With the latest set of rumors about IBM buying Sun, I've been thinking about that phrase and I think IBM can and should make it far more true that Sun ever could.
I've been working with DB2 replication and looking at DB2EE and I've been rather disappointed, so let me let you in on my dream. You have a DB2 installation, and it never goes down, because it covers the globe. When you power on a node, say a single processor intel box, or an RS6000 64-way box or a AS400 system or a mainframe cluster, it loads DB2, checks the local processing power against the licenced power it's allowed (which also feeds into the "global optimizer") and the connectivity to the other nodes it finds nearby (with rules that can be programmed to the required level of strictness or looseness) then it checks the local data it has stored against what is found in the cloud of other nodes and replicates updates back and forth then it picks work out of the global run queue and continues the work of the installation. So say your thousand node intel mixed linux and windows DB2 "system" is a little sluggish at certain queries. You call in an IBM consultant and he looks through the logs and discovers that you have a locking problem with certain very large joins and he advises that you add a mainframe cluster to your network and adjust certain rules. The next week the new machines are added and load starts to shift over to them. The users see ZERO downtime, the system just starts working faster. This is all the next step up from IBM's current work on self-healing systems. (Which would definitely play a part in this as a node could say "Hi gang, I'm having a little bit of a problem with my RAID array, perhaps I should only do Read-only queries for a while here.") -HJC - ::: When replying to the list, please use 'Reply-All' and make sure ::: a copy goes to the list ([EMAIL PROTECTED]). *** To unsubscribe, send 'unsubscribe' to [EMAIL PROTECTED] *** For more information, check http://www.db2eug.uni.cc
