[openstack-dev] NFV in OpenStack use cases and context

Yathiraj Udupi (yudupi) yudupi at cisco.com
Mon Jun 16 14:32:55 UTC 2014


Hi Sylvain, 

The Smart Scheduler work should work along withGantt easily, as it is just
another scheduler driver, instead of using the Filters, it will have a
separate mechanism of solving the placement problem all at once.
So just like how it currently works along with Nova scheduler, it should
easily be integrated into Gantt, is what I would imagine. I have not
looked very closely at the Gantt work, but I assume if you claim, no
behavior changes in Gantt, and if you are continuing to use
FilterScheduler, SmartScheduler should fit in too.

I will try to update the nova-spec ([1]
https://review.openstack.org/#/c/96543/) for Smart Scheduler, and let the
review continue sometime soon.

Thanks,
Yathi. 



On 6/13/14, 12:37 AM, "Sylvain Bauza" <sbauza at redhat.com> wrote:

>Hi Yathi,
>
>Le 12/06/2014 20:53, Yathiraj Udupi (yudupi) a écrit :
>> Hi Alan, 
>>
>> Our Smart (Solver) Scheduler blueprint
>> (https://blueprints.launchpad.net/nova/+spec/solver-scheduler ) has been
>> in the works in the Nova community since late 2013.  We have demoed at
>>the
>> Hong Kong summit, as well as the Atlanta summit,  use cases using this
>> smart scheduler for better, optimized resource placement with complex
>> constrained scenarios.  So to let you know this work was started as a
>> smart way of doing scheduling, applicable in general and not limited to
>> NFV.  Currently we feel NFV is a killer app for driving this blueprint
>>and
>> work ahead, however is applicable for all kinds of resource placement
>> scenarios. 
>>
>> We will be very interested in finding out more about your blueprints
>>that
>> you are referring to here, and see how it can be integrated as part of
>>our
>> future roadmap. 
>
>Indeed, Smart Scheduler is something that could help NFV use-cases. My
>only concern is about the necessary steps for providing such a feature,
>with regards to the scheduler breakout that is coming.
>
>Could you please make sure the current nova-spec [1] is taking in
>account all other efforts about the scheduler, like scheduler forklift
>[2], on-demand resource reporting [3] or others ? It also seems the spec
>is not following the defined template, could you please fix it ? It
>would be easier to review your proposal.
>
>Gantt and Nova scheduler teams are attending a weekly meeting every
>Tuesday at 3pm UTC. Would you have a chance to join, it would be great
>to discuss about your proposal and how we can identify all the
>milestones for this and potentially track progress on it.
>
>Thanks,
>-Sylvain
>
>[1] https://review.openstack.org/#/c/96543/
>[2] https://review.openstack.org/82133 and
>https://review.openstack.org/89893
>[3] https://review.openstack.org/97903
>
>
>
>> Thanks,
>> Yathi. 
>>
>>
>> On 6/12/14, 10:55 AM, "Alan Kavanagh" <alan.kavanagh at ericsson.com>
>>wrote:
>>
>>> Hi Ramki
>>>
>>> Really like the smart scheduler idea, we made a couple of blueprints
>>>that
>>> are related to ensuring you have the right information to build a
>>> constrained based scheduler. I do however want to point out that this
>>>is
>>> not NFV specific but is useful for all applications and services of
>>>which
>>> NFV is one. 
>>>
>>> /Alan
>>>
>>> -----Original Message-----
>>> From: ramki Krishnan [mailto:ramk at Brocade.com]
>>> Sent: June-10-14 6:06 PM
>>> To: OpenStack Development Mailing List (not for usage questions)
>>> Cc: Chris Wright; Nicolas Lemieux; Norival Figueira
>>> Subject: Re: [openstack-dev] NFV in OpenStack use cases and context
>>>
>>> Hi Steve,
>>>
>>> Forgot to mention, the "Smart Scheduler (Solver Scheduler) enhancements
>>> for NFV: Use Cases, Constraints etc." is a good example of an NFV use
>>> case deep dive for OpenStack.
>>>
>>> 
>>>https://urldefense.proofpoint.com/v1/url?u=https://docs.google.com/docum
>>>en
>>> 
>>>t/d/1k60BQXOMkZS0SIxpFOppGgYp416uXcJVkAFep3Oeju8/edit%23heading%3Dh.wlbc
>>>la
>>> 
>>>gujw8c&k=oIvRg1%2BdGAgOoM1BIlLLqw%3D%3D%0A&r=%2FZ35AkRhp2kCW4Q3MPeE%2BxY
>>>2b
>>> 
>>>qaf%2FKm29ZfiqAKXxeo%3D%0A&m=vTulCeloS8Hc59%2FeAOd32Ri4eqbNqVE%2FeMgNRzG
>>>Zn
>>> 
