In my post, Single system image clusters: an idea whose time has come and gone?, I asked the rhetorical question "Are single system image clusters an idea whose time has come and gone?" I must admit that I had a somewhat proprietary interest in the technology. At an early point in my career, I was the North American VAXcluster programs manager for Digital Equipment Corporation's Software Services.
I was pleasantly surprised by the strong level of response to that question. Although each of the messages I received were worded differently, they all carried the same basic message. SSI clusters are still relevant in a Web 2.0 world because 1) they make fantastic environments for established, business critical applications and 2) they're strong, secure, high performance platforms for database software for today's applications.
It's good to know that something I worked on back in the early 1980s remains relevant today.
HP, today's shepherd of DEC's SSI clustering, Oracle, developer of Real Application Clusters (RAC), and Sun, developer of the Solaris Cluster, are strong proponents of "shared everything" clustering and I hope to have a chance to speak with them about their views on the future of SSI clustering.
What questions should I put to these suppliers when I next speak with them?