[cap-talk] Domain change (IPC?) overhead

[Jonathan S. Shapiro]

On Wed, 2008-03-26 at 03:12 +0000, David-Sarah Hopwood wrote:
> Kevin Reid wrote:
> > On Mar 25, 2008, at 17:13, David-Sarah Hopwood wrote:
> > 
> >> If, for the sake of argument, a system were designed to provide space
> >> accounting at vat granularity but was otherwise identical in semantics
> >> to E, would that system "scale up successfully", in your opinion?
> > 
> > E is not intended to prohibit space accounting at vat granularity; it  
> > is even desirable for implementations to do so.
> 
> I will rephrase: suppose that a system were to *require* space accounting
> at vat granularity, but was otherwise identical in semantics to E, would
> that system "scale up successfully", in Jonathan's opinion?

I don't see why not. The requirement is to be able to draw boundaries at
which space accounting can be performed and GC effects are mutually
isolated. Given a mechanism for accomplishing this, the decision about
how big the domain involved should be is an engineering decision.