tag:blogger.com,1999:blog-2170637149389736039.post2023094898239398840..comments2023-10-30T06:20:46.020-07:00Comments on Simple Oracle Dba: What is SmartDBPdVhttp://www.blogger.com/profile/11518325134965208858noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-2170637149389736039.post-28309211828213687352018-10-05T02:26:11.880-07:002018-10-05T02:26:11.880-07:00Yes, the Oracle database is supported on Windows 7...Yes, the Oracle database is supported on Windows 7. You can submit SQL statements either as a client to a server or run the entire database on Windows 7, such as Oracle Database Express Edition 11.2. <br /><br /><a href="https://mspsupport.co/" rel="nofollow">managed services support</a>Anonymoushttps://www.blogger.com/profile/05956946459332823545noreply@blogger.comtag:blogger.com,1999:blog-2170637149389736039.post-72525024024646782402018-04-17T02:49:48.654-07:002018-04-17T02:49:48.654-07:00Piet
I understand that #SmartDB is more than just...Piet<br /><br />I understand that #SmartDB is more than just ACID. Having had some discussions with many intractable MVC/OO developers the #SmartDB advocates will need to come up with some very compelling reasons to move to this architecture.<br /><br />I liked your 'Requirements, Reasoning, Recommendations' approach and this was my attempt at a first requirement. A number of these will need to be formulated in order to stand any chance of convincing others to use this architecture. Both to move the work into the database and then expose functionality via a PL/SQL (Oracle) or stored procedure (Other) shell.<br /><br />To me, correctness is the ultimate requirement and so is the one I commented on. There are other requirements - performance, security, etc - which should have their own 'Requirements, Reasoning, Recommendations'.<br /><br />Even having proved the incorrectness of concurrently processing data using unit tests in MVC/OO architecture - and also its correctness in a #SmartDB architecture; and having presented on the mathematical and logical basis underlying the relational model (to which, in my experience the vast majority of Database Developers are ignorant) I have still failed to convince management to move to a #SmartDB approach.<br /><br />We will certainly have a hard-time convincing others of the benefits of the #SmartDB architecture in the current climate.<br />Johnnoreply@blogger.comtag:blogger.com,1999:blog-2170637149389736039.post-49043498921141732602018-04-16T08:06:45.339-07:002018-04-16T08:06:45.339-07:00John, Toon,
Thx for your reactions. I'm resur...John, Toon,<br /><br />Thx for your reactions. I'm resurrecting this topic because, well.. "Good Architecture" remains an issue. <br /><br />I will be presenting in Riga (RigaDevDays) and Bulgaria (BG-OUG) on #SmartDB. I noticed Bryn Llwellyn is also speaking at RDD, and I know he has specific ideas on #SmartDB. <br /><br /><br />@John: I Agree with your statements, but I have the impression your reasoning-paragraph describes "ACID" and then some. I would add that #SmartDB is more then just ACID. It is also a concept whereby you maximize the work/logic/code IN the database and minimize the work/logic/code "Outside" of the database. In the extreme case (Bryn?) this gives you a database that only exposes schemas that can do a given number of "whitelisted" calls to PL/SQL. Security as an added bonus.<br /><br />Any other thoughts out there? PdVhttps://www.blogger.com/profile/11518325134965208858noreply@blogger.comtag:blogger.com,1999:blog-2170637149389736039.post-89793852860631548322018-02-09T07:13:01.424-08:002018-02-09T07:13:01.424-08:00Hear, hear, John.
You should love SQL assertions, ...Hear, hear, John.<br />You should love SQL assertions, once they arrive (if ever).<br /><br />;-)Toon Koppelaarshttps://www.blogger.com/profile/08153913435188725112noreply@blogger.comtag:blogger.com,1999:blog-2170637149389736039.post-38784281224237566372018-01-07T11:34:16.403-08:002018-01-07T11:34:16.403-08:00Requirements
To ensure that the data conforms to t...Requirements<br />To ensure that the data conforms to the business rules 100% of the time.<br /><br />Reasoning<br />A changeable database is mutable state. Simultaneous access to the database results in parallel processing of this state. The combination of parallel processing and mutable state results in non-deterministic behaviour. A non-deterministic system is essentially broken. In order to fix this broken non-deterministic system either the database needs to be made immutable or the parallel access to the mutable state needs to be synchronised to make it single threaded. Since an immutable database is not possible in most systems, concurrency control mechanisms need to be included to synchronise access to the mutable state. The mutable state includes both the data and any business rule that could be formally expressed in the logical model of the database.<br /><br />Recommendations<br />The business rules are implemented within the DBMS declaratively where possible; assuming the implementation of the declared rule includes such concurrency control mechanisms as required to make all changes to the data within the database deterministic.<br />Where a declarative implementation is not possible the business rules are implemented procedurally, also explicitly including such concurrency control mechanisms as required to make all changes to the data within the database deterministic.<br />For an example procedural implementation in Oracle please see the examples Advance Mathematic for Database Professionals.Johnnoreply@blogger.com