Legacy Information Systems: Issues and Directions
Bisbal, J.; Lawless, D.; Bing Wu; Grimson, J.
Term Paper Done by Tay Pei Fang (S11)
A Legacy information system represents a 2.3 User Interface
massive, long-term business investment. User Interface is the main access to the
Unfortunately, such systems are often brittle, function of the software system. For most
slow, and non-extensible. Capturing legacy legacy systems, they usually have character-
system data in a way that can support based or command interfaces, which are
organizations into the future is an important but very complex and needs a long time to learn.
relatively new research area. The authors offer Reengineering these interfaces to make
an overview of existing research and present them more users friendly would extend the
two promising methodologies for legacy life of these legacy systems.
information system migration.
3 New Contribution to Software Engineering
Legacy systems, reverse engineering, New ways of managing the legacy systems
redevelopment, wrapping, migration, “Chicken have emerge due to the high maintenance
Little Strategy”, “Butterfly Methodology” cost and also the appearance of cheaper
and more flexible alternatives such as the
1 Introduction COTS.
Legacy Information Systems was written years Different types of reengineering the legacy
ago using outdated techniques, yet many are systems, such as Reverse Engineering, and
still mission critical and have a great impact on Forward Engineering have also slowing
business. These systems usually run on becoming a continuing part of the process of
obsolete hardware. They are usually very software development and evolution.
difficult and expensive to maintain in both
hardware and software. Legacy systems may New LIS migration methodologies, which
represent the years of accumulated experience offer more flexibility, better system
and knowledge. Thus it is not easy to upgrade understanding, easier maintenance and
and migrate several hundred programs and reduced costs.
databases without disrupting data-processing
services. With all these methods and solutions emerging
to take care of the legacy systems, maintenance
2 Topics Related in Lecture cost and efforts and systems failures will
gradually decreased. Thus offering long-term
2.1 Commercial Of The Shelf (COTS) benefits.
It discusses how COTS can be constructed
from the existing computing units embedded As a result there will be more business
in the legacy information systems. The opportunities, which in turn, leads to the
existing legacy systems embody a collection increase of the company and system’s values.
of large properties of algorithms and
functions, which could be reusable in new 4 Techniques Discussed in the Article
systems development. Legacy application
systems, particularly some of their functions, This article basically discussed about the basic
are too valuable to be discarded and too problems and several solutions to Legacy
expensive to reproduce (Bennett 1995). Information Systems (LIS). Legacy Information
Systems usually run on obsolete hardware that
2.2 Software Maintenance is difficult and expensive to maintain. Software
Preventing system performance from maintenance is also expensive due to the lack of
degrading to unacceptable level and understanding of the system. LIS usually have
maintaining controls over the modifications poor interfaces that are difficult to use and
made to the system are especially needed in integrate with other systems. Furthermore, they
LIS. This is because the designs to those are difficult to extend.
systems are usually using ad-hoc solutions,
which would make the maintenance and There are three ways to overcome the problems
testing very difficult in the long run. and they are redevelopment, wrapping and
migration. Based on the complexity and risk of information system. With the interoperation of
failure in LIS projects, wrapping is considered as both the systems, the process could be complex
a short-term solution and it can complicate the and slow.
LIS maintenance over a long term.
Redevelopment on the other hand is far too risky As for the Butterfly Methodology, it assumes that
for most organization but this provides a long- although the LIS must remain operable
term plan. As for migration, it is a middle ground throughout migration, the LIS and target system
by providing a long-term solution while mitigating need not interoperate during the process.
the risk of system shutdown.
It is basically like a parallel run where both the
There are two migration methods discussed in LIS and target system are run at the same time
the article. These two approaches are the after the legacy databases are transfer over to
“Chicken Little Strategy” and the “Butterfly the target database. However, any transactions
Methodology”. The first approach, “Chicken during the transfer are logged in temporary
Little Strategy”, lets the LIS and target systems tables. This table will be transfer to the target
interoperate during migration using a mediating database later. During the transfer of the
module. This strategy rebuilt the LIS applications temporary tables, any further manipulations of
gradually on the target system. The target the data are stored in another temporary tables.
system is initially quite small but it grows as the This iteration would go on until the last temporary
migration progresses. This method starts with tables reaches a stage where the migration of
the analysis of the LIS followed by the designing these last tables would not cause any serious
of the target interface, database and application. inconvenience to the core business.
