Hi KwangKoog Lee, I do hear your concern about the transport and congestion control. Different DSCPs may make sense for the I2RS protocol; the control part with write operations makes sense to have at a high priority. However, getting the RIB updates as an information stream may not want to have the same priority. Congestion control is likely to be needed to handle congested situations that may occur. For instance, consider how I2RS and BGP should interact in terms of QoS.
I do have concerns about the head-of-line blocking that happens with TCP given congestion, but I think there are both other transports and that multiple transport channels can be used. Thanks, Alia On Thu, Aug 1, 2013 at 4:24 AM, KwangKoog Lee <[email protected]>wrote: > I also support WG adoption and agree with other members' comments. > Also, I have a small comment about the Sec. 6.1. > > "Whatever transport is used for the data exchange, it must also support > suitable congestion control mechanisms." > > Instead of this sentence, I think it needs to mention that I2RS protocol > messages must be treated by the highest proirity even if the control > network is under a congestion situation so that the communication is very > highly reliable between I2RS client and I2RS agent. > I think congestion control is little bit risky to I2RS operation because > those messages have very impotant information and also problem > statement document already mentions that responsiveness and reliability of > i2rs is important. > Thanks. > > Sincerely, > Kwnag-koog Lee (Ph.D) > KT (Korea Telecom) > > > On Wed, Jul 31, 2013 at 8:14 PM, Carlos Pignataro (cpignata) < > [email protected]> wrote: > >> >> On Jul 31, 2013, at 12:57 PM, Joe Marcus Clarke <[email protected]> >> wrote: >> >> Section 2: >> >> read scope: ... >> >> write scope: ... >> >> JMC: Should there be an event or notification scope in addition to read >> and write? >> >> >> +1 to this point. I think that it should also include some words about >> pub/sub. >> >> Section 6.6 covers some of this though, so perhaps a forward reference. >> >> Thanks, >> >> -- Carlos. >> >> _______________________________________________ >> i2rs mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/i2rs >> >> >
_______________________________________________ i2rs mailing list [email protected] https://www.ietf.org/mailman/listinfo/i2rs
