I have reviewed this document as part of the Operational directorate's ongoing
effort to review all IETF documents being processed by the IESG.  These comments 
were written primarily for the benefit of the operational area directors.  
Document editors and WG chairs should treat these comments just like any other 
last call comments.

I was not assigned this ops-dir review, but was assigned a secdir review, so I had already reviewed this document and thought I would comment.

I think this document is Ready for Publication, with one OPS&Mgmt concern.

Section 7 constrains ftp duplication to managed networks, and discusses the need for knowledge of network capacity, and observing the losses in each stream separately to fine-tune the duplication process. But there is no recommendation about what information should be made accessible by midpoints, or how such information should be made available. 

It happens often that the real-time community rejects requests for standardization of management, because there are many different environments in which RT transport occurs. I am not recommending that this document try to standardize a particular management protocol or data model. (Even though I think such standardization would be a good thing to do.)

Section 7 contrains this spec to managed networks, but section 7 is about congestion control. I am not sure I understand the proposal to make capacity and congestion information observable by the endpoint so it can fine-tune the duplication process. Is this a recommendation to use layer 7 management protocols, such as snmp or ipfix, or to use layer 4 congestion management protocols such as ECN/PCN? 

I think there should be a discussion of the information model for information that should be made available by the elements in the managed network, and possible mechanisms for delivering that information to the endpoint, to support ftp duplication.

David Harrington

ietfdbh at comcast.net