Hybrid Business Intelligence

In the current scenario many small and medium business have their BI analytics on cloud hosted services, while still the large Organizations maintaining data in on-premise data stores for their BI requirements. There are various reasons and financial benefits for Organizations to have their BI on cloud, the majority being:

  • 24×7 availability of environment without any infrastructure dependency
  • Time taken to deploy BI components drastically reduces
  • Flexibility of scaling up or scaling down the infrastructure based on requirement
  • Cost effective in terms of infrastructure procurement and huge spending on license
  • Supports mobile and Web based remote access to control database access, user security and perform analytics

Aside from the above, there are few concerns for most of the Organizations to have their BI on Cloud as stated below:

  • Data security is a major concern for many organizations to move to Cloud as it implicates their confidential/sensitive data to be placed on Cloud
  • There are many cloud BI vendors on the market which confuses the Organizations to choose the appropriate vendor based on the Organization needs and it is difficult to measure cloud BI vendor capability
  • Performance limitations of data warehouse on the Cloud
  • There is no standard prizing model which makes it difficult for the customers to choose the right cloud BI.

Considering the above pros and cons, many Organizations currently migrate to hybrid BI scenario where you have a combination of cloud and on-premises approach in place.

Hybrid Business Intelligence (BI) has an advantage of a centralized approach with that of a decentralized or distributed one.  Many organizations have comprehended there are various concerns that must be addressed when incorporating these two forms of BI.  Consequently, implementing appropriate Hybrid BI harvests more information than traditional or cloud BI.

Hybrid Approach to Business Intelligence provides more flexibility for Organizations analytical and reporting needs.  It is recommended to have a centralized architecture of distributed data, which enables the business to work with any BI tool.

In a Hybrid BI environment business can practice Pareto’s 80/20 principle in which 20 percent of data extended from distributed BI in respective departments is channeled into a central repository. This 20% of data that is mission-critical and highly relevant to a particular business process can be maintained in a centralized repository and the remaining 80 percent of the data can remain with the departments that procured it in terms of relevant business data marts. It is up to the Organizations to determine which data should be maintained in a centralized repository and the sensitive/business specific data in the decentralized environment.

The harness of hybrid approach includes:

  • Quick development with an enterprise perspective
  • Converging the warehouse on data that provides value and not data that happens to be available
  • Implementing consistent meta data by maintaining a centralized repository
  • Eluding redundant extracts and inconsistent transformations through the implementation of a physical data warehouse
  • Utilizing the business resource in a very concrete way to define the data quality and transformation rules

With hybrid approaches, the physical architecture can be discussed upfront. If the ultimate size will demand a separate physical staging area or a data mart layer, they could be implemented early or the program intermittently monitored for the addition of these layers.

In a hybrid approach, data sourcing or subject areas undergo much iteration with the decision based on providing large business value early and quite often.  Through conformed and changing dimensions approaches, hybrid architectures ease a high level of sharing and provides single version of information for most business subjects by leveraging a physical data warehouse layer that provisions enterprise requests but also directs the shared data to the data marts.

Attention is also given to metadata, performance planning and user training in hybrid approaches.

Conclusion:

Today’s most effective business intelligence (BI) solutions are not built by traditional approaches, but by functional, hybrid, modern approaches that helps to craft a design that  meets the longer-term as well as the short-term needs of an Organization.




image

Study on Re-engineering the Business Intelligence

We Re-engineering Business Intelligence tool and gave new usability standard with better look and feel.