Home : Introduction requirements issues capabilities ECM proposal project life metrics CMMI benefits motivation author academic lessons learned policy
Demonstrated: distributed database conveyor emulation

Distributed Database

An aircraft required an embedded distributed database which is shared among 5 systems all but one had multiple processors within it. Extra effort placed up front in infrastructure paid off considerably. A single electronically readable description file was designed which fully describes the data base, including what each system contained and the connections between the systems. A custom tool parsed and verified the file before it was allowed to be inserted into the system. Verification included band width and memory utilization The payoff includes:

Any one on the project can modify the data distribution with complete confidence, if it gets past the tool, it will work
Generated distribution source code including the structures needed by the applications
Accurate documentation and reference material
Instant bus bandwidth, max response timing, and memory utilization reports
Precise communications between each software group without many meetings
Allowed our Systems types to support the customer with minimal support from software
Eliminated all risk for software relating to the distribution
Truly supported the highest operationally with multiple hardware degraded modes
Major functionality can be added with very low cost and risk
Source code became easily traceable to requirements

Note: this aircraft had a particularly nasty one line requirement: Mission Replay. Which in the past translated into little pieces of code everywhere. Using this approach, a single software module is required and the data is entirely under the control of the systems person. By giving more direct control to system's engineers the software group had more time to do software.