Aijun, Any network slicing proposal is going to require active management of the underlay network in response to changes in the requirements of the overlay VPNs. The text you quote is designed to address a different issue which none of the other network slicing proposals address, viz, in a heterogeneous mix of regular and enhanced VPNs, how do you ensure that regular VPNs, that know nothing about the enhanced VPNs, don’t use the resources that are supposed to be used only by enhanced VPNs? See also: https://tools.ietf.org/html/draft-dong-spring-sr-for-enhanced-vpn-07.
Yours Irrespectively, John Juniper Business Use Only From: Aijun Wang <wangai...@tsinghua.org.cn> Sent: Thursday, March 26, 2020 9:37 PM To: John E Drake <jdr...@juniper.net> Cc: Joel M. Halpern <j...@joelhalpern.com>; xie...@chinatelecom.cn; lsr <lsr@ietf.org> Subject: Re: [Lsr] Using IS-IS Multi-Topology (MT) for Segment Routing based Virtual Transport Network [External Email. Be cautious of content] Hi, John: In your proposal, there is the following text “ the network controller SHOULD ensure that the IGP and TE metrics for these resources is higher than the metrics for the underlay network resources allocated to non-enhanced VPNs.” Considering these resources will span across the network and be changed upon the slicing requirements , will such arrangement make the metric allocation within the network a mess? If the above statement can’t be met, how you ensure the traffic that pass the P router use the dedicated resource(for example, bandwidth)? Aijun Wang China Telecom On Mar 26, 2020, at 23:31, John E Drake <jdrake=40juniper....@dmarc.ietf.org<mailto:jdrake=40juniper....@dmarc.ietf.org>> wrote: Hi, As Joel notes, it is true that enhanced VPNs require the use of specific underlay network resources, either dedicated or shared, but the this needs to be done without installing overlay VPN awareness in the P routers, which is inherently unscalable and operationally complex. Also, since VPNs span multiple ASes, putting overlay VPN state in an IGP doesn't work. Please see: https://tools.ietf.org/html/draft-drake-bess-enhanced-vpn-02 Yours Irrespectively, John Juniper Business Use Only -----Original Message----- From: Lsr <lsr-boun...@ietf.org<mailto:lsr-boun...@ietf.org>> On Behalf Of Joel M. Halpern Sent: Thursday, March 26, 2020 9:36 AM To: xie...@chinatelecom.cn<mailto:xie...@chinatelecom.cn>; lsr <lsr@ietf.org<mailto:lsr@ietf.org>> Subject: Re: [Lsr] Using IS-IS Multi-Topology (MT) for Segment Routing based Virtual Transport Network [External Email. Be cautious of content] In once sense, the statement is inherently true. A VPN technology without underlay support would seem to have significant difficulty in consistently meeting an SLA. Having said that much, the rest does not seem to follow. Yours, Joel
_______________________________________________ Lsr mailing list Lsr@ietf.org https://www.ietf.org/mailman/listinfo/lsr
