Legacy System Migration: Strategies, Risks, and Benefits

Introduction

Legacy systems continue to be widely used by many corporations and government entities for a variety of reasons, with complacency and inertia at the top of the list. Many outdated computer systems, databases, and other elements of information architecture are still fully functional, and thus — despite the creation of more efficient, more effective, and more relevant replacement systems and technologies — many information officers and other decision-makers have opted to retain their current systems rather than migrate to newer ones. The complacency that this decision indicates is hugely damaging to overall efficiency for many such entities, and the problem of migration from legacy systems to newer platforms is an issue that is beginning to be addressed in academic as well as practical literature.

There are also very real and pragmatic issues that sometimes act as barriers to migration and support the continued reliance on legacy systems, and these are compounded by the complacency noted above. The costs of adopting new technologies and information systems and migrating data can be quite substantial, and lost productivity during periods of transition — which often include long periods of system downtime as well as training and adjustment periods for personnel — can also be quite inhibitive. Developing strategies for data migration that minimize these risks while also maximizing the potential benefits of migration is key to convincing relevant personnel to abandon legacy systems.

This paper presents an analysis of current trends in data migration and findings in the literature regarding adaptations and replacements for legacy systems. Literature is reviewed that examines different perspectives toward change in information architectures and computer systems, and different specific pilot studies in the field are also summarized. This provides evidence as to current strategies employed in data migration and the transition away from legacy systems, and the degree of success that these various strategies and interventions have been able to achieve. From these examinations and descriptions, as well as other situations where migration and legacy system replacements have been observed, recommendations as to the most effective migration strategies and the most beneficial changes to information architecture can be made.

Current Migration Strategies: Examples and Evidence

A variety of different migration and legacy system replacement strategies have been employed by different organizations, with highly varied results. Some studies have demonstrated major positive changes to organizations following the transition from a legacy system, while other changes to information architecture have apparently had less of an impact. There is also evidence that changes to technology have occurred at a highly uneven rate even among entities that are highly similar in nature, suggesting that personal and subjective decision-making at each organization is the primary reason for changes to — or stagnation in — computer systems and information architecture.

The need for migration away from legacy systems — as well as many of the practical and bureaucratic barriers that can exist to this migration — is quite clearly demonstrated in a case involving New York State that began to evolve a little over a decade ago. In Part I of a two-part case, Fisher and Bradford (2005) detail the difficulties brought about by the inadequacies of the financial management systems used by individual departments and in their interfaces with the state's Comptroller. Though the system was still functional, it was not fully compatible with other technologies in use in individual offices, nor had it been adapted and upgraded to continue serving changing needs (Fisher & Bradford, 2005). Though these major and frequent problems with the legacy system were noted by state officials, the costs, complexities, and interdepartmental politics that became issues forestalled any significant action for a considerable period (Fisher & Bradford, 2005).

A survey conducted in 2008 found that many states still estimated that at least forty percent of the computer and information systems in place in their various departments would be classified as legacy systems in need of migration (NASCIO, 2008). Enterprise Resource Planning (ERP) migration strategies were viewed by many agencies responding to this study as largely ineffective and at best temporary solutions that could be used to guide migration, but inadequate in and of themselves (NASCIO, 2008). This finding supports the view of ERPs found in the earlier New York case study, where only one department attempted to make use of an ERP and was ultimately unable to make this feasible (Fisher & Bradford, 2005).

Enterprise Resource Planning Systems and Their Limitations

An extensive study focusing solely on Enterprise Resource Planning systems found that, while there did tend to be a large rate of failure both in the design and in the implementation of these systems, there were definite successes as well (Singla & Goyal, 2006). These researchers found that in most cases of failure they examined, an accurate assessment of risks based on the complexities and interdependencies of the various systems currently in place was not taken into account when designing specific ERPs, and that this lack of knowledgeable and comprehensive assessment carried over into implementation (Singla & Goyal, 2006). This, in turn, caused failures in system performance as well as failures to meet capability expectations, leading to a highly negative experience of and assessment with ERPs by the majority of organizations in the study that attempted to employ them as migration tools (Singla & Goyal, 2006).

Even simple changes in information architecture can prove to have large ramifications and hindrances, as well as questionable efficacy. A small pilot study was conducted in one organization with the introduction of extra graphics and icons warning of various risks to information exposure that were encountered as users navigated their standard information system (Pattinson & Anderson, 2007). The results of this study were ultimately inconclusive. Differences in reaction and security-taking measures between groups exposed to the extra graphics and a control group experiencing the security interface in a standard way were large enough to convince the authors that something had occurred, but without creating a statistically significant difference in the small-population study (Pattinson & Anderson, 2007). Evidence such as this is one of the major reasons that so many information officers are content to let their "good enough" legacy systems remain in place — the evidence for significantly increased efficacy simply is not there in many technological changes and advancements.

Conclusion

The case for migrating information away from legacy systems and outdated information architectures is quite strong, but it is not supported in current literature in a way that appears convincing to a large number of information system decision-makers. Changing this requires a direct approach that demonstrates the tangible improvements in information systems designed for current needs and technologies. This will also tend to present long-term cost savings and efficiency improvements to organizations that execute effective migrations.

References

Fisher, I. & Bradford, M. (2005). New York State agencies: A case study for analyzing the process of legacy system migration: Part I. Journal of Information Systems, 19(2), 173–89.

NASCIO. (2008). Digital states at risk! Modernizing legacy systems. National Association of State Chief Information Officers.

Pattinson, M. & Anderson, G. (2007). How well are information risks being communicated to your computer end-users? Information Management & Computer Security, 15(5), 362.

Singla, A. & Goyal, D. (2006). Managing risk factors in ERP implementation and design: An empirical investigation of the Indian industry. Journal of Advances in Management Research, 3(1), 59–67.

Van der Walt, P. & Du Toit, A. (2007). Developing a scaleable information architecture for an enterprise-wide consolidated information management platform. Aslib Proceedings, 59(1), 80.