News
Proof of Concept grant for Andy Tanenbaum
Prof. Andy Tanenbaum recieved a Proof of Concept grant from the European Research Council for his project about moving the MINIX 3 system (www.minix3.org) to the ARM processor. The Proof of Concept grant is for 150.000 euro.
The Proof of Concept grant is only open to people who have already won an ERC Advanced Grant or an ERC Starting Grant, so the applicant pool is very strong. There were 78 proposals submitted. Prof. Tanenbaum was rewarded an Advanced Grant of 2.5 million euro earlier. The purpose of the Proof-of-concept grant is to allow winners to move the results of their research closer to the market.
Andy Tanenbaum: "In my case, the Advanced Grant is about making operating systems more dependable. Part of this acknowledges that software contains errors (bugs). No amount of effort can produce error-free software. We have 50 years of experience with this. What the research is trying to do is find ways to build fault-tolerant software, that is, software that continues to function in spite of errors. Aircraft, for example, contain a lot of redundant hardware so if one component fails, there is a backup that can be used. Operating systems don't have any kind of fault tolerance. If a fault is triggered, the system just crashes (e.g. a Windows blue screen). What we are doing in the research is making it possible to recover after a fault and continue execution.
Another facet of our research is about being able to upgrade the operating system without stopping it. With current systems, to upgrade you have to stop the computer, then reboot it with the new one. For servers at many companies, banks, etc., stopping the computers is not acceptable, so our work on live update could benefit systems that must never go down."
MINIX 3
"The Proof of Concept grant is about moving our system, MINIX 3 (www.minix3.org) to the ARM processor, which is widely used in embedded systems. These are systems run by a computer but don't look like a computer. A modern car has probably 100 computers in it. An airplane might have 1000. Telephones, televisions, pacemakers, MRI scanners, hearing aids, and thousands of other devices have an embedded computer. What we aim to do is bring the results of the fundamental research closer to use in these embedded systems.
As a simple example, if a bug if found in the software of a pacemaker, it would undesirable to turn off the pacemaker then reboot it with the new software because the patient might die while it is off. Clearly a live update of the software while the system is running is desirable here. No current system can do this. We hope our research will show the way."
