« Innumeracy and Business Intelligence | Main | Non-Linearity - Why should BI Practitioners know about it? »
April 27, 2010
ON-DEMAND WITH ON-PREMISE - IT'S GAME ON!
I have been encountering 'this' situation in BI consulting engagements with increasing regularity and thought that it merits a blog post. 'This' situation mentioned above is a scenario where enterprises are utilizing a combination of On-Demand cloud applications (say for CRM functionality) with On-Premise systems (Traditional BIG ERP's like SAP, Oracle Apps etc.) with exchange of data between these systems on a regular basis. What's more! - It's the ubiquitous ETL tools like Informatica, SQL Server Integration Services, etc. that are used to exchange data between the systems.
Here are some instances of such a scenario, that I have seen recently:
1) Large Healthcare information processing company uses SalesForce.com as the cloud based CRM system and needs to interact with On-premise SAP system, using Informatica as the ETL engine.
2) Leading Airline Company uses Oracle CRM On-demand and wants to exchange data with Oracle Apps which is installed in the company's data center. The enterprise standard ETL tool is SQL Server Integration Services and that needs to be leveraged for data integration.
3) A large Telecom company, which uses Jobstreet as an on-demand recruitment engine wants to integrate the recruitment data with Oracle HR Analytics (BI Apps) module installed on-premise, to develop a comprehensive BI platform for HR data.
I am very sure that there are many other scenarios where the integration between on-demand and on-premise software is required, and this trend is bound to accelerate in the future. The good news for BI practitioners is the fact that such integration (for any scenario) can be accomplished with a good understanding of web-services in the context of ETL platforms.
At a high level, the data integration architecture for the On-demand plus On-premise scenario, is as described below:
1) On-demand applications provide webservices for insert, update & delete for each entity
2) Each WSDL file has a set of methods that needs to be understood for its functionality
3) ETL tools have the capability to call webservices within its flow
4) ETL tools increasingly are providing specific integration packs with On-demand solutions that provide for an easier & more comprehensive way of integration. For example, Informatica provides a connector for SalesForce.com that just makes the integration all that more easier.
BI practitioners would do well to understand the business imperatives behind the on-demand with on-premise scenario and think through a solid ETL technology architecture to enable it.
Before signing off, I would like to introduce the EbizQ forum where I have been a forum contributor for the past few months. Given below is a forum question and my reply, that is relevant in the context of this blog post.
Thanks for reading. Please do share your thoughts.
Question:
Is There a Certain Size Business or Certain Vertical Industry for Which SaaS BI Makes Most Sense?
By ebizQ on Mar 1, 2010 at 10:05 AM
- Karthikeyan Sankaran | March 2, 2010 12:16 AM | Reply
I think SaaS BI has a role to play in organizations across industries and of various sizes. Every organization has a set of business functions (Finance, Marketing, Operations, Sales, Strategy etc.) each with its own characteristics that dictate the business decisioning requirements. The applicability of SaaS BI is more a function of these characteristics than the organization themselves. That the characteristics themselves might be dependent on the industry or size is definitely a valid argument but I will leave that for another discussion.
For example, certain business decisions require a tightly coupled information chain (from ERP to DW to Reporting) and this is best served by in-house / on-premise BI platforms. On the other side of the spectrum, certain decisions are to be taken based on a loosely coupled information chain and these decision makers would be well served by on-demand / SaaS platform.
Bottom-line, I expect, every organization, big or small, to have a mix of on-premise and on-demand BI platforms each serving a specific business community and opportunity.
Posted by Karthikeyan Sankaran at April 27, 2010 12:00 AM
Twitter 
LinkedIn 
YouTube