Dear Piotr, all, I am reading up on draft-dulinski-alto-inter-problem-statement-02, and here is some initial comment.
My high-level feedback is that this is a very valuable piece of work. In particular, Section 2 gives a good starting point to discuss related basic concepts; Figure 1 starts to show the basic structure; and the core section (Section 3) gives good examples to motivate the benefit for inter-ALTO communications. Personally, I see inter-ALTO as a natural direction for ALTO extension (e.g., a potential for a WG work item), and this document can be a good base. At a high level, to form such a base, I see some additions to this document where the following issues are discussed: - Additional analysis on why existing mechanisms are not sufficient. For example, do the authors conclude that RFC7285 does not provide enough capabilities to support inter-ALTO? If so, which features are missing, according to the use cases? Have the authors evaluated other candidate mechanisms (e.g., applying BGP would be a first consideration)? - It may help if the framework provides some discussions on information aggregation. For example, BGP uses import and export to implement policy routing. I see that similar mechanisms may be needed in inter-ALTO as well--the draft discussed about polices already, and what I am asking for is more elaboration. More details: - The partitioned use case is a bit terse. It helps to elaborate more. Minors: - The draft has "Due to various reasons" at multiple places. At some places, examples are given and at others no. It helps to be specific. - Abstract: "ALTO servers are limited by the fact that it may not be possible for an ALTO server to compute costs for source/destination pairs correctly if a source and/or a destination is outside ..." The word correct is strong. Correct implies there is an objective value, and in some context, the cost can be "subjective". Hence, a weaker word can be helpful here. - Introduction: "Topology- and policy-related information may be supplied through ALTO in a proactive or in a reactive way." I found this usage of proactive and reactive quite interesting. - Section 3: "It can be shown that without additional information on ..." Can you give an example to support it can be shown? - Section 3.1: "Moreover, there is a significant disproportion between availability of information on upstream and downstream paths." How about giving definitions on upstream vs downstream, as some readers may not be familiar with the terms? - Section 3.1: "To mitigate this situation, the inter-ALTO communication framework may be used to exchange information on downstream paths between two interested parties. " My understanding is that this may require a recursive process, all the way to the final local administrative domain, right? Cheers, Richard
_______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
