The work of information technology (IT) managers for food, chemical and life science product manufacturers seems to take on an added urgency given all the data gathering and reporting required for regulatory compliance. To handle the need, manual systems given way to digitized systems that then often reach into other systems—sometimes peacefully coexisting and sometimes making their users’ lives miserable. The goal of a single enterprise-level system that handles everything seamlessly is often only a mirage—despite the impressive intentions of enterprise resource planning (ERP) systems
Most manufacturers still only use ERP systems to manage accounting transactions and financial goals. Product lifecycle management (PLM) systems emerged to tackle the single-system problem from the bottom up: They start with the product being made and manage it through research and development, prototyping, manufacturing and phase-out. Of course, lots of other software has sprung up in between, handling the specifics of operations management, quality management, laboratory information management and more. Unfortunately, much of the data managed by these systems overlap, and the systems themselves don’t share information easily.
When Roka Bioscience, a Warren, N.J. manufacturer of food safety diagnostic systems, needed to integrate the various systems in its supply chain, it found a metadata-based integration platform from Magic Software Enterprises to pull together its Oracle ERP, PLM and third-party systems data.
“Streamlining administrative processes and eliminating duplicate data entry is key for improved productivity and data accuracy,” says Daniel Churco, Roka Bioscience senior manager for information technology. “However, when we tried to connect our Oracle Agile PLM system with our Oracle JD Edwards EnterpriseOne ERP system, we found out just how difficult an ERP-to-PLM integration can be.”
Roka Bioscience uses the Agile PLM system to guide its product development through final product manufacturing. It’s also where the company stores all of its product lifecycle information: documents, design history files, manufacturing records, parts, bills of materials (BOMs), etc. They also rely on Agile PLM for all quality management system (QMS) information.
The JD Edwards ERP system uses a lot of the same information as Agile, such as item and catalog numbers, lot information, serial numbers, BOMs and supplier data, among other details. “Instead of doing manual double-entry, which takes time and can introduce costly transcription errors, we decided to integrate and automate the two systems so that once information is approved in Agile, it can flow into JD Edwards,” says Churco.
Using integration packs over the course of a very frustrating year, the integration failed. So Churco had to find something else. “Via a simple Google search for the word ‘integration,’ I discovered the Magic xpi Integration Platform,” he says.
Magic xpi is an enterprise-grade integration platform with native certified adapters for Oracle JD Edwards products and other leading IT systems. It supports Windows, Linux, AIX, IBM i, and Solaris. More than 50 technology adapters allow it to integrate a wide variety of operating systems, databases and standards, including Webservices, messaging, FTP and directories, LDAP, .NET, Java, RPG, and HTTP, among others.
“Magic xpi is system-agnostic and has achieved Oracle Validated Integration for JD Edwards,” says Churco. “This—together with its code-free, metadata-based approach and visual drag-and-drop studio—allows us to work with the multiple systems easier than using native Oracle tools or custom integration. The GUI lets you look at your flows and understand them visually, which helps build the project and makes trouble shooting a lot easier than looking at lines of code.”
If you include the previous failed projects, Churco says the entire integration project took over two years, but “using Magic we were able to integrate Agile into JD Edwards in only nine months.”
The next phase of the project is to automate new business processes. For example, tech support data is currently entered into JD Edwards, and the same data is also manually entered into Agile so it can be formally assessed and managed by the Quality Systems department. “Automating these processes with Magic xpi will save time and reduce the possibility of transcription errors, which can present a compliance risk,” says Churco. “We expect this project will go even more quickly because the base is already in place.”
