Think Cloud Integration is Pie in the Sky? Say Hello To – Hybrid Integration!

“Cloud is the Future” -  has become a cliché in the technology arena. Implementation of cloud technologies has been touted as a means to level the playing field for all. As of now, it has been successful in commodity type applications that are mostly mass market and directly facing the consumer. Most of them have evolved, successfully causing disruption to the traditional software models.

However, Enterprise Applications, ERPs’ and Integrations (EAI, EDI & ETL) with on-premise applications and it's use cases for being able to leverage the cloud vary greatly across the spectrum. While the industry consolidates and evolves specifically around Integration using cloud platforms, here are four things to consider before moving your data integrations to any type of cloud platform.

For the purposes of this article, when we reference “The Cloud” or “Cloud Platforms”, we mean Software as a Service (SaaS) for the data integration space.

Top FOUR things to consider before moving to SaaS / Cloud for data integration

Your business has unique needs.
Varied data formats such as EDI, XML, IDOCS, AgGateway, cXML, CIDX, PIDX. RosettaNet etc. along with company specific ERP systems make integrations unique. Additionally, the respective business logic on how companies operate is either a competitive advantage or market differentiator. A Cloud Platform integration that is generic enough to support all businesses including potential competitors may take away market differentiators. 

The cloud becomes a black box.
If integrations are completely moved to a SaaS provider platform on the Cloud, the companies are setting themselves up for a dependency on the SaaS provider – where all business logic, data mappings and BPM rules are externalized and out of control. The change in mass market platforms are not performed on individual customer basis.

Ongoing costs add up quickly.
Generally, EDI cloud solutions are priced as per transaction or by message volume/size. As each message received/sent - in/out of the ERP system to/from external Trading Partners, it spins off several new messages.  Each of these messages has to make those round trips to and from the SaaS black box. This means a multiplier cost structure for each transaction.

Take for example an incoming EDI Purchase Order from a customer going into the ERP system for fulfillment, will generate Order Conformation, Order Change, Shipment Notice, Invoice transactions and all the Acknowledgements for all the mentioned documents in both directions.

Support and maintenance needs.
Support and maintenance will each cause more headaches and business continuity will always be at risk. Besides most SaaS platforms for integrations will come with a learning curve for existing technical staff plus the cost of vendors trying to help out. Due to the sensitive data coming from ERP systems there is a huge security risk when using Internet protocols with Cloud platforms for batch or real-time processing of data from trading partners.

Hybrid Data Integration

When dealing with data integration, the ideal choice we recommend is separation of integration functionality into two keys services. Communication Services and Business Process Management [BPM]. While Communication Services can live on the cloud or “IaaS”, BPM Services should reside inside the firewall and on premise data center.

By separating out these two functions and applying different technology solutions, the drawbacks listed above for pure cloud platforms can be avoided.

This is the hybrid solution provides the best of both worlds. Companies can now leverage the current infrastructure investments while preparing themselves for the cloud integrations in the future. Besides integrating with ERP systems such as SAP, PeopleSoft, JD Edwards, Lawson, MS Dynamics, Oracle, it will be much simpler if the integration platform that handles the data transformation and BPM functions is on premise, within the firewall. This is helpful to avoid data latency and ERP systems are not expected to make round trips to the cloud platforms to perform basic integration functions.