Hi, Aijun:
Thank you again for your interest in our draft! Hope the following answer is helpful. 1) How you deal with the congestion links in the intermediate nodes? Currently, our mechanism focuses on making adjustments proactively on the remote, so that the congestion could be avoid in advance. The adjustment is based on the link quality of the second link. For instance, the load balance on R3 is based on the link quality of R2->R4 in path R3->R2->R4, or the quality of R1->R4 in path R3->R1->R4. As for adjusting from nodes that are further away from the congestion point, this implementation will be more complex. This could be discussed in the future. 2) For the flow based congestion adjustment, how can avoid the redirection of the elephant flow to other links doesn’t lead to new congestion on the such links? When redirecting the flow, the selection of links is based on their remaining bandwidth. There is a mechanism that will periodically update the current link bandwidth to the devices. And if all the bandwidths are inadequate, the packet-based mode will be utilized, another mode in our mechanism, which will be soon released. 3) Even for the link based congestion adjustment, how can avoid the oscillation of link congestion status when the weights are adjusted? Therefore, we have three congestion adjustment mode, and we will soon release an updated version with detailed explanations. 1) Packet-based mode: the traffic causing congestion will be forwarded packet by packet to different links with lower loads. In this scenario, NP is needed to preserve the order of packets. 2) Flow-based mode: the corresponding flow will be redirected to the ECMP links with lower loads. 3) Link-based mode: in this mode the forwarding weights of the corresponding ECMP links will be adjusted gradually, instead of be directly changed to a very small value. This will avoid the oscillation. Please let me know if you have any other questions! -------------------------------------------------------------------------------- Best regards, 叶佳铭/Jiaming Ye中国移动通信有限公司研究院/China Mobile Research Institute Email: yejiam...@chinamobile.com ----邮件原文----发件人：Aijun Wang <wangai...@tsinghua.org.cn>收件人："叶佳铭" <yejiam...@chinamobile.com>,"39yingzhen.ietf39" <yingzhen.i...@gmail.com>抄 送: rtgwg <rtgwg@ietf.org>,rtgwg-chairs <rtgwg-cha...@ietf.org>发送时间：2024-07-05 11:28:16主题：Discussions on draft-cheng-rtgwg-adaptive-routing-framework Hi, Jiaming: After reading your documents, I have the following questions: 1) How you deal with the congestion links in the intermediate nodes? 2) For the flow based congestion adjustment, how can avoid the redirection of the elephant flow to other links doesn’t lead to new congestion on the such links? 3) Even for the link based congestion adjustment, how can avoid the oscillation of link congestion status when the weights are adjusted? Best Regards Aijun Wang China Telecom 发件人: forwardingalgori...@ietf.org [mailto:forwardingalgori...@ietf.org] 代表 叶佳铭发送时间: 2024年7月5日 10:36收件人: yingzhen.ietf <yingzhen.i...@gmail.com>抄送: rtgwg <rtgwg@ietf.org> rtgwg-chairs <rtgwg-cha...@ietf.org>主题: 回复:IETF 120 RTGWG Slot Requests Hi Chairs, I39d like to request a 10 mins slot for a new draft. Please help to arrange it into the agenda of the rtgwg session. The information is as follow: Presenter: Changwang Lin/Jiaming Ye Name: draft-cheng-rtgwg-adaptive-routing-framework Revision: 00 Title: Adaptive Routing Framework Date: 2024-07-04 Group: Individual SubmissionPages: 13URL: https://www.ietf.org/archive/id/draft-cheng-rtgwg-adaptive-routing-framework-00.txtStatus: https://datatracker.ietf.org/doc/draft-cheng-rtgwg-adaptive-routing-framework/HTMLized: https://datatracker.ietf.org/doc/html/draft-cheng-rtgwg-adaptive-routing-framework Thank you very much. -------------------------------------------------------------------------------- Best regards, 叶佳铭/Jiaming Ye 中国移动通信有限公司研究院/China Mobile Research Institute Email: yejiam...@chinamobile.com -----原始邮件-----发件人:yingzhen.i...@gmail.com发送时间: 2024-06-26 02:26:56 收件人：RTGWG,rtgwg-chairs抄送：主题：IETF 120 RTGWG Slot Requests Hi, The draft agenda for IETF 120 has been posted: IETF 120 Meeting Agenda The RTGWG session is scheduled on Monday Session II 13:00-15:00, July 22, 2024. Please send slot requests to rtgwg-cha...@ietf.org before the end of the day Wednesday July 10th. Please include draft name and link, presenter, desired slot length including Q&A. Please note that we only have one 2-hour session, and the agenda will be built based on all the requests we receive. Also having a discussion on the RTGWG mailing list is a prerequisite for a draft presentation in the WG session. If you need any help please reach out to the chairs. Thanks, Yingzhen
_______________________________________________ rtgwg mailing list -- rtgwg@ietf.org To unsubscribe send an email to rtgwg-le...@ietf.org
