By DaveLevy on Nov 15, 2006
Gavin Maltby, presented on Solaris' Fault Management Architecture. It is so hard to really get home how some of the very low level features of Solaris are unique and compelling competitive advantage. We all know hardware fails, but systems don't always have to do so. FMA allows Solaris to fail through some faults that bring other UNIX (or UNIX like) systems down and to recover rapidly from others. The engineering team have revisited the UNIX panic thing and improved Solaris' behaviours and diagnostics. This is an availability technology and as those of us who have worked on Consolidation Economics know, downtime costs money. In fact, the wrong (or too long) downtime causes business to fail. (Ask me about my local video shop some time?)
I am not sure that this meeting was the best for me to attend since most of my time and thinking is much closer to systems users as opposed to the hardware and this software is very close to the hardware design. The hardware designers should now be designing for it. Its one of the technologies that stems from Greg Papadopolous 5th eigenvector. "Best on Solaris". It's because we design the CPU, system and OS, we can create this technology and deliver the improved uptime benefit.
It was good to meet the other attendee and I shall be back for the January meeting. You might like to join me since there's free booze. (There's a December meeting but its on the 20th and I don't think I'll make it.)