[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: hard questions: request routing



Title: RE: hard questions: request routing


> Consider the graph of content nodes.  For each metric, there is
> an optimal tree embedded in that graph. But, there is no guarantee
> that there is one embedded tree that contains the optimal
> tree for every metric.  That's not an indictment of the scheme,
> just a cautionary note.
>

hilarie,

that is not really true, you can traverse a tree based on any single or set of metrics.
a simple example, is just to assign weights to branches in the tree.


> It is interesting to task a single system with the responsibility for
> creating the content tree for each of its content sets.  Still, that
> seems like a lot of tree distributions transactions, potentially.
>
> Hilarie
>

that could do, but my approach do not require that. The size of trees per content are
very small. Encoding techniques could be used to represent the tree.
I could get all the representations that i need may be using less
that 128 bits per content depending on what coding methods I use.

I bet you that the other proposed metrics will cost much more than that.

Furthermore, the other proposed metrics will not do for the following
reasons:

1. Represent a hacked approach
     -  this will not fly in the IETF
2. The metric specify the redirection tree as opposed of tracking content
   - good luck on agreement
   - try to specify the optimal set of metrics.
   - I will gurantee you that the optimal set of metrics will break the black box concept.

3. Changes might be needed in the DNS space.


abbie