11.01.2017 Views

A Technical History of the SEI

ihQTwP

ihQTwP

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Federal Aviation Administration Study<br />

The Challenge: Evaluate a Problematic FAA System Under Development<br />

By 1995, <strong>the</strong> FAA had spent $2 billion on a project to upgrade <strong>the</strong> in-flight air traffic control<br />

(ATC) system. The attempt to upgrade <strong>the</strong> system had gone through many well-documented and<br />

costly delays [Glass 1997]. At that time, it continued to fail some <strong>of</strong> its system tests; as a result, it<br />

still had not been put into service. The existing system, which had been launched a number <strong>of</strong><br />

years earlier, had obsolete hardware and s<strong>of</strong>tware and was becoming very difficult to maintain. In<br />

addition, because <strong>of</strong> rapidly expanding hardware advances, <strong>the</strong>re were not enough replacement<br />

hardware parts available (mostly transistors), and <strong>the</strong> hardware in use was slow.<br />

A Solution: <strong>SEI</strong> Architecture Evaluation Methods<br />

The <strong>SEI</strong> was contacted by senior <strong>of</strong>ficials at <strong>the</strong> U.S. Department <strong>of</strong> Transportation (DoT) and <strong>the</strong><br />

DoD to conduct an independent technical assessment to determine if <strong>the</strong> system could be salvaged.<br />

This work by <strong>the</strong> <strong>SEI</strong>, in a joint project with Lincoln Laboratory, represented one <strong>of</strong> <strong>the</strong><br />

first times that <strong>the</strong> <strong>SEI</strong> had done work on behalf <strong>of</strong> a non-DoD federal agency. Lincoln Laboratory<br />

personnel had domain knowledge <strong>of</strong> ATC systems and knew <strong>the</strong> FAA well, and <strong>the</strong> joint<br />

<strong>SEI</strong>/Lincoln Laboratory team worked toge<strong>the</strong>r very successfully. The FAA also provided <strong>the</strong> team<br />

with three members to contribute insights from <strong>the</strong> FAA perspective and provide unfettered access<br />

to personnel and documents. The team was asked to provide input to FAA management on<br />

<strong>the</strong> technical question <strong>of</strong> whe<strong>the</strong>r <strong>the</strong> upgraded system was salvageable and, if so, what would<br />

need to be done to feel confident that <strong>the</strong> system would perform at an acceptable level. If <strong>the</strong> system<br />

was not salvageable, <strong>the</strong> FAA was prepared to walk away from <strong>the</strong> two billion dollar investment<br />

that it had made.<br />

The <strong>SEI</strong>/Lincoln Laboratory/FAA team split into subgroups to investigate <strong>the</strong> areas <strong>of</strong> modifiability,<br />

availability, scheduling, code quality, documentation, s<strong>of</strong>tware tools, maintenance environment,<br />

testing, change management, and s<strong>of</strong>tware processes.<br />

To address <strong>the</strong> question <strong>of</strong> modifiability, <strong>the</strong> <strong>SEI</strong> developed a set <strong>of</strong> likely change scenarios. At<br />

that time, <strong>the</strong> <strong>SEI</strong> was just beginning to develop scenario-based methods for evaluating s<strong>of</strong>tware<br />

architectures. The prime contractor was asked to demonstrate <strong>the</strong> amount <strong>of</strong> change that would be<br />

required for each <strong>of</strong> <strong>the</strong> likely changes to <strong>the</strong> system. As a result, <strong>the</strong> subgroup determined that<br />

with regard to modifiability, significant changes could be put in place with relatively little disruption<br />

to <strong>the</strong> system. This was <strong>the</strong> first practical application <strong>of</strong> a method for assessing <strong>the</strong> quality attribute<br />

<strong>of</strong> modifiability, and this method formed part <strong>of</strong> <strong>the</strong> foundation for <strong>the</strong> <strong>SEI</strong> S<strong>of</strong>tware Architecture<br />

Analysis Method. In turn, after much future analysis, some <strong>of</strong> <strong>the</strong> ideas for SAAM later<br />

evolved into <strong>the</strong> Architecture Trade<strong>of</strong>f Analysis Method and <strong>the</strong> Quality Attribute Workshop.<br />

The o<strong>the</strong>r subgroups focused on <strong>the</strong>ir respective topics. One group analyzed <strong>the</strong> quality attribute<br />

<strong>of</strong> availability because <strong>the</strong> system had been developed to have extensive hardware and s<strong>of</strong>tware<br />

redundancy, and many <strong>of</strong> <strong>the</strong> decisions on <strong>the</strong> structure <strong>of</strong> <strong>the</strong> system had been made to ensure<br />

availability. The availability assumptions were examined; while many different structures for<br />

availability could have been developed, <strong>the</strong> one in use was judged to be adequate.<br />

CMU/<strong>SEI</strong>-2016-SR-027 | SOFTWARE ENGINEERING INSTITUTE | CARNEGIE MELLON UNIVERSITY 245<br />

Distribution Statement A: Approved for Public Release; Distribution is Unlimited.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!