tag:blogger.com,1999:blog-5942052796969657606.post4361431091901392156..comments2023-03-12T08:56:39.925-05:00Comments on Db2 for i: In Memory...Unknownnoreply@blogger.comBlogger7125tag:blogger.com,1999:blog-5942052796969657606.post-78235657520152445532019-06-06T13:35:47.497-05:002019-06-06T13:35:47.497-05:00Still a great article half a dozen years later. J...Still a great article half a dozen years later. Just had to point a manager to this link again today. Thanks Mike!Ken Kuhlmanhttps://www.blogger.com/profile/02079069537982453408noreply@blogger.comtag:blogger.com,1999:blog-5942052796969657606.post-22413884454230308892018-04-06T10:07:11.844-05:002018-04-06T10:07:11.844-05:00Hi Mike,
Excellent article.
If an object is in mem...Hi Mike,<br />Excellent article.<br />If an object is in memory from an interactive job, (*INTERAACT pool) and a batch job needs the same object, but batch runs in *SHRPOOL1, does the object have to be brought in to memory again because it's in a different pool?<br />With large amount of memory now available, even more on a new P9, I was planning on using SETOBJACC to keep some objects in memory in hoping to improve some batch jobs performance.<br />SETOBJACC needs to point to a storage pool. <br />The objects being considered for SETOBJACC might be used by both interactive (*INTERACTIVE) and batch (*SHRPOOL1)<br />My thoughts were run everything out of *BASE, SETOBJACC would point to *BASE.<br />I'm looking for some confirmation on the use of the multiple memory pools along with SETOBJACC.<br />Paul Steinmetzhttps://www.blogger.com/profile/16604078831928557864noreply@blogger.comtag:blogger.com,1999:blog-5942052796969657606.post-37914776284328069172015-08-05T11:21:17.687-05:002015-08-05T11:21:17.687-05:00Excellent question Darron. In general, IBM i is de...Excellent question Darron. In general, IBM i is designed and implemented to handle the dispatching of work to the processors as well as managing the I/O requests to minimize waiting and maximize throughout. When it comes to DB2 for i, using SQL and good data-centric programming techniques (set-at-a-time for example) is a critical success factor in allowing the system to optimize the query plan and execution. This "optimization" allows for CPU bound plans and smooth I/O, whether this includes in-memory techniques or not. Simply put, the DB2 for i Query Optimizer does what you are describing: maximize the potential of the computing and I/O resources, while being a good neighbor.Mike Cainhttps://www.blogger.com/profile/01481223716996299215noreply@blogger.comtag:blogger.com,1999:blog-5942052796969657606.post-46518532070502540252015-08-03T20:40:50.664-05:002015-08-03T20:40:50.664-05:00Hi Mike,
it's been a while since you wrote thi...Hi Mike,<br />it's been a while since you wrote this but a question for you anyway... despite working on the 'i' for 2 decades, I still like database tech and recently investigated an in-memory database for features (VoltDB in this case). <br /><br />Apart from #1. declaring the demise of the dinosaur databases (Oracle & DB2 were in that list - no mention of DB24i ) and #2. having a product to go the hard sell; there were some interesting aspects of the database that could contextually apply in the 'i' which you haven't addressed here. Being in-memory isn't the end of the equation.<br /><br />In particular is the concept of machine sympathy - the idea that the code is written in such a way to maximize the potential of a CPU. An simple example is ensuring that data for a file is always stored in such a way as retrieval of the data is possible with one operation. For an in-memory DB this literally requires that the data to access is contiguous in memory, no fragmentation. <br /><br />With the i, all I know of is that memory paging to/from disk, the fragmentation boosting disk I/O with more disk arms, but disk has always been in the equation. So, if there was a theoretical IBMi setup with pure memory (is that even possible?), does the OS consider contiguous arrangements or is it still page size blocks? Does the fact that 'i' is a virtual OS hinder or benefit this equation?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5942052796969657606.post-28770251562442355972013-11-22T05:00:58.406-06:002013-11-22T05:00:58.406-06:00A very good article. Thanks.
A very good article. Thanks. <br />Anonymoushttps://www.blogger.com/profile/15242446802338290545noreply@blogger.comtag:blogger.com,1999:blog-5942052796969657606.post-53127260268090447342013-10-25T00:00:40.112-05:002013-10-25T00:00:40.112-05:00Good article! Thanks!Good article! Thanks!Anonymoushttps://www.blogger.com/profile/11011888467539410354noreply@blogger.comtag:blogger.com,1999:blog-5942052796969657606.post-16131257200459887652013-10-07T14:01:53.507-05:002013-10-07T14:01:53.507-05:00Amen, Mike! Great post..Amen, Mike! Great post..Susan Gantner, System i Developerhttps://www.blogger.com/profile/09368310819738269259noreply@blogger.com