I reviewed the PIE draft and have some comments: http://tools.ietf.org/html/draft-ietf-aqm-pie-00
The basic algorithm is very clear and I think it could easily be implemented by someone else based on this draft alone. Also, the background information in the draft is pretty clear, including the linkage with other documents like the docsis-pie I-D. In general, I think the working group should try to quickly complete reviews and get submit it to the ADs for publication. Section 3 on the design goals would be good to link back to the soon-to-be RFC on AQM recommendations from this working group. As I noted in the similar part of the CoDel draft too, I think that the AQM recommendations grew out of these design goals rather than the other way around, but it would be good to tie the working group items together that are closely related. Plus I would guess that the IESG or some other reviewer might ask how PIE relates to those recommendations previously published. It might be helpful to encourage specific future work by including a section near the end on "Future Research". I certainly had a number of thoughts in this regard when reading section 5 on the enhancements. For instance, there could be other auto-tuning algorithms, sensitivity of the parameters could be looked at further (especially with regard to the burst allowance), etc. Note that the I-D template could in-use is missing some important things, like an "intended status" header, which will eventually need to be fixed so it can be submitted for publication to the ADs. Minor editorial comment: there's an odd line break or two in section 6, 3rd paragraph, when trying to say "beta = 2.5". Editorial comment: "time-stamped" should be "timestamps" in section 6 in the 2nd-to-last paragraph on page 12. -- Wes Eddy MTI Systems _______________________________________________ aqm mailing list [email protected] https://www.ietf.org/mailman/listinfo/aqm
