Review request
draft-ietf-alto-performance-metrics

This document extends the ALTO protocol (RFC7285) with additional costs metrics to be conveyed via an ALTO service. There have been discussions in the ALTO WG whether such performance metrics are reasonable as costs to convey via a service such as ALTO. Potential issues could be (in no particular order): clarity of semantics (e.g. what does it mean if the RTT between two endpoints is x?, is this the average over the last 24h?, last year?, etc.), introduction of feedback loops (e.g. endpoints could select different peers or caches to download from based on metrics that change rather quickly and such a selection could effect the metric result itself in the very near future), privacy considerations (certain metrics might reveal too much about an endpoint, e.g. a high bandwidth Internet connection might reveal something about the endpoint owner's financial situation).

For all the metrics introduced in the document, we would thus appreciate an outside-expert-view: does it make sense to convey such a metric via the ALTO protocol?

Many thanks in advance,
Jan & Vijay (ALTO chairs)