Subject | Re: [IBO] GSG057. Transaction Terminology/Optimistic andPessimistic Locking |
---|---|
Author | Raymond Kennington |
Post date | 2002-10-13T15:43:09Z |
Helen Borrie wrote:
It was the grammar in the sentence that obscured the meaning.
--
Raymond Kennington
Programming Solutions
W2W Team B
>Thanks.
> At 04:19 PM 13-10-02 +0930, Raymond Kennington wrote:
> >What is the following sentence supposed to read and what is it supposed to
> >mean?
> >
> >"InterBase's row-versioning architecture makes it unnecessary for clients
> >don't do
> >locking."
>
> In ISAM databases, you have been used to your "client" application (a
> misnomer) writing a lock record to the database's network lock file in
> order to keep others from updating records that are in the purview of this
> client. InterBase handles locking internally and isolates the work of one
> client (transaction) from another by keeping each transaction's view
> consistent throughout the transaction.
>
> The degree of consistency depends on the isolation level of the
> transactions. That, along with other transaction attributes, determine the
> outcome of attempted updates ("posts") by different transactions that are
> "engaging" the same data. IB clients don't do locking. It's
> unnecessary. The reason it's unnecessary is that the row-versioning scheme
> provides an entirely different mechanism from the client-locking mechanism
> you are used to.
>
> If you want to find out more about the row-versioning architecture, a.k.a.
> "multi-generational architecture", or "MGA", pick up some of the worthy
> documents from www.ibphoenix.com and/or ask your database-specific
> questions on ib-support.
>
It was the grammar in the sentence that obscured the meaning.
--
Raymond Kennington
Programming Solutions
W2W Team B