Hi, Mirja: -----邮件原件----- 发件人: Mirja Kühlewind [mailto:[email protected]] 发送时间: 2017年2月24日 21:11 收件人: Qin Wu 抄送: Randriamasy, Sabine (Nokia - FR); Gurbani, Vijay (Nokia - US); [email protected]; IETF ALTO; [email protected]; Spencer Dawkins at IETF 主题: Re: [alto] Submitting draft-wu-alto-te-metrics as a WG item
Hi Qin, see below. > Am 22.02.2017 um 03:52 schrieb Qin Wu <[email protected]>: > > Hi, Mirja and all: > Sorry for late reply, finally get time to follow up. > I think the section 10 of draft-ietf-alto-performance-metrics-00 > doesn't intend to create new registry for each new performance metric > instead, It intends to use the "ALTO cost metric" registry defined in ALTO > base protocol (i.e.,[RFC7285] section 14.2) and add a few new entries into > this registry. Okay. Actually didn’t see this. > > This registry is not for new performance metric but for new ALTO cost metric > type. > In addition, the registry created in ISIS-TE(i.e.,section 12 of RFC7810) is > for sub-TLV type not for new performance metric type. > So these registries have different purpose, not sure they will conflict with > one another. > My take is > To align with IPPM and ISIS-TE, OSPF-TE, what we need to do is If the > performance metric is measured using IPPM method, we will use the same > measurement method defined for IPPM for this metric. > If the performance metric is measured using ISIS-TE method, we will use the > same measurement method defined in ISIS-TE for this metric. > If the performance metric is end to end metric, we could use method defined > in draft-ietf-pce-pcep-service-aware for this metric. Okay. I would recommend to make sure that you don’t duplicate text but refer to existing RFC instead and potentially even refer to those RFC in the registry (if it’s a one-to-one mapping). [Qin]: Sounds good idea. Thanks for your proposal. > > In addition, I propose to update IANA section to clear confusion as follows: > Section 10 of draft-wu-alto-te-metrics OLD TEXT: > " > IANA has added the following entries to the ALTO cost map Properties > registry, defined in Section 3 of [RFCXXX]. > > " > NEW TEXT: > " > IANA has created and now maintains the "ALTO Cost Metric Registry", listed in > Table 3 of [RFC7285]. This document requests to add the following entries to > “ALTO Cost Meric Registry”. > „ > The original text is fine. See guidance given in draft-leiba-cotton-iana-5226bis (sec 3.1): you should make sure to clearly identify the existing registry; only name it correctly ("ALTO Cost Metric Registry“) and provide an URL. [Qin]: Good suggestion, thanks. Mirja > -Qin > -----邮件原件----- > 发件人: Mirja Kühlewind [mailto:[email protected]] > 发送时间: 2017年2月10日 23:45 > 收件人: Qin Wu; Randriamasy, Sabine (Nokia - FR) > 抄送: Gurbani, Vijay (Nokia - US); [email protected]; > IETF ALTO; [email protected]; Spencer Dawkins at IETF > 主题: Re: [alto] Submitting draft-wu-alto-te-metrics as a WG item > > Hi Sabine, hi Qin, > > adding the ippm chairs and Spencer as the responsible AD. > > My short answer: I bought this up for discussion and I don’t have a good > final answer yet. I would recommend to discuss this with the ippm group > (maybe even ask for further feedback on the ippm mailing list). > > Longer answer: The reason why I brought this up is that you not only define > cost metrics but you also talk about how to measure these values. This seems > duplicating work with ippm and makes it harder to keep different efforts in > the IETF aligned. Yes, you are right not all of your metrics map to ippm work > but in this case I would probably also point to OSPF-TE and ISIS-TE (as you > do in the draft yourself). There are already registries for these value. So > I’m still questioning if an own alto registry is needed…? > > Mirja > > >> Am 10.02.2017 um 04:01 schrieb Qin Wu <[email protected]>: >> >> Align with IPPM, I agree with the second approach, since IPPM is not limited >> to IP based metric but also cover protocol specific metric and application >> specific metric, I think use the IPPM registry >> (https://tools.ietf.org/html/draft-ietf-ippm-metric-registry-10#section-10) >> is pretty much reasonable to me, we can fill additional ALTO related >> Performance Metric Registry Entries in this registry. >> what is not clear to me is the relation between >> draft-ietf-alto-performance-metrics and >> draft-ietf-ippm-initial-registry-02. It seems >> draft-ietf-ippm-initial-registry-02 intends to cover all the performance >> metrics which are defined in the IETF and add registry entry for each of >> these performance metrics. Also I am not sure whether >> draft-ietf-ippm-initial-registry-02 only focus on (1).direct metric can be >> measured in the