The Cloud as a Data Source
Not a week goes by anymore where we are not talking to a customer about needs based on a Cloud paradigm. The term Cloud, like Dashboard, has found a rather ambiguous definition and can mean different things to different people. If you dig deeper into IBM's definition of cloud for example, you will find additional terms like Managed Private Cloud and Hosted Private Cloud and Shared Cloud Services.
We have found the placement of Performance Applications, particularly BI applications and Data Assets, to be slow to move out beyond an internally managed network and set of resources. That certainly doesn't mean the Cloud does not apply to Performance Management.
From a data architecture perspective, we deal with the Cloud every day. We have successfully integrated Cloud-based data sources such as Salesforce.com and NetSuite's OpenAir with companies' other internal data assets. In these cases, we are bringing Cloud-based data back inside the organization and combining it with other data assets to give Business Leaders a more complete view of their businesses, regardless where the source data may resides. Connecting and integrating cloud-based data often presents additional challenges, not dissimilar, from existing challenges:
- Master Data Management and Data Quality - Unless a lot of work went into the implementation of the cloud-based application to ensure it aligned with enterprise data assets, bringing data together after the fact will be much harder and require rigor to ensure accuracy and confidence in the user community. At worst, bad data will drive bad decisions.
- Connecting ETL and other data integration mechanisms to Web Services is not the traditional ODBC/Native connection the traditional RDBMS Database Administrator or Database developer is used to. If a direct connection to a web service is not made, then there will be middleware.
- Timing and performance will continue to be must-haves but when the source of the data is outside of the internal company network, sometimes certain levels of control are just not possible.
The Cloud is certainly not going away and follows other outsourcing trends of the last half century.
