On 7 Jul 2010 07:48:56 -0700, in bit.listserv.ibm-main you wrote: >--------------------------------------<snip>------------------------------------- > >>Ted, >> >>Well, I don't know what else to say Ted. These are real world examples that >>provided significant IO and IO Time reduction for Banking and Credit Card >>applications. I did not say "I recommend" these techniques, I said I have >>"used" these techniques for over 10 years (with the stress on used). >> >>I guess Banking Applications must be really unique in Canada, because this >>method worked for Banking and Credit Card applications for over 20 >>countries. >> >>I've worked for Banks for most of my working life, 20 years out of 32, if >>that provides any weight to my experience. There was life before HDS... >> >>Ron >> >> >-----------------------------------<unsnip>------------------------------------- >I've worked for manufacturing, trade associations and financial firms >and in each case I've known read/write ratios of about 4 reads to every >1 write. > >In my experience, with ARCHIVER, the compression/compaction of the data, >while CPU intensive, has provided a considerable advantage in I/O >reduction. My problem, which led to my initial posting, is that >sometimes I get a long load-module record that expands rather than >shrinks when I try to compress it. > >I may be in real trouble when 64K DASD records show up. :-)
What is interesting is that on the squatty boxes, single records can be multiple megabytes in size (a 22 megapixel RAW photo for extreme example) and we still worry about a 64K barrier. Even JPEGs can be over a megabyte and pictures in catalogs probably can be well over 64K. Even if most of this sort of thing is in Unix files or in databases, it would seem the rest of the systems need to be able to handle much larger logical entities. Clark Morris > >In another post, I noted my reasons for choosing compression/compaction >over segmentation techniques. > >Rick > ---------------------------------------------------------------------- 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
