Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions
On Jun 15, 2009, at 9:54 PM, Ryan, Jim wrote: - There was no standards body associated with the proposal. Clarification question on this -- what exactly does this statement mean? I ask because if you take that statement to the extreme, many things in OFED may not qualify. For example, the uDAPL software API is standardized, so I assume it's fine. The iWARP and IB hardware layers are standardized, but are there standardized software APIs? Specifically: the verbs software stack is not standardized by any standards body, is it? (what is the exact definition of standards body, anyway -- does consensus of OFED members count?) These may be naive questions (I don't follow these APIs at all), but is the RDS API standardized? Is SRP? Are the various tools that are shipped in OFED standardized? I'm sure we don't want to take all these things out of OFED, but some clarification of the rules might be useful. -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
RE: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions
Jeff, thanks, I'll try to keep this as brief as possible and offer more offline if anyone wants it: - The bylaws state the requirement for a standards body to be associated with a proposal. It contains some examples, including the IETF and the IBTA (and others) - I think it's fair to say there's some ambiguity around what a proposal would consist of. That's part of the discussion I've tried to get started in the past, unsuccessfully - The bylaws also say the Board can ID other such bodies as the need is perceived - My best understanding is the IBTA is developing a spec that addresses this area. Work has started, but I don't have a timeline for delivery. An obvious concern to me is having code submitted prior to a spec being agreed Jim -Original Message- From: ewg-boun...@lists.openfabrics.org [mailto:ewg-boun...@lists.openfabrics.org] On Behalf Of Jeff Squyres Sent: Tuesday, June 16, 2009 9:25 AM To: OpenFabrics EWG; OpenFabrics General Subject: Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions On Jun 15, 2009, at 9:54 PM, Ryan, Jim wrote: - There was no standards body associated with the proposal. Clarification question on this -- what exactly does this statement mean? I ask because if you take that statement to the extreme, many things in OFED may not qualify. For example, the uDAPL software API is standardized, so I assume it's fine. The iWARP and IB hardware layers are standardized, but are there standardized software APIs? Specifically: the verbs software stack is not standardized by any standards body, is it? (what is the exact definition of standards body, anyway -- does consensus of OFED members count?) These may be naive questions (I don't follow these APIs at all), but is the RDS API standardized? Is SRP? Are the various tools that are shipped in OFED standardized? I'm sure we don't want to take all these things out of OFED, but some clarification of the rules might be useful. -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions
Understood -- and I'm not debating these questions in light of RDMAoE/ IBoE/whateveroE. I'm asking how tightly these questions have been applied to all the other stacks in OFED. For example, RDS, SRP, and verbs itself. Are any of these published/standardized APIs other than the fact that they're in OFED? What does it mean to have a standard body associated with a proposal? Does associated mean that the proposal/API is a published standard in that standards body? Or some weaker definition? On Jun 16, 2009, at 12:37 PM, Ryan, Jim wrote: Jeff, thanks, I'll try to keep this as brief as possible and offer more offline if anyone wants it: - The bylaws state the requirement for a standards body to be associated with a proposal. It contains some examples, including the IETF and the IBTA (and others) - I think it's fair to say there's some ambiguity around what a proposal would consist of. That's part of the discussion I've tried to get started in the past, unsuccessfully - The bylaws also say the Board can ID other such bodies as the need is perceived - My best understanding is the IBTA is developing a spec that addresses this area. Work has started, but I don't have a timeline for delivery. An obvious concern to me is having code submitted prior to a spec being agreed Jim -Original Message- From: ewg-boun...@lists.openfabrics.org [mailto:ewg-boun...@lists.openfabrics.org ] On Behalf Of Jeff Squyres Sent: Tuesday, June 16, 2009 9:25 AM To: OpenFabrics EWG; OpenFabrics General Subject: Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions On Jun 15, 2009, at 9:54 PM, Ryan, Jim wrote: - There was no standards body associated with the proposal. Clarification question on this -- what exactly does this statement mean? I ask because if you take that statement to the extreme, many things in OFED may not qualify. For example, the uDAPL software API is standardized, so I assume it's fine. The iWARP and IB hardware layers are standardized, but are there standardized software APIs? Specifically: the verbs software stack is not standardized by any standards body, is it? (what is the exact definition of standards body, anyway -- does consensus of OFED members count?) These may be naive questions (I don't follow these APIs at all), but is the RDS API standardized? Is SRP? Are the various tools that are shipped in OFED standardized? I'm sure we don't want to take all these things out of OFED, but some clarification of the rules might be useful. -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
