Business considerations for Enterprise Application Integration

Why should network organisations be interested in EAI?
The introduction of new business drivers is threatening to erode the planned life cycle of legacy¹ corporate systems. It is a recognized fact that these legacy systems do not easily lend themselves to integration with modern systems. However, most organizations know how to achieve each of the integrations required, on a one-at-a-time basis. The maxim heard at the outset of every IT exercise to integrate yet another system has long been ‘This isn’t rocket science’. But given the costs involved, one could be forgiven for thinking that was the case.

The cost of building and maintaining point-to-point solutions has been the major brake on aggressive, large-scale systems integration. E-Business has meant that this can no longer be put off but the costs of integration have not become smaller.

E-Business has also fostered an obsessive drive for innovation. This and the changes associated with innovation are becoming the key drivers for integration. Constant technical innovation is driving how utility and communication organizations compete.

A key question is then how to both reduce the costs of integration and to provide the right environment and IT platform for incorporating innovation. The answer seems to lie in the direction of EAI technology. Meta Group (Waters, 2000) points out that within Communication companies the use of EAI technology is 35 to 40%. On the other hand, Energy IT organizations have been reluctant to embrace EAI due to perceived complexity and lack of EAI understanding and skills.

What are the carrots and what are the sticks?
The ongoing focus on customer oriented business is one of the key drivers behind EAI integration. The customer is the end beneficiary of integrating CRM and ERP applications with each other and applications from partners and suppliers. The drive is towards better results for improved service levels, ones that get noticed and appreciated by the customer. Other business drivers for application integration include corporate mergers and acquisitions, deployment of best of breed packages applications and organizational changes. The need for integration has increased due to e-Business trends such as corporate self service, virtual supply chains, customer relationship management, application hosting and business-to-business commerce.

Building and owning point-to-point solutions is expensive. Anecdotal evidence show that while figures like $1 million to integrate applications from mainstream ERP systems may not be the norm, it is not unusual either. Meta Group (Waters, 2000) has predicted that early adopters of comprehensive ERP solutions (including back office, operational and front office functionality) will re-evaluate the benefits and cost of tight integration and will transition to a distributed, component based architecture that relies on an EAI backbone across the spectrum of energy business segments. The ability to quickly and economically modify business processes and the business applications that support them will increasingly be a key competitive energy company differentiator.

What are the risks and points of failure?
Reducing costs and leveraging e-Business have clear benefits but what are the further pitfalls that Company X and Y need to avoid? And what about Company Z who has chosen the EAI path? The numbers vary from analyst to analyst but there is general agreement that two thirds of new IT application initiatives fail outright or are seriously overdue on delivery time.

One risk that Company Z will immediately encounter is the up front investment. EAI technology means investing in some rather expensive software to accomplish a grand integration scheme. In contrast, with point-to-point solutions, it is sometimes easier to justify single cost – i.e. you get exactly what you pay for. It can appear that with EAI you pay a lot of overhead first. The justification is that the investment is going into foundations and infrastructure for further work.

The Company Y example has some relevance – there is a real danger of EAI projects getting waylaid by the temptation to go and fix or replace legacy systems to make them more amenable for integration. One analyst from Forecross Corp (Estes, 2000) warns against experimenting with business-critical applications. But where necessary, this process can be contained and made manageable by being able to determine the true condition of legacy systems. Then, by initiating a multi-step process, legacy systems can be brought to a globalized state allowing interaction with systems external to the enterprise.

Assuming our Company Z has a ‘normalized’, maintainable, legacy application, it is possible to address the ‘e-Readiness’ of monolithic legacy systems using XML by retrofitting encoding and decoding logic directly into applications. XSLT can provide an external data mapping that allows the organization to implement a true, three-tier data exchange architecture. By utilizing XML and XSLT, Company Z has not only extended the life of its legacy systems but also extended its reach into e-Business.

There is another technical problem that can plague EAI projects. Data model, system differences and semantic differences can be so great that the middleware is little more than a messaging service and one ends up essentially having a point-to-point solution implemented on top of a very expensive messaging service. There is no quick fix to this situation, however, recognizing that it exists early in the project is obviously a cost saving advantage.

Integration of a spatial system poses special considerations and risks, due to:

The complexity of spatial data structures
Integrating a design environment (long transaction) with other systems
Visual or graphic information content of maps and schematics

In addition, the stability of the major EAI vendors is certainly a risk worth considering. There is currently significant consolidation and shift in the EAI market place. EAI market leaders are acquiring new companies and expanding their offerings (e.g. WebMethods acquired Active Software, TibCo acquired InConcert, Neon partnered with Forte). Overall there is no clear single market leader and there are many players – this is an unstable situation that in most markets is a prelude to consolidation.

¹ ‘Legacy’ has many connotations for people. In this paper, it is used to mean, although based on previous generation technology, a working, operational system.