> But if you have two cache servers, perhaps you would not even notice > while the second one is getting the whole repository(ies).
This isn't about rebuilding an entire cache, it's about one off changes to that cache. What I'm hearing is that one-off changes could take minutes, hours, or even days in some circumstances. In a world that moves in the subsecond range, seconds are the minimal requirement, not minutes. There's more here than the startup case. What will be the scaling of getting even a single change out to hundreds of thousands of caches, and from thence to hundreds of thousands of routers, so a change in the correct origin can actually be reflected in the policy tables of the Internet at large? It seems, to me, like a pertinent question --and hand waving isn't a good answer at this point. Russ P.S. That we're discussion how well a system will scale and perform after it has been designed and standardized seems a little odd, at least to me. -- <>< [email protected] [email protected] _______________________________________________ sidr mailing list [email protected] https://www.ietf.org/mailman/listinfo/sidr
