Software tool allows programs to run faster without sacrificing security
By Paul Ridden
April 3, 2011
Protecting computer systems from malicious attack using complex software solutions is a constant, but necessary, struggle. As threats become more sophisticated, the technology used to fight them off can impact more and more on system performance. According to researchers from North Carolina State University, programs that have built-in safety features can be slowed down by as much as a 1,000 percent. To combat this issue they have developed technology that helps such programs to run more efficiently without sacrificing effectiveness.
Safety features known as meta functions "can slow a program down so much that software developers will often leave them out entirely," according to team leader Dr. James Tuck. "Leaving out those features can mean that you don't identify a problem as soon as you could or should, which can be important - particularly if it's a problem that puts your system at risk from attack."
The new tool takes advantage of the multi-core architecture of modern chips to process the safety features over different cores to those used by the main program. Although the multi-core approach is not new, the new helper technology developed by the Department of Electrical and Computer Engineering team is said to significantly improve the efficiency of the parallel processing of off-loaded safety features.
"To give you some idea of the problem, we saw the application we were testing being slowed down by approximately 580 percent," said Dr. Tuck. "Utilizing our software tool, we were able to incorporate safety meta functions, while only slowing the program down by approximately 25 percent. That's a huge difference."
The software tool is implemented as a plug-in for the Gnu Compiler Collection of software tools and functions automatically, requiring no manual reprogramming.
A Paper detailing the new development - entitled Automatic Parallelization of Fine-Grained Meta-Functions on a Chip Multiprocessor and co-authored with Ph.D student Sanghoon Lee - will be presented at the International Symposium on Code Generation and Optimization in Chamonix, France on April 5.
Dr. Tuck says that his team is currently fine-tuning the software ahead of an open source release later this year.