Tein Telecom is the market leader in surveillance infrastructures in the Benelux.
They typically undertake large projects with several dozen or even hundreds of video camera's.
Monitoring is often supported and supplemented by recording and playback facilities as well as image recognition systems that log and draw operator attention to significant events.
SIMBA provides a unified management and operational control platform across these multi-technology, multi-vendor surveillance infrastructures.
It is written in Delphi and is one of Tein's principal competitive advantages in this innovative and rapidly expanding market.
Tein asked me to review the SIMBA architecture.
They wanted to pay specific attention to
- correlating camera representations in the graphical user interface to geographical information,
- ease of integration of new equipment, including isolating the platform from errors in vendor-supplied code,
- openness to third-party security systems,
- scalability. Both at the level of the number of camera's and the number of clients viewing streams and performing command and control actions, SIMBA's architecture needs to scale up by at least one order of magnitude compared to current projects.
Conducting this review with the development team, there seemed to be a consensus on the strenghts and weaknesses of the current architecture and a broad agreement on the evolutionary path to take.
The greatest impediment to the further evolution of the platform was fiercely guarded code ownership and a lack of communication within the development team.
Hence I suggested experimenting with pair programming.
I provided coaching at daily standup meetings together with Vera Peeters of Tryx.
We also introduced iteration retrospectives.
By the end of the engagement, the team felt the development process had been greatly improved.
They were well on their way to making the required architectural adjustments.
Business units also felt that development performance had significantly improved.
Reference: Frédéric Loncour