Chris Craddock writes: >Once you >have to undispatch work and dispatch something else in the interim >it matters a whole lot less whether the thing that the interrupted >work is waiting on lives on the same processor or on Mars. As soon >as two or more pieces of your application are no longer in the same >memory, it doesn't matter a whole hell of a lot how far apart they >are - particularly if you don't control the whole thing.
I was with you up until Mars. :-) Yes, there can be a big(ger) difference between same memory and not. But a radio link to Mars (metaphorically speaking) is also very different than an ethernet cable between two boxes in a server room. If it weren't, nobody would be sweating kilometer/mile counts in synchronous flavors of GDPS, for example. >You seem to be presuppose some innate rightness of having all of >the work in one big box.... Never said "one big box." Didn't even hint at that, in fact. I am hinting at the word "fewer," though. Although boxen isn't the only measure I'd use or even necessarily the best one. I am prepared to advance the "radical" notion that the customer (who shall remain nameless) who has the following logical architecture for Internet banking (*simplified* high-level view): End User -> ... -> firewall -> HTTPS server -> firewall -> front-end presentation server -> back-end presentation server -> MQ server -> front-end Tuxedo server -> mid-tier Tuxedo server -> back-end Tuxedo server -> LU0 -> CICS Transaction Server -> .... ....has completely lost the plot. (And they know it now, which is good.) By the way, most or all of those tiers are clustered, and there are about three widely separated data centers involved in that flow as it executes, excluding DR centers. Yes, that's the (inbound) flow for "account balance, please." Rube Goldberg would be proud. :-) For perspective, later today I'm going to demonstrate my iPod connecting directly to CICS TS doing exactly the same thing. I said: >I'm getting more and more enthusiastic about not making things >more complicated than they need to be.... Einstein said: >"A system (theory) should be as simple as possible, but no >simpler." Einstein said it better, but obviously we agree. :-) For what it's worth, I am in general agreement with your comments about Web services. - - - - - Timothy Sipples Consulting Enterprise Software Architect IBM Japan, Ltd. [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
