At Perficient, we often talk to customers about the technical “requirements” to deliver Performance Management solutions. Every situation is different – based on chosen platform, user counts and profiles, concurrency, hardware, OS, network, and a myriad of other items depending how deep you want to go.
Especially for Business Intelligence, what we always find in successful deployments is another set of software. Software which can be thought of as highly desired, but not absolutely required, for BI programs. The people on the BI teams challenged to succeed will need more than just the chosen BI platform to be effective.
The list below is not provided to make specific software recommendations, but given Perficient’s history working in enterprise accounts and deployments, we inevitably see some more than others and are noted along the way.
- DBMS Client Software – The ability to query the database through a visual tool. SQL Server Management Studio, TOAD, DB Visualizer are common tools we’ve seen. The decision on which to use is largely driven by the database landscape. The most effective BI Designers/Developers have the ability and know-how to run explain plans and tune queries.
- Data Modeling Software – The ability to visually design, update and generate database scripts. Typically located on client machines. Common tools include Erwin and ER Studio. We see higher level versions of Visio (Architect, Enterprise, etc.) used a fair amount and increasingly as its capabilities grow to include those of the other modeling tools on the market.
- Server Monitoring – Since most BI environments currently run on Windows OS, especially in the midmarket, we will stick to that for now. Enhanced monitoring for Windows environments over and above what is provided by Task Manager is provided by Process Explorer. It is not wildly scientific, but gives a lot more information than Task Manager does and its free. Perfmon and BMC are used in a number of large accounts as well. No doubt large IT shops will have a pre-determined standard.
- Performance Testing – A systemic, automated way to drive user load and test the architecture’s scalability. Tools such as JMeter, HP Load Runner and OpenSTA are common.
- Workload Automation – A job scheduling tool functioning as the process controller for data related processes. Often data related processes kicked-off on the DBMS, and upon successful completion issues a command to execute a BI process on an entirely different server. Common BI items you see are running (re)loads for an OLAP Cube or running reports in batch. Control-M and Autosys are common.
If you have other ideas or comments, we’d like to hear about it. Leave a comment below!