I agree with Larry that both L2 over L3 and L3 over L3 should be considered and be included in the charter. However, the term “layer agnostic” seems not concrete enough and hence may cause some confusion. For example, it could be interpreted as “underlying network agnostic” (e.g., L2 over L2, L2 over nickname…)
Best regards, Xiaohu 发件人: [email protected] [mailto:[email protected]] 代表 Larry Kreeger 发送时间: 2012年4月21日 6:48 收件人: Bocci, Matthew (Matthew); John E Drake; David Black; Stewart Bryant; [email protected] 抄送: Yakov Rekhter; Nitin Bahadur; [email protected]; [email protected] 主题: Re: [nvo3] NVO3 charter 1530UK 12April I support the sentence below from Matthew which reads “ The WG will determine whether an IP, and/or an emulated Ethernet service is needed to satisfy the needs of the typical data centre." because I just noticed that the latest charter is missing the following sentence that was in the charter Benson sent out after the BOF which read, “This approach should provide an Ethernet service. It may provide an IP service; an important goal is to develop a "layer agnostic" framework and architecture meeting data center requirements.". I agreed that keeping the architecture “layer agnostic” is a very good idea to leave room for how DC networking may evolve in the future. A Layer 2 service may be needed now (at least for some applications), but layer 3 service may suffice for many applications and give benefits in terms of scale and/or simplification and/or interoperability. Lets not decide at the time of charter. - Larry On 4/20/12 4:31 AM, "Bocci, Matthew (Matthew)" <[email protected]> wrote: John, How about the following changes to the paragraph below? The intention is to clearly phrase the charter in terms of requirements work that the WG will do, while at the same time scope the requirements we will work on, rather than predicate any ultimate solutions. "NVO3 will consider approaches to multi-tenancy that use an encapsulation header that resides at or above the network layer, rather than relying on traditional L2 isolation mechanisms (e.g., VLANs) to support multi-tenancy. The WG will determine whether an IP, and/or an emulated Ethernet service is needed to satisfy the needs of the typical data centre." Regards, Matthew On 19/04/2012 16:24, "John E Drake" <[email protected]> wrote: David, I don’t think you have the authority to change the words in the charter, which reads: “It will also address requirements driven by cloud computing services and data centers as they apply to Layer-2 VPN services." Further, as I have told you and Thomas multiple times, a data center operator is not required to provide a certificate indicating they are a ‘provider’ before they are allowed to deploy L3/L2 VPN technology, and many large enterprise networks consider themselves to be service providers in their own right. Thanks, John Sent from my iPhone From: [email protected] [mailto:[email protected]] Sent: Thursday, April 19, 2012 7:17 AM To: John E Drake; [email protected]; [email protected]; [email protected] Cc: Yakov Rekhter; Nitin Bahadur; [email protected]; [email protected] Subject: RE: [nvo3] NVO3 charter 1530UK 12April John, I think the potential “data center” overlap with L2VPN can be resolved by carefully understanding the first text extract from L2VPN charter. I read the second sentence as implying a couple of crucial words in [square brackets]: "The L2VPN working group is responsible for defining and specifying a limited number of solutions for supporting provider-provisioned Layer-2 Virtual Private Networks (L2VPNs). It will also address requirements driven by cloud computing services and data centers as they apply to [provider-provisioned] Layer-2 VPN services." We can split hairs over what “provider” means, but I believe the primary distinction in initial focus is data center infrastructure vs. network carrier (provider) provisioning and operation of the overlay (or VPN if one wants to use that term). The following definition from RFC 4664, Framework for Layer 2 Virtual Private Networks (L2VPNs), may also help in understanding the L2VPN WG’s scope: The term "provider provisioned VPNs" refers to Virtual Private Networks (VPNs) for which the Service Provider (SP) participates in management and provisioning of the VPN. In this context, many data centers of importance to nvo3 (e.g., enterprise data centers) are not operated by the Service Provider, as the term is used in this RFC 4664 definition. Thanks, --David From: [email protected] [mailto:[email protected]] On Behalf Of John E Drake Sent: Thursday, April 19, 2012 8:06 AM To: Bocci, Matthew (Matthew); Stewart Bryant; Kireeti Kompella Cc: Yakov Rekhter; Nitin Bahadur; [email protected]; [email protected] Subject: Re: [nvo3] NVO3 charter 1530UK 12April Matthew, Snipped, comment inline. Thanks, John Sent from my iPhone Why not just rephrase the paragraph so that it does not appear to prescribe protocol development, but rather scopes the solutions to those that the IETF traditionally deals with and that meet the requirements/gap analysis?: “NVO3 will consider an approach to multi-tenancy that uses a Layer 3 encapsulation rather than relying on traditional L2 isolation mechanisms (e.g., VLANs) to support multi-tenancy, and consistent with a requirements gathering and gap analysis exercise. The approach will provide an emulated Ethernet service capable of satisfying typical data center deployments.” [JD] I also have a problem with the last sentence as it sounds as though NV03 will be encroaching on the charter of the L2VPN WG, which reads, in part: "The L2VPN working group is responsible for defining and specifying a limited number of solutions for supporting provider-provisioned Layer-2 Virtual Private Networks (L2VPNs). It will also address requirements driven by cloud computing services and data centers as they apply to Layer-2 VPN services." And: "5. Ethernet VPN (E-VPN) - An enhanced Layer-2 service that emulates an Ethernet (V)LAN across a PSN. E-VPN supports load-sharing across multiple connections from a Layer-2 site to an L2VPN service. E-VPN is primarily targeted to support large-scale L2VPNs with resiliency requirements not satisfied by other L2VPN solutions." ________________________________ _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
_______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
