MINIX 3: Status report and current research

作者：Andrew S. Tanenbaum, Raja Appuswamy, Herbert Bos, Lorenzo Cavallaro, Cristiano Giuffrida, Tomas Hruby, Jorrit N. Herder, Erik van der Kouwe, David C. van Moolenbroek · 发表于：VU Research Portal · 年份：2010 · 被引用次数：14 · 研究领域：Distributed and Parallel Computing Systems、Big Data and Business Intelligence、Cloud Computing and Resource Management

I n m y F e b r u a r y 2 0 1 0 c o l u m n [ 1 ] , ; L O G I N : J u N e 2 0 1 0 m u s I N G s 5related designs makes development and support (the compilers and libraries) cheaper.In the world of embedded systems, you can order systems-onchip (SoC) with just the processing and I/O support features your application requires.But there will be extra cost for anything customized.Volume manufacturing drives price down quickly.For now, I expect to see multicore CPUs with homogeneous cores.I don't expect to see cache-coherent systems with more than 16 cores (although you can already have quad-CPU systems with more than 16 cores).I say this because of the limitation on the amount of power available in single-die designs and the pressure on memory busses required by cache-coherency.Designs like the Intel Single Chip Cloud use much simpler CPUs and dispense with cache-coherency in exchange for message passing.If you read the MINIX 3 article in this issue or read about Barrelfish in the April 2010 issue, you will see that message passing appears to be the way forward for multicore systems. LineupAndrew Tanenbaum, with help from many co-authors, updates us on what is happening with MINIX 3. MINIX 3 is a modern operating system designed both as a platform for learning about OS design and as a real OS with a BSD-style license, with reliability, flexibility, and security highest on the feature list.I like what I read about the future directions of the MINIX 3 project.And where Linux (or So...