network via 3rd party measurement tools (2).or some >> composite metric or derived metric or aggregated performance metrics? >> If the answer is draft-ietf-ippm-initial-registry-02 only focus on (1), I >> think we can cover (2). >> >> -Qin >> -----邮件原件----- >> 发件人: Randriamasy, Sabine (Nokia - FR) >> [mailto:[email protected]] >> 发送时间: 2017年2月9日 21:53 >> 收件人: Mirja Kühlewind; Gurbani, Vijay (Nokia - US) >> 抄送: [email protected]; Qin Wu; IETF ALTO >> 主题: RE: [alto] Submitting draft-wu-alto-te-metrics as a WG item >> >> Hello Mirja and Altoers, >> >> Should the ALTO WG consider broadening the scope of the ALTO performance >> metrics draft by having at least a section for potential IPPM metrics and >> another section covering other network performance metrics? >> >> Indeed, network performance metrics that are already documented in the IPPM >> initial registry can just be referenced in the ALTO metrics document. As the >> IPPM initial registry is in progress, the ALTO metrics draft may include a >> placeholder for IPPM relevant metrics until they get adopted and specified >> in IPPM. Definitely, ALTO-IPPM coordination will be fruitful. >> >> However, quoting Al Morton's e-mail on July 21st: "Another way to interpret >> what ALTO needs is to look at cost metrics the same as link costs in >> routing. Routing link costs are usually static and the values might be set >> relative to some real metric like latency or max link line rate. These >> involve no measurement at all." >> >> Additionally, some network metric values may stem from both measurements and >> provider policy as for example, many bandwidth related ALTO metrics. ALTO >> may convey such information not available via 3rd party measurement tools. >> Besides, IPPM informational RFC 5136 points the difficulty to have a unified >> nomenclature for network capacity related measurements. >> >> To guide users and applications, ALTO performance Metric values may also add >> abstraction to measurements and/or provide unitless performance scores. >> Last, given the growing WG interest in wireless networks, nothing precludes >> ALTO metrics from covering them. >> >> Thanks, >> Sabine >> >>>> -----Original Message----- >>>> From: alto [mailto:[email protected]] On Behalf Of Mirja >>>> Kühlewind >>>> Sent: jeudi 5 janvier 2017 15:02 >>>> To: Gurbani, Vijay (Nokia - US) <[email protected]> >>>> Cc: [email protected]; Qin Wu <[email protected]>; >>>> IETF ALTO <[email protected]> >>>> Subject: Re: [alto] Submitting draft-wu-alto-te-metrics as a WG >>>> item >>>> >>>> Hi all, >>>> >>>> sorry for being super late. >>>> >>>> One thing is to align with IPPM, there other question is do we need >>>> an own alto registry for these metric or can we use the IPPM >>>> registry? I would prefer the second approach if appropriate which >>>> would also mean that the other metrics that are not part of the IPPM >>>> registry yet, should be registered there. >>>> Given this is still work in progress, we definitely should >>>> coordinate with the IPPM working group and see if the registration >>>> itself should happen in IPPM or ALTO. >>>> >>>> Mirja >>>> >>>> >>>>> Am 29.08.2016 um 21:36 schrieb Vijay K. Gurbani >>>>> <vijay.gurbani@nokia- >>>> bell-labs.com>: >>>>> >>>>> On 08/29/2016 03:47 AM, Qin Wu wrote: >>>>>> Here is the update of draft-wu-alto-te-metrics to address two >>>>>> open issues raised in the last meeting. The diff is: >>>>>> https://www.ietf.org/rfcdiff?url2=draft-wu-alto-te-metrics-09 >>>>> >>>>> Qin: Thanks. More below. >>>>> >>>>> Folks: Please take a look at the above diffs that address the >>>>> concerns in [1]. >>>>> >>>>> Authors: Please wait a few days for any feedback from the WG and >>>>> incorporate feedback in a -00 version. Irrespective of any >>>>> feedback, please generate a WG -00 and submit by next week (mid- >>>>> to late next week). >>>>> >>>>> [1] >>>>> https://www.ietf.org/mail-archive/web/alto/current/msg03277.html >>>>> >>>>> Thanks, >>>>> >>>>> - vijay >>>>> -- >>>>> Vijay K. Gurbani, Bell Laboratories, Nokia Networks >>>>> 1960 Lucent Lane, Rm. 9C-533, Naperville, Illinois 60563 (USA) >>>>> Email: [email protected] / [email protected] >>>>> Web: http://ect.bell-labs.com/who/vkg/ | Calendar: >>>>> http://goo.gl/x3Ogq >>>>> >>>>> _______________________________________________ >>>>> alto mailing list >>>>> [email protected] >>>>> https://www.ietf.org/mailman/listinfo/alto >>>> >>>> _______________________________________________ >>>> alto mailing list >>>> [email protected] >>>> https://www.ietf.org/mailman/listinfo/alto > _______________________________________________ alto mailing list [email protected] https://www.ietf.org/mailman/listinfo/alto