It will then start to migrate the legacy databases
and finally leads to the cut over to the target
5 Related work that has been published
Author(s) Year Article Description How does it relate to
your main article
Keith Bennett 1995 Legacy Systems: Discuss the alternatives and It describe in details
Coping with solutions to legacy systems. how legacy systems
Success Demonstrate the progress made comes about and why
in managing legacy systems. there is a need to come
Legacy dilemma on whether to out with new ways or
trade off the cost of continuing to solution to extend the
cope with the legacy system life of legacy system.
against the investment needed to Main article is
improve it and the benefit of describing 2 of the
easier subsequent maintenance. methodologies that
helps in extending the
life of legacy systems
Harry M. Sneed 1995 Planning the Issues and efforts to consider Extending on the idea of
reengineering of before the start of transforming migration
Legacy Systems the legacy systems. Issues of legacy
Steps on the reengineering systems
process of the Legacy system Before addressing the
Different definition of problems of the legacy
reengineering to different people. systems, a thorough
E.g. Business reengineering and analysis of the legacy
Software reengineering. BR will systems is needed on
only indirectly affect the software the cost, benefits and
but SR involves the renovation of project justifications of
programs, procedures and data the systems. These are
structure etc. parts of the
discuss in both article.
The objectives of both
the articles are to
reliability of the system.
Ettore Merlo, Pierre- 1995 Reengineering The article describes how the Extending on the idea of
Yves Gagne, Jean- User Interfaces process of turning a character- interface reengineering
Francois Girad, based interface into a graphical or migration
Kostaas on can be partially automated. The advantages of
Kontogiannis, Laurie This partially automated reengineer the
Hendren, Prakasb approach separates the interface interfaces (extend the
Panangaden, and and application code, so that new life of legacy system
Renato De Mori interface code for a system can etc.).
be generated without modifying
T.Skarmstad, Md. 1999 Constructing How COTS can be constructed Extending on the idea of
Khaled Khan, Md. Commercial Off- from the existing computing units interface reengineering
Abdur Rashid the-Shelf from embedded in the legacy Discuss a similar
Legacy Systems: information systems. approach of migration
A Conceptual Describe the steps in reverse
Framework engineering or migrating
Describe the steps for Wrapping
6 Relation of Techniques to Lab Project How certain technologies have emerge due to
the legacy systems. E.g. Web technologies
Basically this technique is not related to my project
on implementing a new ticketing system for LRT in Approaches to “freeze” and encapsulate the
NTU. This is because, my project is a non-existing legacy system as a component in a new
system. We have to base on the client implementation. E.g. Controls over the
specifications to develop the system instead of modifications made to the system
analyzing from an existing old system. In this
case, redevelopment, wrapping and migration
does not comes into the picture. Furthermore, my
system does not require a database and thus 8 Comments on Notations/Diagrams
migration of database and applications are not
applicable also. 8.1 Complex diagrams
It is more suitable for readers with certain
How these articles help in Lab Project knowledge or experience in the transformation
of huge Legacy systems. It should provide a
It serves as a reminder to us that the analyzing more general diagram for the approaches
and designing stage is much more important than discuss in the article on top of the in-depth
coding stage. This is because today’s successful diagrams for novice readers.
software systems are still likely to turn into
tomorrow’s legacy system. In order to reduce
maintenance cost and enhance reusability and 8.2 Content Well Organized
easy future add on to the system, it is best to do a It first gave an overview of what is a legacy
good design which is reusable and easy to system followed by the needs to redevelop or
maintain rather than going straight into coding reengineer the system. It then illustrates and
staging with careful analysis of the requirements. discusses the solution to the legacy systems
and also gave the cost and evaluation for each
7 Extension to Article solution. After the introduction, it then
describes in details how the 2 migration
Differentiate the types of reengineering methods works. Lastly, it rounds up by
methods and evaluates on each methods. summarizing the needs for LIS migration.
