because it is the gateway to increased market share.  It is also a platform to which people have grown accustomed.  

UI also refers to integrated development environments (IDE's) for application developers.  This speaks to cross-training wherein someone who is familiar with an adept at using a Java Development Environment, for example, can more easily work in a mainframe environment because of the familiarity.  This in turn, would somewhat mitigate the mainframe skillset shortage.  Some of the mainframe development environments currently available can be found in my October 13, 2014 post, "COBOL Development Environments."

#2 - Maintainability

A second driver of modernization, according to the article, is maintainability.  I found this argument a bit hard to agree with.  It alludes to the possibility of millions of lines of spaghetti code holding an application together.  The Year 2000 problem was the perfect opportunity to rewrite spaghetti code to a more structured, object-oriented codebase.  I would like to think that is more or less what happened. The second driver, did however, underscore the mainframe skill shortage and its effect not only on modernization efforts but also with regards to effecting change driven by regulatory agencies or business.  The idea here is modernizing legacy applications will make it easier and quicker to respond to changes in the regulatory or business environment.  There is a degree of doubtfulness within this assumption.  In the absence of a methodology and transparency, it would appear to be just as easy to write bad code in a legacy language as a modern language.  Modernization alone will not prevent the introduction of spaghetti code but when combined with the utilization of a development methodology and software framework may reduce its incidence.

#3 - Performance

Ms. Savage's third driver of legacy modernization, performance, speaks to difficulties in scalability which can arise from the mainframe.  I see a few problems with this assumption.  The first assumption is that the machine itself is 100% utilized hence the need to scale the application to accommodate increased workload.  This argument conflicts with the many articles I have read which cite C-Suite executives wanting to use more of the mainframe in order to improve TCO, Total Cost of Ownership.  In essence, they want to run more applications from different business groups in order to improve ROI and spread the cost of maintaining a mainframe.  Indeed, approximately 46% of respondents to BMC's 2014 9th Annual Mainframe Survey expected mainframe growth in MIPS consumption in the next 24 months.  Also, with the advent of zIIP and other specialty engines it is possible to free up general mainframe capacity and improve performance.    With regard to performance, then, clients can become frustrated at not being able to scale up as quickly as they like, but there are solutions such as zIIP which can solve or mitigate that problem.  With fore-knowledge of the client's business goals and advanced planning, the scaling problem may never occur at all.

#4 - Database

The fourth business driver for modernization centers around database performance.  I agree with this one completely.  Databases need to accommodate data from myriad sources in order to enable quick business decisions based on the big picture.  Refer to my December 14, 2014 post, "2014 BI Leadership Summit" for a discussion of enterprise architecture supporting Big Data.


#5 - Globalization

Globalization is cited as the fifth business driver in support of modernization.  I agree with this somewhat however, I do believe that some of the examples (need for multi-currency capability) are a bit dated.  As far away as the late '90's and early 2000's, I worked on mainframe trading applications (Commodities, Unit Investment Trust, Money Markets) which received daily pricing feeds.  The ability to quote prices in various currencies is already an integral part of mainframe computing. However, in my more recent work with hosted applications, changing from one currency to another, as well as from one language to another can often be accomplished by changing an installation setting.  The same should be true for the mainframe.


Summary


In summary, IBM Systems Magazine's Updating Applications Makes Good Sense is a good starting place for beginning the discussion on modernizing legacy applications.  It ends with some good advice on adopting a modular approach which both minimizes risk and makes good sense. Attempts to modernize an entire application would most likely take longer than most business users are willing to wait, especially if they see no results.  

Bye for now,

Brenda J. Christie

No comments:

Post a Comment