Hi Qiao, Thanks for your review. See my quick response inline.
Thanks, Jensen On Tue, Feb 8, 2022 at 9:26 PM Qiao Xiang <[email protected]> wrote: > Dear all, > > The following are some comments on draft-zhang-alto-oam-yang-01: > > 1. Title: Yang -> YANG > Good point. > > 2. Introduction: "The basic propose" -> "The basic purpose" > > 3. Section 4: "Table 2: Support configuration for information resource > generation algorithms" > Later in 4.1: "This document does not define any data model related to > specific implementation, including: > * Specific algorithms for ALTO information resource generation." > And later in 5.3.1., an example of information resource creation > algorithm is given. > > This is confusing. Does the ALTO OAM data model support or not support > algorithm implementation? > The data model does not support algorithm implementation. But it allows people to define API for specific algorithm implementation. > What are the differences between "configuration for information resource > generation algorithm", "information resource generation algorithm" and > "information resource creation algorithm"? > We will make the text consistent. > > > > 4. Section 5.3 " For some resource-type, the parameter of the intent > interface MUST also include the a dependency node containing the > resource-id of the dependent ALTO information resources (See Section 9.1.5 > of [RFC7285])." > > This is vague. What is "some resource-type"? Is it an information > source that has "uses" attribute? Need to make it clear. > Good point. Yes, we are talking about the information resource that has "uses" attribute. Will make it clear. > > 5. Section 5.3 "One categories" -> "One category" > "The other categories" -> "The other category" > "altorithms" -> "algorithms" > > 6. Section 5.3 "Except for the ird resource, all the other types of > resource entries have augmented algorithm node." > -> "an augmented algorithm node"? or "may have an augmented algorithm > node"? > That is a "MUST" statement. Will make it clear. > > 7. Section 5.4 "To use the reactive update, the reactive attribute MUST > be set true. To use the proactive update, the poll-interval attribute MUST > be greater than zero." ... > "If reactive is false or poll-interval is zero, the ALTO server will > not update the data source." > > What if reactive attribute is set to true, and the poll-interval > attribute is also set to be greater than zero? > It will not append. The "(update-policy)" node is a "choice" node. It means that either "(reactive)" case or "(proactive)" case can be present. If both of them are present, the message violates the syntax. > > 8. Section 5.4.1 "Yang" -> "YANG" > > 9. Section 5.4.2: what is the purpose of defining a Prometheus data > source? What about other time-series database, e.g., InfluxDB? > Thanks for the suggestion. This part is not complete. But this document will not define data models for all the potential data sources. This section will only cover examples for some popular data sources based on our experience. But which one should be put on this section is an open topic so far. > > > Best > Qiao > -- > Qiao Xiang > Professor, Xiamen University > _______________________________________________ > alto mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/alto >
_______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
