Section 9.5. ESA in action: Agile Solutions Ltda

9.5. ESA in action: Agile Solutions Ltda

SAP customers are eagerly anticipating Project Mendocino. One such customer is Agile Solutions Ltda, which is adding this application to its in-process ESA adoption plan. Agile is a pioneering systems integrator (SI) and custom solutions developer in Latin America, the first in the region to create composite xApps with SAP NetWeaver for its customers. Founded in Brazil in 1999, today Agile Solutions has annual revenues of $35 million and 600 employees located in Chile, the U.S., India, and Australia. Its customers include the commodities giant BHP Billiton, Codelco, Samarco Mineradora, Empresas CMPC, Masisa, and Concha y Toro. Despite its deceptively small size, Agile Solutions is one of the world's leading creators of composite applications, and it is currently developing some of the first composite applications based on Project Mendocino.

Agile Solutions's development philosophy in the dawning age of ESA is to focus almost exclusively on the business objectives of its customers, the functionality needed to meet those objectives, and then building the composite application containing that functionality, drawing upon whatever systems and resources are at hand to complete the task.

This has meant that Agile Solutions has tended to downplay discussions of architecture with customers and describe the benefits of composites and ESA in purely business-focused terms. The first customer of its xApps was Agile itself, which implemented the SAP xApp Resource and Portfolio Management in 2002 to integrate its own project workflows. Agile Solutions knew its project management execution was embodied in Microsoft Project and that its data was stored in SAP R/3. What was important was that the two systems be integratednot how, precisely, that would be done. Now Agile Solutions is able to share the status of projects with its customers and partners via an online portal. The portal offers a unified view of the data from the underlying applications, incorporating customers' and partners' feedback and knowledge into the development process.

In practice, Agile Solutions's development strategy demands aggressive timetables and huge skill sets on the part of its developers, who may be called upon to write iViews in Java, services and interfaces in ABAP, and similar tasks in non-SAP applications, all in the service of a single composite application. While Agile Solutions has found it is key to always keep a complete picture of the desired application in mind (just as an architect stays focused on the image of a house), peeling the needed layers of functionality away from their original applications (the metaphorical bricks of the architect's house) calls for developers capable of understanding and transforming each layer. As ESA and the service grid evolve, this process will become simpler, and Agile Solutions's methodologyto focus on the goal and what it takes to achieve it, rather than what is easy to do within the existing technologywill become easier over time.

In the case of Project Mendocino, Agile Solutions is once again using itself as a test bed for new functionalityin this case, a series of knowledge bases that will offer suggestions to and guide future users of Project Mendocino when they make mistakes or confess to the system that they are stuck in a process. To do this, Agile Solutions will create a knowledge base containing insights from its own implementation of Project Mendocino and then test its suggested solution on its own developers during their second, third, fourth, and fifth customer implementations, and so on, incorporating more insights each time.

The resulting composite application will hide the deep functionality from the user and offer context-specific advice instead, essentially automating the search engine concept to deliver knowledge at the precise moment and the necessary level of understanding.

Thanks to the reusability of enterprise services-enabled composites, Agile intends, as a next step, to recycle the application to train new employees in its development process, and to manage business knowledge to create new solutions for customers in fields such as medical diagnostics, where the system could suggest tests or medications in response to user-submitted systems in a predefined process using pattern recognition and search engines.