Re: Apple Crate 2e for auction???

["mdj" <mdj.mdj@gmail.com>]

Michael J. Mahon wrote:

> In the more general case, if the amount of computation to be done on
> a machine exceeds the amount of time required to package and send the
> "problem" and get and parse the "result", then it's practical to
> distribute the application.
>
> This is true at all scales, so it is a "fractal" truth.  ;-)

Michael, that's an excellent turn of phrase :-)

Somewhat off-topic anecdote:

Back in the early of web application development, application server
vendors adopted a partitioning strategy of separating an application
into web, application, and data (being actual database or other data
source), one which paralleled the typical conceptual partitioning of
presentation, logic, and data/model.

At the time, the big draw card of this approach was that you could then
'scale' any of those three tiers, depending on which required more
horsepower.

Unfortunately for those who followed this approach, this model defied
the above mentioned "fractal truth" and distributed tasks that were
more optimally performed locally across address boundaries and onto
networks with very high contention (that increased exponentially as
requests rose linearly)

Somewhat disappointingly, the optimal partitioning scheme for such
applications, which is to rendezvous only when data is committed to
storage, is seldom implemented - generally due to the high cost of
distributed filesystems or horizontally scalable DBMS's. The next most
optimal scheme reigns due to the extra hardware requirements not
exceeding the software cost of the 'better' implementation.

Matt