Determine the implications, cost and benefits 8.3 Explanation Clear and easy to understand
of transforming the legacy systems. The article does not use bombastic words or
jargons. It is easy for reader with novice level
Compare the solutions for short-term benefits of software engineering.
to long-term benefits in terms of effort, time
taken and cost for maintenance.
9 Validation and Expansion to Article had spent millions of dollars fixing the problem.
Therefore, those solutions or methods
The 2 methods discuss in the article are meant for mentioned in the articles are definitely needed
huge legacy systems which are time critical and to solve this problem. These companies may
transactions intensive. We can also use the 2 have either redeveloped a new system or just
methods for smaller legacy systems. In this case patched up the database depending on the
the database would not be so huge and the time both the financial, urgency and system life
taken to transfer it would be much shorter than factors.
those big system.
11.2 Relevance to immediate future
For example, a huge Legacy system like the bank Beside the Y2K problem, there are still a lot
financial systems using mainframe and more legacy problems. Furthermore today’s
unstructured COBOL would required the “Butterfly successful software systems still have the
method” by transfer the database and application tendency to turn into tomorrow’s legacy
to the target system gradually. system. Therefore, there is not only a need to
capture the legacy system data in such a way
For a small legacy system such as accounting that it can support organizations into the future,
packages for e.g. NTU Alumini Club, it is not so there is also a need for more ways to manage
time critical and the number of transactions is a system to prevent it from degrading.
minimal as compare to the bank systems.
Therefore, the butterfly method can be hasten up Besides solving legacy problems, the two
by selecting a time e.g. 3am in the morning when methods discussed in the article can also be
the club is not in operation to transfer the used for example in the merger of banks or
database. In this way, all the necessary data can financial institutions where there is a need to
be transferred to the target database at one go. study the systems in details and redraft on the
There is no need to synchronize the database business policies. In this case, these
further and it also will not affect the operation of organizations will needs to either migration the
the club. Although the database is transfer over, database from one to another or have a new
the legacy system is not cut off immediately. system and database to be able to cater to all
Usually a parallel of at least 2 months is needed to the different business policies.
test the target system. The whole operation will
only be cut over when the target system is stable Thus a further research into this area will
and functioning correctly. In this way, the definitely prove to be of great contributions to
migration is then considered successful. the society.
10 Comments on the S/W Project, Process, 12 References
Platform, People & Product of the Article
1. T.Skarmstad, Md. Khaled Khan, Md. Abdur
This article describes the solutions to the legacy Rashid (1999). Constructing Commercial Off-
systems in a very general way where it may not be the-Shelf from Legacy Systems: A Conceptual
suitable for certain platform, application or Framework [World Wide Web page]. Available:
software. It should give a guideline on which type <
of application, platform, software or database is http://www.nepp.nasa.gov/DocUploads/94D7C
more suitable to which approach and the must and 357-E3A2-46E8-
must not dos for the migration. 9F43EF7FF2DF2034/Constructing%20Comme
Furthermore, as the platform and software gets cy%20systems.pdf >
more sophisticated each day, it will be even more
difficult to migrate a system. It should not only 2. N.Serbedzijia (2000). Engineering Web
discuss the methods on migrating systems and Technologies for Legacy Systems [World Wide
database but also some ways of maintaining and Web page]. Available: <
controlling the systems to prevent it from further http://ausweb.scu.edu.au/aw2k/papers/serbedz
deteriorating as prevent is always better than ija/paper.html >
3. M.Bieber (1998). Web Engineering [World
11 Comments Wide Web page]. Available: < http://www-
Legacy problems have already exists for 4. S.Santhanakrishnan (2001). Re-Engineering
several decades and one the most significant Legacy Applications [World Wide Web page].
problem is the Y2K bug. Many big companies Available: <
5. D. O’ Sullivan, R. Richardson. Applications of
Case Based Reasoning to Legacy System
Migration [World Wide Web page]. Available: <
6. B. Wu et al. (1997) The butterfly Methodology:
A Gateway-Free Approach for Migrating
Legacy Information Systems, IEEE Computer
Soc. Press, Los Alamitos, Calif