On Mar 23, 2013, at 10:38 AM, Arturo Servin <[email protected]> wrote:

<snip>

>       This would help to move a problematic CA out of the monolithic repo.
> However, this would make a cache to create multiple rsync sessions to
> retrieve objects and lowering its performance (according to Rob's
> measurements with RIPE NCC repo). Isn't it?

Just to make a general observation: I'm a little worried that this is a case of 
the tail wagging the dog…  The issue that Chris seems to be raising (at least, 
to me) is one of scale and resilience… It seems to me that our efforts to 
optimize _rsync's_ performance at all costs is starting to require us to relax 
our security and stability concerns.  At a high level, I think that actually 
points to rsync as a suboptimal component.  Regardless of the transport used, 
Chris' comments seem sound to me, and I haven't seem too many comments that 
rebuff his general concerns.  The fact that rsync doesn't fit well into the 
model of securing repositories while still making them lively enough for caches 
is worrisome to me.  I would think that it ought to suggest something other 
than relaxing our security and stability concerns…  Just my 0.02…

Eric
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to