L2 adjacency would be ok. BR, John -----Original Message----- From: Pablo Camarillo (pcamaril) <[email protected]> Sent: Monday, June 14, 2021 2:25 PM To: Kaippallimalil John <[email protected]>; Jeffrey (Zhaohui) Zhang <[email protected]> Cc: [email protected] Subject: RE: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11
Hi John, Many thanks for detailed your answer. Do you think it would be fair to use "L2 Adjacency" to refer to it in the draft? Thanks, Pablo. -----Original Message----- From: Kaippallimalil John <[email protected]> Sent: lunes, 14 de junio de 2021 20:57 To: Pablo Camarillo (pcamaril) <[email protected]>; Jeffrey (Zhaohui) Zhang <[email protected]> Cc: [email protected] Subject: RE: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 Hi Pablo, There is no IP adjacency between UE - gNB. It is a data link protocol and the user plane stack for adjacency is as in figure below (see TS 38.300. 6.6 L2 Data Flow). IP flow is packaged in SDAP SDU, then in PDCP layer with RoHC, security & RLC with seg/ARQ. This RLC SDU is carried in a MAC transport block /multiplexed with other RLC SDUs. Addressing, mapping and (de)multiplexing are handled by these layers and related control plane. UE gNB [SDAP] <----------------------> [SDAP] [PDCP] <----------------------> [PDCP] [RLC] <----------------------> [RLC] [PHY] <----------------------> [PHY] BR, John -----Original Message----- From: Pablo Camarillo (pcamaril) <[email protected]> Sent: Monday, June 14, 2021 12:15 PM To: Kaippallimalil John <[email protected]>; Jeffrey (Zhaohui) Zhang <[email protected]> Cc: [email protected] Subject: RE: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 John, Indeed, the point was on the downlink, whether there is an IP adjacency in between the gNB and the UE. Many thanks for your answer. If it is not an IP adjacency, what is the type of the adjacency between the gNB and the UE? Thanks, Pablo. -----Original Message----- From: dmm <[email protected]> On Behalf Of Kaippallimalil John Sent: viernes, 11 de junio de 2021 18:43 To: Jeffrey (Zhaohui) Zhang <[email protected]>; Pablo Camarillo (pcamaril) <[email protected]> Cc: [email protected] Subject: Re: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 Hi Pablo, Jeffrey, If the question is on IP adjacency in 3GPP, it is between the UE and UPF (PSA). The gNB switches radio bearer (DRB) to IP flow (GTP-U) based on instructions from SMF (SMF --> AMF --> gNB). The first router for the user's IP flow is the UPF. An easy to read reference is RFC 6459 (refers to 4G, but the concepts in this case are similar) MME (instead-of SMF) provisions eNB (i-o gNB) to switch radio bearer to GTP-U bearer, and IP adjacency is between UE - PGW (i-o UPF). Or 3GPP TS 23.501, 5.8 (and other sections) provide more details. BR, John -----Original Message----- From: dmm <[email protected]> On Behalf Of Jeffrey (Zhaohui) Zhang Sent: Friday, June 11, 2021 11:15 AM To: Pablo Camarillo (pcamaril) <[email protected]>; Jeffrey (Zhaohui) Zhang <[email protected]> Cc: [email protected] Subject: Re: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 Hi Pablo, I have to say that we completely disagree on this one. Someone else will have to chime in if we want to get a conclusion 😊 Please see zzh> below. -----Original Message----- From: dmm <[email protected]> On Behalf Of Pablo Camarillo (pcamaril) Sent: Friday, June 11, 2021 10:16 AM To: Jeffrey (Zhaohui) Zhang <[email protected]> Cc: [email protected] Subject: Re: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 [External Email. Be cautious of content] Jeffrey, -----Original Message----- From: Jeffrey (Zhaohui) Zhang <[email protected]> Sent: viernes, 4 de junio de 2021 22:06 To: Pablo Camarillo (pcamaril) <[email protected]> Cc: [email protected] Subject: RE: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 Hi Pablo, The IP adjacency to the UE is from the UPF, not from the gNB. While gNB is an IP router, it is at the underlay. This is similar to the following situation: Host1 ---- EVPN PE1 --------- EVPN PE2 ---- Host2 Host1 and Host2 have IP adjacency, even though EVPN PE1 and PE2 are IP routers. Host1 and EVPN PE1 are not IP adjacent. [PC] Your examples assume PDU Session Type is L2. In which case indeed there is no IP Adjacency (there is only a L2 adjacency). [PC] If the PDU Session Type is either IPv4 or IPv6, then I believe the interface at the gNB facing towards the UE is IP enabled, and therefore there exists an IP adjacency. Similarly, in your example above there would be an IP adjacency in between Host1 and PE; as both interfaces (Host1 towards PE1, and PE1 from Host1) are IP based. Zzh> In the UE---gNB---UPF scenario, even for IPv4/6 PDU session, there is no IP adjacency between UE and gNB. Zzh> In the EVPN scenario above, there is no IP between Host1 and PE1 or between Host2 and PE2 either. It's just Ethernet. IP is between Host1 and Host2. [PC] Also, for