Brad Chapman wrote: > > It seemed to me the plan is pretty controversial since it proposes > that all remote communication occur through the definition layer (the > middle scripting layer) and not at the level of node processing. It > was based on my attempts (probably failed) to learn about methods > for implementing distributed systems based on Jarl's suggestion that > all remote note authentication should occur in the definition layer. > The proposal is basically derived from some systems I read about where > the corba/distributed layer was implemented separately from the layers > that do the actual work (the processing layer in our case). I tried to > adapt it to the vsh setup. I thought communication (between cores) would occur in the brokering layer. This is what GMS was going to handle, right? If communication occurs through the Python-based definitions layer, what is the broker layer used for? It must be used for more than communicating with Overflow (processing layer). > In addition, since this system would probably require a > multithreaded server in the definition layer (to deal with possibly > simultaneous communication with the processing layer and multiple > remote implementations), this proposal led me to worry about using > ORBit for all our communication. As an aside, we will eventually need a multithreaded connection between the DL and the front ends. Cheers. Jeff -- +----------------------------------+ | J.W. Bizzaro | | | | http://bioinformatics.org/~jeff/ | | | | BIOINFORMATICS.ORG | | The Open Lab | | | | http://bioinformatics.org/ | +----------------------------------+