Red Hat has apparently also understood the necessity of real-time in financial trading – FSMLabs main current focus. Unfortunately for them, they are using inappropriate technology as shown by the over 1.2 million lines of code that they boast has gone into soft-real-time Linux. The dual kernel virtualization approach we invented and pioneered has the advantage of a correct design. There are definitely places where the “keep hacking on it until it’s ok” method of building software is good, but real-time operating systems and real-time networking is not one of those places. The fundamental problem we identified in the “mixed use real-time/non-real-time” operating system 12 years ago cannot be hacked through.
British Telecom breaks with the party line. This is the intersection of a business/political dispute about what business telecos should be in and a technical issue of QOS. Traffic shaping is a blunt instrument for QOS and part of the problem is that nobody bothers to try to characterize scheduling requirements more precisely than “needs to be fast”.
He confessed to programming malicious software codes into computers that track Navy submarines in May 2006 while in Naples. He told Navy investigators that he was upset that his company’s bid on a project was passed over. Sylvestre had fled Italy after he entered the codes.
The guilty party, one Richard Sylvestre, was a sys-admin at a US Navy System based in Naples Italy and he had a top-secret security clearance. So, without being to crass about it, the sales pitch about security as a property of who works on the software is once again shown to be simply ridiculous – see my old note on the subject. Security is a reliability issue that needs to be addressed in design and administration procedure. The network that was hacked was too trusting of its operators and the people who were responsible for administering the sys-admins were not doing their jobs.