IPv4 or IPv6, Host1 and Host2 are not adjacent as per RFC1812, since the PE processes/forward the IP packet. [PC] Please let me know whether you agree. It would be great if you can send me some pointer that I could read if you still think Im wrong. Zzh> For the EVPN example, EVPN is emulating an Ethernet network - no different from Host1 and Host2 being connected via two Ethernet switches. Zzh> The UE case is similar, though it is not Ethernet below the IP, but a lower layer consisting of one segment being PDCP (3G case at least) and the other segment being GTP-U. I could only find a 3G picture here: https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.researchgate.net%2Ffigure%2F3GPP-protocol-stack-see-9_fig1_224357167&data=04%7C01%7Cjohn.kaippallimalil%40futurewei.com%7Ce49b237ceede443f7c0408d92f6a2d34%7C0fee8ff2a3b240189c753a1d5591fedc%7C1%7C1%7C637592955355021142%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=hniWKZOX2j7cBTy2GAe0Iz2XIoA%2BzmKj0aW8u6iIwis%3D&reserved=0. Zzh> Jeffrey [PC] Many thanks. Jeffrey -----Original Message----- From: Pablo Camarillo (pcamaril) <[email protected]> Sent: Friday, June 4, 2021 12:52 PM To: Jeffrey (Zhaohui) Zhang <[email protected]> Cc: [email protected] Subject: RE: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 [External Email. Be cautious of content] Hi Jeffrey, Why do you say there is no IP adjacency from gNB to UE? In the downlink, the gNB receives an IP packet from the UPF destined to itself. The gNB decapsulates such packet, and forwards the inner-exposed packet to the UE. Assuming the inner-exposed packet is IPv4, in such case the IPv4 Destination Address is the one of the UE. RFC1812: "Adjacent - reachable without going through any IP routers" As far as I know, there is no IP router in between the gNB and the UE. Can you please clarify? Cheers, Pablo. -----Original Message----- From: Jeffrey (Zhaohui) Zhang <[email protected]> Sent: martes, 1 de junio de 2021 16:04 To: Pablo Camarillo (pcamaril) <[email protected]>; [email protected] Subject: RE: [DMM] [spring] note: WGLC on draft-ietf-dmm-srv6-mobile-uplane-11 Hi Pablo, Let me pull this up: ----------- Zzh4> gNB is certainly an IP device but its IP adjacency is in the "underlay" (transport network) not at the "overlay" (towards the UE). [PC5] The End.DX4 behavior (or the others) are not limited or restricted by "underlay" vs "overlay" IP adjacencies. The behavior is the same: remove the encaps, forward on a particular IP adjacency (regardless of its type). Zzh4> In the wireline/IETF VPN case, there is PE-CE IP adjacency. Traffic from CE is forwarded based on IP lookup in the VRF, whether the traffic to the CE requires IP lookup or not (i.e. whether per-CE or per-VPN label is used, or whether END.DT4/6/X is used). Zzh4> In case of gNB, UE-gNB is not IP adjacency and traffic to/from a UE does not have IP lookup (based on the inner IP header) at the gNB. --------- As I pointed out in zzh4>, the point is that there is *no IP adjacency* between UE and gNB. Practically, a device can implement the END.DX4/6 behavior to forward IP traffic over non-IP adjacencies. It's just that RFC8986 specifically calls out IP adjacencies. You may want to point it out and see if SR folks have any concerns. Jeffrey Juniper Business Use Only Juniper Business Use Only _______________________________________________ dmm mailing list [email protected] https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Furldefense.com%2Fv3%2F__https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Fdmm__%3B!!NEt6yMaO-gk!UbiidCykEZnbLyjo8UMPaSlhktUmVxQv7QR3n6XJAj5q053iHL7Vj9-HloD_jsGX%24&data=04%7C01%7Cjohn.kaippallimalil%40futurewei.com%7Ce49b237ceede443f7c0408d92f6a2d34%7C0fee8ff2a3b240189c753a1d5591fedc%7C1%7C1%7C637592955355021142%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=SFFuux5asd1dsRdHRjyVcqfwwGtFZMfRxZIJf%2B3SYFA%3D&reserved=0 Juniper Business Use Only Juniper Business Use Only _______________________________________________ dmm mailing list [email protected] https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Fdmm&data=04%7C01%7Cjohn.kaippallimalil%40futurewei.com%7Ce49b237ceede443f7c0408d92f6a2d34%7C0fee8ff2a3b240189c753a1d5591fedc%7C1%7C1%7C637592955355031097%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=uzxLYduPhKkEVqWbbZSRS59A1TLOwODBxRQVLwJVCLA%3D&reserved=0 _______________________________________________ dmm mailing list [email protected] https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.ietf.org%2Fmailman%2Flistinfo%2Fdmm&data=04%7C01%7Cjohn.kaippallimalil%40futurewei.com%7Ce49b237ceede443f7c0408d92f6a2d34%7C0fee8ff2a3b240189c753a1d5591fedc%7C1%7C1%7C637592955355031097%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=uzxLYduPhKkEVqWbbZSRS59A1TLOwODBxRQVLwJVCLA%3D&reserved=0 _______________________________________________ dmm mailing list [email protected] https://www.ietf.org/mailman/listinfo/dmm
