I'm moving it to the Infra wiki space it relates to more than just the Pharos project. thanks
https://wiki.opnfv.org/display/INF/CNTT+RI+Lab+Infrastructure+Discussion/ On 11/6/19 5:42 AM, Lincoln Lavoie wrote: > Hello All, > > I've created the Wiki discussion page, as requested by > Mark: > https://wiki.opnfv.org/display/pharos/CNTT+RI+Lab+Infrastructure+Discussion > > I've added your comments and tagged in this page, if you've responded, > so you probably had more emails as well (sorry about that). > > Please use this page to continue the discussions. > > Cheers, > Lincoln > > On Wed, Nov 6, 2019 at 7:58 AM SHOSTAK, MARK <ms7...@att.com > <mailto:ms7...@att.com>> wrote: > > +Ahmed > > > > *From:* SHOSTAK, MARK > *Sent:* Wednesday, November 06, 2019 05:45 > *To:* georg.k...@ericsson.com <mailto:georg.k...@ericsson.com>; > gergely.csat...@nokia.com <mailto:gergely.csat...@nokia.com>; > lylav...@iol.unh.edu <mailto:lylav...@iol.unh.edu> > *Cc:* opnfv-tech-dis. <opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org>>; FIX, MICHAEL A > <mf4...@att.com <mailto:mf4...@att.com>>; fuq...@chinamobile.com > <mailto:fuq...@chinamobile.com> > *Subject:* RE: [opnfv-tech-discuss] Lab hardware requirements for CNTT > > > > Hi All, > > > > I’d like to ensure there’s traceability and a basis/justification > for all RI/CIRV lab h/w requirements we specify, as well as > assumptions, so we know we’ve covered everything and can > substantiate our conclusions and ultimate decisions. > > > > Gergely wrote: > > > /If we would like to test both profiled from the Reference Model > I think we need 2 * 20C/40T CPU./ > > *Gergely,*I don’t contest this, but can you connect the dots for > me on how you extrapolated socket/thread counts? > > > > I can make a table, unless someone beats me to it, as a start I’m > envisioning: > > - Hardware item requested > > - Rx reference, if applicable > > - Basis // logic used to extrapolate h/w > > - Assumptions // such as, this item will allow us to test > x, y, z > > > > Thoughts/suggestions on the table? > > > > *Lincoln*, Can we host this table on the Infra wiki, or where > would you recommend? > > > > *Georg*, The network area is evolving, so without going into the > history, you are correct, the speed is intended to reflect a > virtual/vNIC rate. References to physical NICs have been removed > from the abstraction context. There are still open questions in > this area in regard to what bandwidth is required of the physical > NICs for the lab machines, as well as updates for the RM itself. > As you noted you don’t see the value in those numbers, can you > revise this section of the RM? If so, let’s discuss objectives/vision. > > > Thanks! > > -Mark > > > > > > Gergely wrote: > > > If we would like to test both profiled from the Reference Model I > think we need 2 * 20C/40T CPU. > > as > > > > *From:* opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org> > [mailto:opnfv-tech-discuss@lists.opnfv.org] *On Behalf Of *Georg > Kunz via Lists.Opnfv.Org <http://Lists.Opnfv.Org> > *Sent:* Wednesday, November 06, 2019 02:51 > *To:* gergely.csat...@nokia.com > <mailto:gergely.csat...@nokia.com>; lylav...@iol.unh.edu > <mailto:lylav...@iol.unh.edu>; opnfv-tech-dis. > <opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org>> > *Cc:* opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org> > *Subject:* Re: [opnfv-tech-discuss] Lab hardware requirements for CNTT > > > > Hi all, > > > > Regarding the network interfaces, the numbers listed in the > reference architecture cannot be mapped to physical hardware, > simply because no NICs supporting those bandwidths are widely > available (maybe 100Gbs). This table seems to refer to the > bandwidth of virtual NICs in a flavor. I personally don’t see the > concrete value (and use case) of these numbers yet, but if I am > missing anything here, please let me know. > > > > Best regards > > Georg > > > > *From:* opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org> > <opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org>> *On Behalf Of > *Gergely Csatari via Lists.Opnfv.Org <http://Lists.Opnfv.Org> > *Sent:* Wednesday, November 6, 2019 6:45 AM > *To:* lylav...@iol.unh.edu <mailto:lylav...@iol.unh.edu>; > opnfv-tech-dis. <opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org>> > *Cc:* opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org> > *Subject:* Re: [opnfv-tech-discuss] Lab hardware requirements for CNTT > > > > Hi, > > > > If we would like to test both profiled from the Reference Model I > think we need 2 * 20C/40T CPU. > > If we would like to test all the flavors listed in the Reference > Model we need hardware what can run flavors with 32GB RAM and 320 > GM local disks. > > If we would like to test all the network options from the > Reference Architecture we need hardware capable to run instances > with 1*100, 1*200, 1*300, 1*400, 1*500, 1*600 Gbps interfaces. > > > > Br, > > Gerg0 > > > > > > *From:* opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org> > <opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org>> *On Behalf Of > *Lincoln Lavoie via Lists.Opnfv.Org <http://Lists.Opnfv.Org> > *Sent:* Wednesday, November 6, 2019 2:56 AM > *To:* opnfv-tech-dis. <opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org>> > *Cc:* opnfv-tech-discuss@lists.opnfv.org > <mailto:opnfv-tech-discuss@lists.opnfv.org> > *Subject:* [opnfv-tech-discuss] Lab hardware requirements for CNTT > > > > Hello All, > > > > We (Infra Project) would like to have some discussions on how to > best provide the laboratory resources for the CNTT Reference > Implementation (CIRV OPNFV Project). Within this context, we are > considering CNTT as an upstream project, where the outcome of this > discussion would "drive" a patch to the upstream CNTT > documentation here > > (https://github.com/cntt-n/CNTT/blob/master/doc/ref_impl/cntt-ri/chapters/chapter04.md > > <https://urldefense.proofpoint.com/v2/url?u=https-3A__protect2.fireeye.com_v1_url-3Fk-3D1309b30f-2D4f8391db-2D1309f394-2D0cc47ad93e6a-2Daa2f2e3f2efdb8e7-26q-3D1-26e-3D50b69976-2D937a-2D4dc2-2D9da9-2D0b538ef598b7-26u-3Dhttps-253A-252F-252Fgithub.com-252Fcntt-2Dn-252FCNTT-252Fblob-252Fmaster-252Fdoc-252Fref-5Fimpl-252Fcntt-2Dri-252Fchapters-252Fchapter04.md&d=DwMGaQ&c=LFYZ-o9_HUMeMTSQicvjIg&r=jpZpdbY46Bqv_Ghv8dHXcbwtfXtibdpcqKx0s1QrMGM&m=Ik2JEX2Ed56yh6UbEChl4lji3Tyn71RVr3uq7Y0Ojxw&s=TGm1FkyHxJCd1rzUyOkhUWu39MqpWv308sib8lwpNHQ&e=>). > > > > One aspect we have discussed within the Infra team is to develop a > road map for the labs to come online to the full / final set of > hardware needed by the CNTT requirements. As part of this road > map, I think we should get some agreements in place for the > Release 1 hardware as part of the reference implementation work. > The CIRV team will need to ensure this will meet their needs in > working with the installers, etc. If we make these agreements as > a community, we should be able to feed these upstream to CNTT as a > change request / proposal to the section cited above. > > > > For lab release 1, can we agree to a common requirement for the > controller and compute nodes (i.e. a server, meeting the > requirements, could serve either role)? > > > > For lab release 1, can we agree the jump host can run as a virtual > machine, instead of bare-metal? > > > > For lab release 1, can we agree to a homogenous set of hardware > (i.e. there is no difference between B / C / N flavors in the > compute nodes)? > > > > For lab release 1, can we agree to the following specification for > bare-metal nodes: > > * 2.2GHz 24C/48T > > * 512GB RAM > > * 3.2 TB SSD > > * 4x 25Gbps Network Ports > > > > For lab release 1, can we agree to the following specification for > virtual jump hosts: > > * 2x vCPU cores > > * 8GB RAM > > * 1TB HDD > > * 2x virtual NICs > > > > This is a fair bit to discuss, but let's try to keep it all within > this thread and on the list, so the community can easily > participate in the discussion. I'm also happy to host a couple of > IRC "meet ups" if it would make an easier discussion as well > (outputs / minutes of those discussions would be posted back to > this thread. > > > > Cheers, > Lincoln > > > > -- > > *Lincoln Lavoie* > > Senior Engineer, Broadband Technologies > > 21 Madbury Rd., Ste. 100, Durham, NH 03824 > > lylav...@iol.unh.edu <mailto:lylav...@iol.unh.edu> > > https://www.iol.unh.edu > > <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.iol.unh.edu&d=DwMGaQ&c=LFYZ-o9_HUMeMTSQicvjIg&r=jpZpdbY46Bqv_Ghv8dHXcbwtfXtibdpcqKx0s1QrMGM&m=Ik2JEX2Ed56yh6UbEChl4lji3Tyn71RVr3uq7Y0Ojxw&s=750IpFfdq5dKaUFrdFAmysopS_N1JGwuXEAHxC29O7A&e=> > > +1-603-674-2755 (m) > > Image removed by sender. > > <https://urldefense.proofpoint.com/v2/url?u=https-3A__www.iol.unh.edu_&d=DwMGaQ&c=LFYZ-o9_HUMeMTSQicvjIg&r=jpZpdbY46Bqv_Ghv8dHXcbwtfXtibdpcqKx0s1QrMGM&m=Ik2JEX2Ed56yh6UbEChl4lji3Tyn71RVr3uq7Y0Ojxw&s=-Qs3P664KoN80zbXBDnTLUwrZVOE3op-YzWkwWMK-o0&e=> > > > > -- > *Lincoln Lavoie* > Senior Engineer, Broadband Technologies > 21 Madbury Rd., Ste. 100, Durham, NH 03824 > lylav...@iol.unh.edu <mailto:lylav...@iol.unh.edu> > https://www.iol.unh.edu > +1-603-674-2755 (m) > <https://www.iol.unh.edu/> > > -=-=-=-=-=-=-=-=-=-=-=- > Links: You receive all messages sent to this group. > > View/Reply Online (#23707): > https://lists.opnfv.org/g/opnfv-tech-discuss/message/23707 > Mute This Topic: https://lists.opnfv.org/mt/42392960/1198612 > Group Owner: opnfv-tech-discuss+ow...@lists.opnfv.org > Unsubscribe: https://lists.opnfv.org/g/opnfv-tech-discuss/unsub > [j...@jento.io] > -=-=-=-=-=-=-=-=-=-=-=- -- Thank You Jack Morgan
signature.asc
Description: OpenPGP digital signature
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#23710): https://lists.opnfv.org/g/opnfv-tech-discuss/message/23710 Mute This Topic: https://lists.opnfv.org/mt/42392960/21656 Group Owner: opnfv-tech-discuss+ow...@lists.opnfv.org Unsubscribe: https://lists.opnfv.org/g/opnfv-tech-discuss/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
