[opensource-dev] Side topic related to cache (was Re: User Story: Improved Cache
jbhancroft at gmail.com
Sat Sep 18 08:37:50 PDT 2010
I was just sketching this up yesterday, in a brainstorming session with some
folks about how
we thought the v.w. systems would be morphing over the next 2-5 years.
A co-operative P2P network of users who would be willing to opt into
"participate in cache sharing"
would be a good idea. Having the cache be fairly intelligent - broadcasting
requests and then maintaining
a list of "fast and available" similarly co-operative viewers. It definitely
offloads comm work from always
going back to the SL servers over and over. This problem has been solved
many times by many projects.
It seems to me that it would improve the user experience "Fast"... and also
impact LL's operational budget
significantly (in a good way :).
On Fri, Sep 17, 2010 at 5:37 PM, Altair Sythos <sythos at gmail.com> wrote:
> On Fri, 17 Sep 2010 17:14:01 -0400
> Ponzu <lee.ponzu at gmail.com> wrote:
> > I have also thought a bit about the situation where two or more users
> > are on the same subnet. Think of a dozens of people in the same LAN
> > in a meeting, for example.
> > Current system has each user building an almost identical cache
> > locally.
> > Instead, what if all the users went through a proxy, and the proxy did
> > the caching? Wouldn't that be a big win?
> by the way viewer support proxy you can use a local caching proxy
> (oops, squid, sort off) and already extend *ALL* you want (HTTP get
> textures.... cached in a standard HTTP cached server)
> Policies and (un)subscribe information available here:
> Please read the policies before posting to keep unmoderated posting
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the opensource-dev