RE: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions
What does it mean to have a standard body associated with a proposal? Does associated mean that the proposal/API is a published standard in that standards body? Or some weaker definition? There's a significant difference between an API and a wire protocol, and even most wire protocols shouldn't be subject to any standards body. We need to be realistic here. ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
RE: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions
Verbs is a software implementation of the IBTA and iWarp verbs for Linux, although there may be some features that are a superset of the IBTA and iWarp specifications, and/or were implemented first in OFA and then submitted to IBTA for standardization. SRP is an ANSI T10 standard. IPoIB is an IETF standard. There is no standard for RDS that I know of, but I think it was OFA that asked Oracle to submit it. The real question is does OFA want to limit itself to accepting only code that is part of some standard ? As for kernel.org, they typically do not follow such a rule, they do what they think is right for Linux, sometimes the code is associated with a standard, sometimes it is not. I think that the ofa-general list pretty much has been following the kernel.org model. I think that in any case, OFA needs to have a consistent policy, and if we allow something that is not a standard for one member, it should be allowed for all members. my 2 cents. woody -Original Message- From: general-boun...@lists.openfabrics.org [mailto:general-boun...@lists.openfabrics.org] On Behalf Of Jeff Squyres Sent: Tuesday, June 16, 2009 11:24 AM To: Ryan, Jim Cc: OpenFabrics EWG; OpenFabrics General Subject: Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions Understood -- and I'm not debating these questions in light of RDMAoE/ IBoE/whateveroE. I'm asking how tightly these questions have been applied to all the other stacks in OFED. For example, RDS, SRP, and verbs itself. Are any of these published/standardized APIs other than the fact that they're in OFED? What does it mean to have a standard body associated with a proposal? Does associated mean that the proposal/API is a published standard in that standards body? Or some weaker definition? On Jun 16, 2009, at 12:37 PM, Ryan, Jim wrote: Jeff, thanks, I'll try to keep this as brief as possible and offer more offline if anyone wants it: - The bylaws state the requirement for a standards body to be associated with a proposal. It contains some examples, including the IETF and the IBTA (and others) - I think it's fair to say there's some ambiguity around what a proposal would consist of. That's part of the discussion I've tried to get started in the past, unsuccessfully - The bylaws also say the Board can ID other such bodies as the need is perceived - My best understanding is the IBTA is developing a spec that addresses this area. Work has started, but I don't have a timeline for delivery. An obvious concern to me is having code submitted prior to a spec being agreed Jim -Original Message- From: ewg-boun...@lists.openfabrics.org [mailto:ewg-boun...@lists.openfabrics.org ] On Behalf Of Jeff Squyres Sent: Tuesday, June 16, 2009 9:25 AM To: OpenFabrics EWG; OpenFabrics General Subject: Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions On Jun 15, 2009, at 9:54 PM, Ryan, Jim wrote: - There was no standards body associated with the proposal. Clarification question on this -- what exactly does this statement mean? I ask because if you take that statement to the extreme, many things in OFED may not qualify. For example, the uDAPL software API is standardized, so I assume it's fine. The iWARP and IB hardware layers are standardized, but are there standardized software APIs? Specifically: the verbs software stack is not standardized by any standards body, is it? (what is the exact definition of standards body, anyway -- does consensus of OFED members count?) These may be naive questions (I don't follow these APIs at all), but is the RDS API standardized? Is SRP? Are the various tools that are shipped in OFED standardized? I'm sure we don't want to take all these things out of OFED, but some clarification of the rules might be useful. -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg -- Jeff Squyres Cisco Systems ___ general mailing list gene...@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
RE: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions
