>>> On Fri, Aug 31, 2007 at 6:08 PM, in message <[EMAIL PROTECTED]>, Tom Marchant <[EMAIL PROTECTED]> wrote: > On Fri, 31 Aug 2007 15:07:53 -0600, Mark Post wrote: -snip- >>Even on midrange systems it isn't all that cheap when you get >>into shared/virtualized environments. If the admin wants to give >>64GB to every guest, you won't get many guests. > > Oh, really? Sure you can. z/VM can handle it very nicely, thank you.
Not without a lot of paging space to back it up, and not without a severe performance hit if the machine only has, for example, 64GB of real to start with. My point being that the "RAM is cheap" mantra starts to fail when a box is only so big, and it needs to be shared among numerous entities. We've been doing it in the mainframe world for a long, long time. The midrange folks are just now starting to get an inkling of what shared/virtualized means in terms of application design, system requirements, etc., etc. Oh, and if you create a 64GB z/VM guest, shame on you. As someone who is very heavy into z/VM performance once told me, "z/VM is very good at managing large numbers of small things. It's not so good at managing a smaller number of very large things." I tend to agree. The z/VM scheduler isn't too happy about guests with large working sets. -snip- >>To answer your question, >>yes, that would mean they have no virtual storage. > I take it you disagree with Tony Harminc's post. Paging is only one part of > virtual storage. The other part is the mapping of virtual addresses to real > addresses. I would take the position that the two are both necessary, but not sufficient. I really think you need both of them to achieve what most people think of as "virtual storage." Mark Post ---------------------------------------------------------------------- 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
