Bob Richards writes: >It has been awhile since someone has referenced the >"Mainframe" site: > http://mainframe.typepad.com/ >Timothy Sipples has placed a nice post out there called >"How Many Mainframes Do You Need?" >It is worth the read, even for us old, crusty types who >sometime forget we didn't always know everything! <smile>
Thanks, Bob! I hope to encourage some "new thinking" with that post. You may also enjoy the embedded YouTube video. Shane writes: >The more the better ... >Unfortunately a shrinking pond is ultimately going to do >none of us any good. Physical infrastructure parsimony is a very good thing. Keep in mind what's ultimately relevant and important: how broad and how deep the end-user value is of the system(s). Why wouldn't you want to deliver the same (or more) value to users if you can do it with fewer boxes? What, is there some contest I don't know about to see who can collect the most frame metal? :-) No, save some of the money, and spend some of the money on delivering more value to your users. More value includes things like better/more application development capabilities, business information intelligence, improved operations and systems management, removing remaining (and inconvenient) service interruptions, additional consolidation (from other server types), and so on. "Do more with less" is what managers always want and what mainframers can deliver better than anyone. I agree with other commenters that each new model -- now the z10 -- should challenge previous assumptions, including the "how many?" assumptions. For example, several analysts figured out that the System z10 BC is the "consolidation platform for the rest of us" (for those of us not big enough to justify an EC machine). I agree with that. In writing that post, I am coming from the perspective of chatting with certain customers who, for example, still think that they should run their compilers on a separate physical machine -- probably even separately backstopped with its own disaster recovery machine! -- because somehow the compiler could magically leap across an LPAR boundary and threaten their production workloads. Never mind the fact that those production workloads are often running within the context of a physical Parallel Sysplex -- and sometimes even a 3-machine Sysplex! Come on, isn't this the year 2009? :-) Anyway, perhaps it strikes you as odd for the "IBM guy" to say "you don't need very many," but that's exactly what I'm saying. How many? "It depends," but I offer some of the typical thought patterns in that blog post, at least to get you thinking. - - - - - Timothy Sipples Consulting Enterprise Software Architect IBM Japan, Ltd. e99...@jp.ibm.com ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html