[email protected] (David Devine) writes:
> Z/series has had such nice to have's as GDPS (about 10-15 years)
> multiple pathing to devices and system manged storage (25 +).
> It's only in the last few years that other platforms have started to
> catch up in these area's and their idea of multiple paths is generally
> 2.
> (This is a broad sweep, there may well be kit out there thats all singing and 
> dancing)
> (Ibm I series & P series could be classed as junior mainframes having
> evolved from System 34 & 36 (cut down System 360's) and are sloooowly
> getting Z/series features.)
>
> Staff costs? 
> once you've got a Z/series site setup which has skilled support staff
> (not including application programmers & developers) you can pretty
> much expand up to 10 times the kit and plex's (and probably a lot
> more) with minor staff increases if at all.
>
> How many people does it take to manage windows or unix estates ? where
> i've worked over the years you are talking 4 or 5 times as many people
> as mainframe support staff.
> And thats just support. 
>
> Once you include the dozens of Android, java & C++ developers and
> proggies you are going to need to actually produce something
> worthwhile, you can only afford to buy cheap kit!
>        
> This is why you need to consider "Total cost of ownership" and it is
> not solely limited to financial payback period and capital
> depreciation write off's; staff & running costs are often overlooked
> and reliability freqeuently is.

re:
http://www.garlic.com/~lynn/2013b.html#5 mainframe "selling" points
http://www.garlic.com/~lynn/2013b.html#6 mainframe "selling" points
http://www.garlic.com/~lynn/2013b.html#7 mainframe "selling" points

disclaimer, in the late 70s, I had done some work for LLNL doing
some 4341 benchmarks when they were looking at a large compute
farm of 4341 machines ... reference in the ibm-main "millions of
cores" thread
http://www.garlic.com/~lynn/2013.html#77 OT: but hopefully interesting - 
Million core supercomputer

and over the yrs I was involved in various & sundry other things with
LLNL ... including being asked in 1988 to help LLNL standardize some
serial technology they had ... that morphs into fibre-channel standard
(later the enormously heavy-weight FICON standard is layered ontop of
FCS ...  which significantly cuts throughput).

the majority of the i86 server chips (as opposed to desktop, laptop,
personal, etc chips) are going into the large cloud (public & private)
mega-datacenters (along with various large supercomputers).

for other drift recent (a.f.c.) discussion of s/38 evolving into as/400
... with various features to also support s/34&s/36:
http://www.garlic.com/~lynn/2013b.html#3 New HD

initially power was targeted at server market and power/pc was at the
personal market ... including use by apple. apple then moved over into
personal i86 platform when it didn't look like there would be the
necessary investment to support power/pc in the personal market
... especially lower powered laptops.

IBM has been doing about $5B/annum each in the three hardware platforms
(mainframe, power/risc, i86) ... for mainframe that works out to be
approx the equivalent of 180/yr fully configured 80processor z196 @$28M.
... at 50BIPS/per ... that would come to 9TIPS/annum. By comparison, the
e5-4600 blade is coming in around 1TIP (and e5-2600 blade at 527BIPS). A
single rack of e5-4600 blades is likely more processing power than the
total of all z196 machines sold.

the millions of cores ibm-main post mentions having done the cluster
scaleup for the IBM HA/CMP product ... but then that was transferred and
we were told we couldn't work on anything with more than four
processors. past posts mentioning ha/cmp
http://www.garlic.com/~lynn/subtopic.html#hacmp

I had also coined the terms "disaster survivability" and "geographic
survivability" out marketing ha/cmp. Somewhat as a result, I was asked
to write a section for the corporate continuous availability strategy
document ... but then the section got pulled when both pok and rochester
claimed that their products couldn't meet the objectives.  misc. past
posts mentioning availability:
http://www.garlic.com/~lynn/submain.html#available

-- 
virtualization experience starting Jan1968, online at home since Mar1970

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to