Early this year, the mid-market and vertical-centric enterprise applications vendor promised to release the details of its SOA strategy at its user conferences later in the year. It has kept its promise with details and initial delivery dates being presented to audiences at its string of Inforum user conferences. Delivering on the commitment was significant because it is a proof point that there is more to Infor than M&A expertise and that it has the will and resources to invest in R&D.
Its Open SOA program is the opposite of the approaches adopted by SAP and Oracle, and different again to Lawson. Rather than delivering a core platform that users need to license and adopt to move to SOA, it has opted to bake SOA into each application. It is providing an integration backbone, courtesy of Progress’s Sonic ESB, that applications and services can plug into to enable message-based, event-driven, publish-and-subscribe SOA.
The technology is being added to existing applications and those applications are being service-enabled through add-on features. Both aspects of the SOA technology will be provided as part of the normal maintenance fee, which will cover integration between Infor applications. Customers who to want to bring third-party applications to the equation will have to pay a license fee.
The Open SOA strategy is based on the event-based model outlined by OAG, and uses the OAGIS 9 business object document specification. It is based on the concept of event-based architecture that embeds much of the capability within the applications and also means implementations can be SOA-enabled without having to upgrade as a whole or rewrite the code. In theory, components can be swapped on a piece-by-piece basis, allowing a gradual transition and co-existence of old, new, and third-party applications.
CTO Bruce Gordon refers to it as invisible SOA. Customers want the benefits but they don’t want to know about it, he said. With a customer base of mid-sized business with potentially limited resources, Infor’s challenge was to reduce the complexity and minimize the cost of SOA. Although SOA is supposed to simplify integration, it remains a complex proposition. Gordon admitted that it is impossible to remove all the complexity, but said he believes the Infor approach will meet the requirements for flexibility and agility without delivering too much complexity.
It is more important to deliver solutions that can be implemented through feature packs that do not disrupt the business. [Organizations can] focus on what they are doing, not an IT department that changes all the time, he said. Open SOA is customer-focused. We are doing what our customers indicated they wanted. We are evolving existing [applications] not creating a platform lock-in. There is no need to move to a new platform. We enable co-existence. There is no charge for SOA infrastructure. There is no major reimplementation. It is event-driven SOA, It allows you to upgrade by components, yet still get more efficiency because you are using events.
The company has spent the last few years doing behind-the-scenes work and is now at the stage where it can start delivering. The EBS is now available, and each application in its cavernous portfolio will receive the SOA treatment over the next few years. SOA components will be released to the market during 2008, with the first rolling out in June and a rush slated for December.
The company has demonstrated the first components that will make use of Open SOA. Role-based home pages are scheduled for June 2008 for the ERP LN suite. It is a new style of user interface using Ajax-based web pages that will be able to present data from multiple disparate systems in a more intuitive way. Pages will be able to be personalized and there will be support for enterprise mash-ups for to improve data visualization.
Multi-books accounting facility is another SOA component development. This a new general ledger facility that tackles the problem of global companies’ requirements to conform to multiple accounting standards and regulations. It can be run concurrently with existing general ledger systems from both Infor and third parties, or as the sole system, and provides the facility to support multiple books without having to swap out applications or upgrade. Its is scheduled for delivery in December 2008. It is an example of how Infor plans to develop SOA components that can plug into existing systems, add value, and provide an evolutionary path to SOA.
Our View
Adapting existing technology always means compromises or inefficiencies, and it will be no different for Infor’s Open SOA. At the moment Infor appears to be bearing most of the cost through its commitment to extend its existing applications and the heavy workload that entails. The most obvious downside is that customers may have to wait a few years before they can access the SOA world.
Infor’s SOA strategy and technology is highly pragmatic for itself and its customers. With its variety of different platforms and code bases, it would be impracticable for Infor to try and create a common SOA platform and move customers over to it. However, Open SOA allows it to continue to develop each of the application lines independently while also guaranteeing that they will work together no matter what technology base they use. Customers will benefit from the ability to continue with their chosen applications, whether they want to take the SOA route or not. When and if they do, they should not be faced with the risk of a new implementation.
Infor’s acquisition strategy determined its SOA design, whereby Open SOA is effectively the technical manifestation of its business strategy. In that sense Infor may not have had free choice when it came to choosing its SOA approach, but that does not preclude it from being successful.
