Thoughts on Microsoft cluster offerings

I haven’t seen it yet. Eventually when I get more time, I want play with it, but from what I have heard it is not ready for prime time yet. That said, I would like note that the Cygwin tools are really good. I just built LAM-7.1.1 using them. Tried it out and it works quite nicely (at least on this box). A major coup for Microsoft would be if they tossed their existing SFU bit and used Cygwin. What they get right away is lots of codes running natively on windows via Cygwin. Would would be even better is if they spent some programmer group cycles to help make that better (it leverages the windows platform), so that end users can more easily move codes between unix/linux/windows. Think about it. This would lower the barriers to adoption, as you could easily make the case that OS support for a number of things is not an issue. We would look to use Cygwin as a platform for this on windows. For one thing, we could start up an sshd daemon (remote shell login) which curiously solves many problems that the linux/unix world needs. If Microsoft hooked that sshd to use the local windows authentication system …. (sound of hints being dropped, hopefully Redmond is listening). All of these things would help to lower the barriers that exist to adopting windows as a computing platform for clusters. Then there are the issues of running virus checkers on every node … There may be ways around that as well. Some more annoying than others. Short version: if you remove all the virus vector entry points, and have all important code run in a least priveledge environment (which should be the norm!!!) then you ought to able to prevent most of the attacks from escaping a single machine. Which is what you want for a cluster. Scan the binary upon entry into the cluster file system, and then run in a minimal priveledge mode. Even run in a nice chroot-like container if possible. Well that’s our free advice for Redmond. If they want more, they are going to have to hire us to consult for them.