On Feb 7, 2009, at 11:04 AM, Jan Lehnardt wrote:
Hi Damien,
On 7 Feb 2009, at 16:47, Damien Katz wrote:
So I propose supporting 2 different transaction models:
[...]
With these 2 transactions models, it's possible to deploy the same
apps on a single machine or a huge partitioned cluster. To support
the current model, it's only possible to deploy apps on a single
machine. I propose we drop the current model as bulk transactions
are not supportable in clustered or replicated set ups.
To be absolutely clear:
You propose to drop the old behaviour in favour of the two
new ones that get their own distinct API endpoints (which
are to be determined). The user could then choose which
type of trade-offs to take by choosing one of the endpoints.
You are _not_ proposing to drop the old behaviour and we
now have to decide on one of the two new behaviours to
replace it.
I am proposing removing the old behavior completely, in favor of the
two new ones, which will be selectable by the client.
Is that correct?
Cheers
Jan
--
