Re: The TransRelational Model: Performance Concerns
Date: Fri, 12 Nov 2004 17:48:33 +0100
Message-ID: <p8q9p0to4pgntrt4j7se906n4dv4pdqv71_at_4ax.com>
On Fri, 12 Nov 2004 14:37:01 +0100, Troels Arvin <troels_at_arvin.dk> wrote:
>> Curious, they were rather easy to solve to me.
>
>How? - Did you simply reserve room for expansion in newly allocated blocks?
Well, this is more or less the solution propossed in the patent's text: to reserve room for expasion in new allocated blocks and to periodically reorganize the old blocks in order to have room for expansion. Nightly reorganizations for instance.
But I choosed a different approach based on paged BTrees (Perhaps I should patent it :).
It is not more complicated, we don't have the problem of ocasional exhausted room, periodical reorganizations are no longer needed and the use of memory is more efficient.
>Apart from that, your comment makes me curious: You seem to have been
>involved in an implementation of TRM.
Yes, but unfortunately I am having important financial problems to finish it :(. It is not easy to find support for something like this in a semi-developed country :(
> Would it be rude to ask which
>product that is?
No, but it is still a prototype.
Regards Received on Fri Nov 12 2004 - 17:48:33 CET