|
Business Intelligence (BI) and data warehousing have undergone a significant evolution in the past 10 years. There was a time when BI was a standalone, isolated world of time-dated, historical snapshots of data for analytics. BI implementers extracted all the data they wanted from the operational environment, and then disappeared with it into a world of their own.
Tactical and strategic analyses were performed on daily, weekly, and perhaps monthly snapshots of data. Query performance was important but not a critical factor in the environment’s success. The audience was limited to business analysts and highly technical statisticians and researchers. The most difficult problem faced by the implementers was the ability to give these users flexibility in terms of handling unplanned or unusual queries.
Then along came the notion of operational BI. While it seemed innocent enough – just another form of BI – operational BI has turned out to be a major disruptive force in BI environments today. Perhaps it can best be summed up in one characteristic: Operational BI must actively support operational decision making.
What does this mean to your existing data warehouse environment? Plenty! Operational BI has put significant and increasing pressures on the BI environment to meet these new business requirements. This white paper discusses the pressures and their impact on data warehouse environments. We end the paper will suggestions on how to make your infrastructure more “dynamic” and future-proof.
Operational BI Pressures on Today’s BI Environments
There are five major categories of pressures that we will discuss.
Data Currency and Volumes
The first major pressure from operational BI users comes in the form of the currency of the data they use. To make good operational decisions, the data must be much fresher or current than that normally found in traditional data warehouses. Detailed intraday snapshots of data are trickle fed into data warehouses or operational data stores, allowing operational personnel the ability to analyze events occurring during the day of the event. As you can imagine, this means that the volume of data being stored increases substantially. Data warehouses now contain tens of terabytes to hundreds of terabytes (even petabytes!) to support all forms of BI now. And it should be noted that the granularity of the data for operational BI must be at the lowest level of detail. All this means that, not only does the data warehouse infrastructure have to handle faster, more frequent loads of data, but seamless scalability is mandatory – whether it is for processing the data, storing the increased volumes, or maintaining the integrity of the environment (backups, failovers, etc.)
Performance
While performance in traditional BI environments has always been important, it is now a critical success factor for operational form of BI. The mere name, operational BI, should conjure up visions of users requiring sub-second response times – enough to make the stoutest BI implementers’ hearts quiver!
Now add to this the fact that the environment must still support the more traditional BI users (tactical and strategic) with appropriate response times. This one facet of the evolving data warehouse environment – handling a mixed workload environment – has stumped many database vendors. A mixed workload environment means the technology must have the ability to prioritize queries, not only according to their importance to the enterprise but also their response requirements.
Number of users
Before the advent of operational BI, most decision support environments appealed to a relatively small number of users – perhaps less than 10% of the entire workforce. With a focus on operational decisions, the new form of BI attracts a great deal more users, maybe even all of the business users.
These new users may have very different interface requirements. If they are front line personnel, they are familiar with drop down menus, rigid edit checking, in-line help functions, and so on. These are unlike anything traditional BI implementers have had to deal with before. It means that BI implementers must rethink how BI is delivered, what the interface looks like, how support should work, and where and when the users will use this capability.
|
