On Wed, 27 Aug 2003, [iso-8859-1] Red K wrote:
Yes, I also mentioned that proxylet is not discussed
for a long time. But what is the reason for this?
Is it related to system efficiency and simple
maintainance? I mean to setup one callout server is
easy to maintain and higher efficient to carry out
specific service than put them onto a single host.
I do not think that proxylets are always more or less efficient than
callout services. Same for the ease of maintenance. It all depends on
a particular environment and workload. Moreover, it is possible (if
not desirable!) to have a proxylet that communicates with the OPES
processor via OCP and, hence, is also a callout service.
Alex.
On Wed, 27 Aug 2003, Markus Hofmann wrote:
Reason is that the OPES WG is chartered to specify the callout
protocol and the rules language, but *not* a proxylet interface or a
proxylet architecture - this would be a local implementation
decision or standardized elsewhere.
The OPES architecture does *not* rule out proxylet, but the focus of
the WG is on the callout mechanism.
On Wed, 27 Aug 2003, Abbie Barbir wrote:
At this stage our charter does not focus on proxylet or how rules
are loaded to the opes processor.