>>>z4%3D%0A&s=836991d6daab66b519de3b670db8af001144ddb20e636665b395597aa1185
>>>38
>>> f
>>>
>>> Thanks,
>>> Ramki
>>>
>>> -----Original Message-----
>>> From: ramki Krishnan
>>> Sent: Tuesday, June 10, 2014 3:01 PM
>>> To: OpenStack Development Mailing List (not for usage questions)
>>> Cc: Chris Wright; Nicolas Lemieux; Norival Figueira
>>> Subject: RE: NFV in OpenStack use cases and context
>>>
>>> Hi Steve,
>>>
>>> We are have OpenStack gap analysis documents in ETSI NFV under member
>>> only access. I can work on getting public version of the documents (at
>>> least a draft) to fuel the kick start.
>>>
>>> Thanks,
>>> Ramki
>>>
>>> -----Original Message-----
>>> From: Steve Gordon [mailto:sgordon at redhat.com]
>>> Sent: Tuesday, June 10, 2014 12:06 PM
>>> To: OpenStack Development Mailing List (not for usage questions)
>>> Cc: Chris Wright; Nicolas Lemieux
>>> Subject: Re: [openstack-dev] [NFV] Re: NFV in OpenStack use cases and
>>> context
>>>
>>> ----- Original Message -----
>>>> From: "Steve Gordon" <sgordon at redhat.com>
>>>> To: "Stephen Wong" <stephen.kf.wong at gmail.com>
>>>>
>>>> ----- Original Message -----
>>>>> From: "Stephen Wong" <stephen.kf.wong at gmail.com>
>>>>> To: "ITAI MENDELSOHN (ITAI)" <itai.mendelsohn at alcatel-lucent.com>,
>>>>> "OpenStack Development Mailing List (not for usage questions)"
>>>>> <openstack-dev at lists.openstack.org>
>>>>>
>>>>> Hi,
>>>>>
>>>>>     Perhaps I have missed it somewhere in the email thread? Where is
>>>>> the use case => bp document we are supposed to do for this week? Has
>>>>> it been created yet?
>>>>>
>>>>> Thanks,
>>>>> - Stephen
>>>> Hi,
>>>>
>>>> Itai is referring to the ETSI NFV use cases document [1] and the
>>>> discussion is around how we distill those - or a subset of them - into
>>>> a more consumable format for an OpenStack audience on the Wiki. At
>>>> this point I think the best approach is to simply start entering one
>>>> of them (perhaps #5) into the Wiki and go from there. Ideally this
>>>> would form a basis for discussing the format etc.
>>>>
>>>> Thanks,
>>>>
>>>> Steve
>>>>
>>>> [1]
>>>> http://www.etsi.org/deliver/etsi_gs/NFV/001_099/001/01.01.01_60/gs_NFV
>>>> 001v010101p.pdf
>>> To try and kick start things I have created a table on the wiki [1]
>>>based
>>> on the *DRAFT* NFV Performance & Portability Best Practises document
>>>[2].
>>> This really lists workload types rather than specific applications,
>>> although I've put in an examples column we can populate with them.
>>>
>>> I find it a useful way to quickly break down some of the
>>>characteristics
>>> of NFV applications at a glance. What do people think of this as
>>> something to start with? Remember, it's a wiki! So anyone is welcome to
>>> either expand the table or start adding more concrete user stories
>>>(e.g.
>>> around ETSI NFV use case number 5 that Itai and I have been referring
>>>to,
>>> or any other VNF for that matter) in this section (we may/want need to
>>> create a separate page but for now it seems OK to get started here).
>>>
>>> Thanks,
>>>
>>> Steve
>>>
>>> [1] https://wiki.openstack.org/wiki/Meetings/NFV#Use_Cases
>>> [2] 
>>> 
>>>http://docbox.etsi.org/ISG/NFV/Open/Latest_Drafts/NFV-PER001v009%20-%20N
>>>FV
>>> %20Performance%20&%20Portability%20Best%20Practises.pdf
>>>
>>> _______________________________________________
>>> OpenStack-dev mailing list
>>> OpenStack-dev at lists.openstack.org
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>
>>> _______________________________________________
>>> OpenStack-dev mailing list
>>> OpenStack-dev at lists.openstack.org
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>
>>> _______________________________________________
>>> OpenStack-dev mailing list
>>> OpenStack-dev at lists.openstack.org
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>
>> _______________________________________________
>> OpenStack-dev mailing list
>> OpenStack-dev at lists.openstack.org
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>
>_______________________________________________
>OpenStack-dev mailing list
>OpenStack-dev at lists.openstack.org
>http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev




More information about the OpenStack-dev mailing list