I think that the resources will be determined in great deal by how your your application handles it's data and processes. Remember that much of the interaction towards twitter Api will use network (bandwidth) resources and very little machine (CPU, Ram) resources. In our personal experience much of our resource optimization has been geared towatds minimizing API calls (Process waiting for api responses) and Not Hardware resources. Hope this helps.
BTW We are looking into optimizing our algorithm for API calls. Anyone want to discuss what were doing? On Jan 18, 5:50 am, techtimes <techf...@gmail.com> wrote: > Hi --- > > Is their any benchmark that would allow us to plan well into the > future for server resources? > > example: > > : we would be using the real time streaming API --- > > : 5000 users use our service: all would need to see and interact with > their "Home statuses time line-- > > : 1 to 2% are power users that have more then 1K +++ followers such > as R.Scoble --- and friends > > : would a - 4 core XEON 8GB - machine be enough for a plan of 5000 > users - and their respective followers time time status? > > Thanks > Regards > Joao