RESEARCH ARTICLE
*Email Address: ranaehsan@gmail.com
1710 Adv. Sci. Lett. Vol. 24, No. 3, 2018 doi: 10.1166/asl.2018.11142
Copyright © 2018 American Scientific Publishers Advanced Science Letters
All rights reserved Vol. 24, 1708-1711, 2018
Printed in the United States of America
A Review of Cloud Migration Techniques and
Models for Legacy Applications: Key
Considerations and Potential Concerns
Muhammad Ehsan Rana*, Wan Nurhayati Wan Ab Rahman
Faculty of Computer Science and IT,
Universiti Putra Malaysia,
43400 Serdang, Selangor, Malaysia
Legacy systems refer to the applications designed for a particular client that have been in use for a long period of time and
developed using obsolete technologies. They are often business-critical systems, therefore any changes here inevitably will
affect the other parts of the system. Although these systems are considered to be outdated but are too costly and risky for an
organization to replace it. Cloud computing provides a new platform for organization that promises flexible scalability,
business agility, high availability and reduction in costs. Considering these benefits, migration of legacy systems to cloud is a
lucrative option for many organizations. However the architecture of these legacy applications require a tested, fool-proof
and risk free approach for migration. Limited migration models or frameworks have been proposed which caters issues of
legacy systems migration to cloud platform. Many of these cloud adoption techniques and models emphasize on the generic
phases and procedures on migrating the applications and data to cloud. However some of these models are more flexible, and
provide better approaches compared to others. This research explores the issues associated with the legacy applications with
regards to their migration on cloud and reviews the existing techniques and models that have been proposed in this context.
Keywords: Legacy Systems, Cloud Computing, Cloud Migration Techniques, Cloud Migration Models.
1. INTRODUCTION
A. Legacy Systems
Legacy system refers to a computer-based system
that may still in use but no longer receiving vendor
support (cannot be upgraded). Not only hardware and
software are included in the system, legacy processes and
procedures are also part of the system since the ways of
doing things are hard to be changed as they rely on the
legacy software. As a result, legacy system is usually
incompatible with newly purchased systems and may be
problematic due to compatibility issues, obsoleting and
lack of security support.
These are often business-critical systems and
changes on the system inevitably will affect the other
parts of the system. This requires the system to be
maintained although no updates are available from the
vendors, but is too risky to be replaced from the business
perspective. For example, bank customer accounting
system mostly are one of the earliest system which
organizational policies and procedures will rely on. Bank
might be having high business risk, if the system replaced
did not work as it should be, when the bank replaces the
existing system which may run on expensive mainframe
hardware. Addition, the changes on existing procedures
may cause difficulties with the bank’s auditors
1-3
.
COBOL based systems are good examples of legacy
systems. COBOL was introduced in late 1950s but there
are estimated 200 billion lines of COBOL codes still
running in business and enterprise systems
4
. These