On 15/11/2010, at 5:31 PM, Sandra Murphy wrote: >> >> I last ran this around June 2007 btw. > > Can you share the results? > > --Sandy >
Nothing formal written up sandy. And, Randy is right. the size of the datapool is not my 'bang all resources in one roa' but the total number of discrete routes in the BGP cloud: if you fold them all into one uber ROA you fate share the age date of the resources. And, not all resources lie under one cert chain anyway so you wind up unable to combine them into one cert. So, the experience of a 30k tree won't inform the 300,000 tree, or the million-tree. Not enough. The churn rate is a function of several things, every CRL is updated, as is a manifest, each cycle time. So, there are many non-ROA objects changing. No manifests or Roas in my testbed. But, on the plus side, rsync is pretty quick. -G _______________________________________________ sidr mailing list [email protected] https://www.ietf.org/mailman/listinfo/sidr
