Integration is a key attribute of successful user adoption of cloud solutions. As enterprise cloud adoption increases, the need for integrating to multiple backend data sources is inevitable. In the case of Salesforce integration - the majority of enterprises require data to be in sync (one-way or bi-directional) between on-premise and other cloud systems.
Treating cloud systems like another on-premise solution and approaching cloud integrations using traditional enterprise application integration (EAI) paradigms is not the path to success. The new generation of integrations should take advantage of what cloud solutions offer and not weigh down the agility of cloud solutions with too much dependence on on-premise data.
4 things to consider for successful cloud integration projects:
1. The Limitations are different:The number of APIs called, governor limits, amount of resources consumed should be considered before architecting solutions. Most of the cloud platforms provide standard APIs and multiple transport layers which makes integration easier. However each platform has their enforcements and limits which need to be well-understood before architecting any solutions, using the right API for the specific integration needs. Not all integrations have to be SOAP webservices based and real-time.
2. Pace of innovation in Cloud is staggering:Workday and SFDC deliver quarterly releases with 150+page release notes. Google delivers constant innovation in even more regular releases of functionality for Google Apps. From an integration standpoint, this means APIs and limits are changing fast and keeping up with constant change is key to being successful, embracing agile to keep up with the changes and adopting new features as they come. Keep an eye on backward compatibility and the ability to support multiple API versions.
3. Integration solutions and tools are different:There’s a new breed of integration solutions that have come up to help with cloud to ground integrations, presenting many options to choose from:
- Mature enterprise-grade ESB platforms - e.g webMethods, Tibco, IBM Websphere and MuleESB.
- Lightweight cloud to ground integration tools with different form factors - appliance, VM and cloud. e.g., Cast Iron, Pervasive, Jitterbit.
- Pure cloud integration solutions offered as a service, e.g., Boomi, Snaplogic, Skyvva and Informatica Cloud.
- ETL and simple batch based data export - Salesforce Data loaders, Informatica ETL, Talend, DBAmp.
- Use of vendor-specific integration solutions - e.g Workday integration network, Workday API, Workday Studio, use of right SFDC API - SOAP/REST, Bulk, Streaming API.
4. End game is process integration:Perhaps most important, the integration architecture needs to support integration across multiple cloud and ground applications. The business process should define the integration type and ensure its keeps the right level of interaction without too much dependency between the systems. The integration should not be myopic at the object level but rather defined at the process level then drilling down to the objects.
Sample Integration Roadmap
Ultimately, as much focus as is given to selecting the middleware, data mapping and use of APIs, equal importance should be given to the integration process, making use of the advantages you have in cloud adoptions. The objective is to make integration transparent so that an end-user has a simple experience that seamlessly brings together the right data from across multiple back-end applications. This has always been the goal of successful integration projects but it’s more achievable than ever in the cloud!