System Optimization and Quality
The System Optimization and Quality (SOQ) department focuses on developing technologies, tools, methods and services, for optimizing system performance and assuring software quality for the benefit of IBM system designers, partners and customers.
We specialize in developing and deploying technologies for identifying and eliminating software defects across all phases of the software life cycle. Our work involves innovative tools and methods for designing reliable and correct systems, using advanced software test and analysis techniques that aim to extend state of art in the field of software quality. All in all, our work has resulted in widely recognized contributions to the reliability and quality of software systems developed by IBM, customers and partners. These contributions (e.g., combinatorial test design) have been recognized by many papers and patents, and were recently highlighted in IBM's Annual Investor Briefing (New York, April 2012).
Some of the tools and methods we've developed over the years for improving the quality of software are:
- Code coverage and efficient test selection techniques (Focus, active project)
- Detection of software errors in concurrent systems (Concurrent Testing, completed project)
- Efficient collaborative code reviews (CCRT, completed project)
Our team is also known as an IBM center of competence in code optimization and compiler technology. Recently, we engaged in several exploratory projects where we aim to deliver breakthrough performance through dynamic recompilation and optimization of binary code. A key part of our work in code optimization is the area of post-link optimization (i.e., FDPR), which optimizes executables to gain additional performance improvements beyond that provided by compilation. Our work in this area resulted in significant performance improvements to the object code of many important applications and systems, measured on benchmarks such as SPEC, TPC-C, and EEMBC. With the emergence of cloud computing and virtualization as mainstream computing paradigms, our researchers extend our base set of tools with cloud-ready capabilities. Most recently, we began to work on workload modeling and generation, aiming to enable the study of performance patterns and limits cloud-based systems.
The SOQ team routinely works with IBM development groups to support customers who demand high performance of complex applications running on IBM platforms. Our team is extensively engaged in various activities for IBM and its customers in this area. By inventing, deploying and supporting innovative tools and methodologies, the team has improved the performance of the IBM platforms and the quality of the software they run.
The SOQ department provides direct consultancy and services to customers in the areas of software quality and performance of complex systems. Contact us for more details.
Groups & Teams