Associated is the strong form of the range you suggest. OFA is not in the spec creation biz. We look to other orgs to create the spec that we implement to. Again, I'm trying to tell you what our bylaws tell us. Certainly they can be changed, and I'd never claim we followed them with 100% diligence Jim -Original Message- From: Jeff Squyres [mailto:jsquy...@cisco.com] Sent: Tuesday, June 16, 2009 11:24 AM To: Ryan, Jim Cc: OpenFabrics EWG; OpenFabrics General Subject: Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions Understood -- and I'm not debating these questions in light of RDMAoE/ IBoE/whateveroE. I'm asking how tightly these questions have been applied to all the other stacks in OFED. For example, RDS, SRP, and verbs itself. Are any of these published/standardized APIs other than the fact that they're in OFED? What does it mean to have a standard body associated with a proposal? Does associated mean that the proposal/API is a published standard in that standards body? Or some weaker definition? On Jun 16, 2009, at 12:37 PM, Ryan, Jim wrote: Jeff, thanks, I'll try to keep this as brief as possible and offer more offline if anyone wants it: - The bylaws state the requirement for a standards body to be associated with a proposal. It contains some examples, including the IETF and the IBTA (and others) - I think it's fair to say there's some ambiguity around what a proposal would consist of. That's part of the discussion I've tried to get started in the past, unsuccessfully - The bylaws also say the Board can ID other such bodies as the need is perceived - My best understanding is the IBTA is developing a spec that addresses this area. Work has started, but I don't have a timeline for delivery. An obvious concern to me is having code submitted prior to a spec being agreed Jim -Original Message- From: ewg-boun...@lists.openfabrics.org [mailto:ewg-boun...@lists.openfabrics.org ] On Behalf Of Jeff Squyres Sent: Tuesday, June 16, 2009 9:25 AM To: OpenFabrics EWG; OpenFabrics General Subject: Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions On Jun 15, 2009, at 9:54 PM, Ryan, Jim wrote: - There was no standards body associated with the proposal. Clarification question on this -- what exactly does this statement mean? I ask because if you take that statement to the extreme, many things in OFED may not qualify. For example, the uDAPL software API is standardized, so I assume it's fine. The iWARP and IB hardware layers are standardized, but are there standardized software APIs? Specifically: the verbs software stack is not standardized by any standards body, is it? (what is the exact definition of standards body, anyway -- does consensus of OFED members count?) These may be naive questions (I don't follow these APIs at all), but is the RDS API standardized? Is SRP? Are the various tools that are shipped in OFED standardized? I'm sure we don't want to take all these things out of OFED, but some clarification of the rules might be useful. -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions
On Jun 16, 2009, at 2:42 PM, Woodruff, Robert J wrote: I think that in any case, OFA needs to have a consistent policy, and if we allow something that is not a standard for one member, it should be allowed for all members. Agreed; I think that this is my central point -- thanks for saying it succinctly! Regardless of whether OF asked Oracle to submit RDS or not, it's not associated with any standard (I'm not picking on RDS; I'm picking on the OF rules and selectively applying them). Therefore, either the bylaws are wrong of OF/EWG is wrong. I see several possibilities: 1. EWG needs to come into compliance with the OF bylaws 2. The bylaws need to be changed to match what EWG does (has done) 3. Throw the bylaws out (rules are useless if they don't guide what you do) 4. Some combination of 1 and 2 My $0.02. I leave it to others to figure out what to do. -- Jeff Squyres Cisco Systems ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet-- some procedural questions
I think that in any case, OFA needs to have a consistent policy, and if we allow something that is not a standard for one member, it should be allowed for all members. Agreed; I think that this is my central point -- thanks for saying it succinctly! Regardless of whether OF asked Oracle to submit RDS or not, it's not associated with any standard (I'm not picking on RDS; I'm picking on the OF rules and selectively applying them). Therefore, either the bylaws are wrong of OF/EWG is wrong. The bylaw in question seems pretty silly given the lack of control or involvement that OFA has with Linux kernel development. Given that RDS is not standardized at the API or wire level and given that RDS is included in the Linux kernel, what options does the OFA have for enforcing its bylaws? Removing RDS from OFED? Once OFED has *fewer* features than the standard kernel it becomes pretty pointless; maybe the logical conclusion is that OFA should get out of the distribution business (my feelings about ending OFED are well-known I think). The same applies to IBoE; if (and that really is if since I don't think a conclusion about merging IBoE support into the kernel has been reached) IBoE goes into the kernel but OFED can't or won't distribute it, then the relevance of OFED becomes marginal. - R. ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
Re: [ofa-general] RE: [ewg] [PATCH 0/9] RDMAoE - RDMA over Ethernet -- some procedural questions
On Mon, Jun 15, 2009 at 06:54:49PM -0700, Ryan, Jim wrote: Recall the bylaws of OFA requires that any ULPs that OFA supports will be produced by some recognized standards organization. No such organization was known to be associated with the Mellanox proposal. Is this being submitted as something to be included with OFA/Linux or is this just a technology demonstration? Personally, I don't think the proliferation of new networking protocols is something to be encouraged. If this is really what people want it should be IBoE and just like FCoE all the issues regarding interworking with actual IB networks must be delt with. Jason ___ ewg mailing list ewg@lists.openfabrics.org http://lists.openfabrics.org/cgi-bin/mailman/listinfo/ewg
