kind of distributed-computing system; if considering abstract say (TE) info from domain internal to parent PCE, the fundamental question might be what almost-exact information that are gonna be abstracted and also need to make sure by abstracting the global end-to-end optimal pach can still be achieved with no scalability concerns ...
Regards, Peng 2010/1/10 赵永利 <[email protected]>: > Hi Dan > Thank you very much for your answers. > Ok. The Parent PCE can determine a likely domain path based on its > knowledge of the inter-domain links. However, the domain path may not be > optimal because it does not know the detailed information in each domain. > Maybe there is a balance between the optimization and the complexity. Do you > think so? > I am always wondering about the optimal end-to-end path in the current > framework. > Another question. Can we abstract the detailed information in each > domain into some simple forms and provide them to the parent PCE? For > example, if we compute the shortest path according to the hop numbers. We > can gain the hop numbers between each border nodes pair in each domain by > the child PCE, and provide them to the parent PCE. Then the parent PCE can > gain the optimal the end-to-end path across multi-domains. > > Best regards. > > Yongli Zhao > > Beijing University of Posts and Telecommunications > Beijing, China > > > 2010/1/8 Daniel King <[email protected]> >> >> Hi Yongli, >> >> Thanks for reviewing the draft. In response to your questions: >> >> 1. The Parent PCE will determine the likely domain paths from the source >> to the destination. The Parent PCE will then send an edge-to-edge request to >> a Child PCE in the ingress domain (ingress-to-edge), and requests to a Child >> PCE in each candidate transmit domain (edge-to-edge) and finally a Child PCE >> in the egress domain (edge-to-egress). Once the responses are collated a >> suitable path can be identified and selected by Parent PCE. The Parent PCE >> will then send a response back to the Child PCE in the ingress domain. The >> Child PCE will then send an ERO response to the PCC that requested the >> inter-domain path. >> >> 2. H-PCE has been designed so the Parent PCE will not have the same >> detailed internal domain topology information of the Child PCE's it >> services. However the Parent PCE does know capabilities of the links that >> connect the domains together. The inter-domain link information would be >> reported by the Child PCE during the initial Child and Parent PCE >> information exchange. The Child PCE can learn about its inter-domain links >> via its domain IGP. During "step 9" of the example procedure the Parent PCE >> has the opportunity to apply network or policy constraints based on its >> knowledge of the inter-domain links selected for the candidate path. >> >> So in summary, one aim of H-PCE is to avoid the need for anyone one entity >> to have a detailed topology view of each transit domain. A Child PCE per >> domain, is used to compute a segment of the end-to-end path. The Child PCE >> can also replace its path segment information with a path-key [RFC5520] to >> further maintain confidentiality. >> >> On review of the step 9 procedure, we will update the text so the process >> and actual network topology view is clearer. >> >> Br, Dan. >> >> From: 赵永利 [mailto:[email protected]] >> Sent: 03 January 2010 09:45 >> To: [email protected] >> Cc: [email protected] >> Subject: question on draft-king-hierarchy-fwk-03.txt >> >> Hi Dan, >> I have two questions on the document >> "draft-king-hierarchy-fwk-03.txt". >> (1) In the procedure of section 5.6.2, three domain paths have been >> gained, and PCEs 1 to 4 have received the path compuation requests from the >> parent PCE. Does it mean that the parent PCE will send edge-to-edge path >> computation requests to all the PCEs along all the domain paths? >> >> (2) In step 9 of the same section, it seems not very clear about the >> compuation responses in the sentence of "PCE 5 correlates all the >> computation responses from each child PCE". Does it mean the detailed >> constraint-based path segments in the local domain? >> Thank you very much. >> >> Best regards. >> >> >> Yongli Zhao >> >> Beijing University of Posts and Telecommunications >> Beijing, China >> > > > _______________________________________________ > Pce mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/pce > > _______________________________________________ Pce mailing list [email protected] https://www.ietf.org/mailman/listinfo/pce
