Re: [bmwg] A new draft aimed at bmwg

["MORTON, ALFRED C (AL)" <[e-mail]>]

Mr. Lee and co-authors,

I read your draft: it is a good start for this topic,
and have a few comments to share.

General:
Instantiation of containers, pods, and internetworking
with K8s is somewhat non-deterministic and apparently
NUMA-Node agnostic, which is one way of saying that 
performance will likely vary from one instantiation to
the next, and the next, etc.  It would be good to 
specifically mention repeated instantiation and testing 
to quantify the performance variation under these 
conditions.

Related to the Memory Management Unit POV, section 3.1:
Life-cycle operations usually require considerable memory transfers.
Does this also imply that memory access times in Container VNFs
will be be affected when host resources are committed to Life-cycle
operations on other VNFs (such as instantiation/migrqation/deletion)?
IOW, will the effects of Life-cycle management be greater in the 
containerized infrastructure than for hypervisor with VM-based VNFs?

thanks,
Al
(as a participant)
 

> -----Original Message-----
> From: bmwg [mailto:[e-mail]] On Behalf Of MORTON, ALFRED C
> (AL)
> Sent: Friday, March 8, 2019 4:52 AM
> To: WB Lee <[e-mail]>; [e-mail]
> Subject: Re: [bmwg] A new draft aimed at bmwg
> 
> *** Security Advisory: This Message Originated Outside of AT&T ***.
> Reference http://cso.att.com/EmailSecurity/IDSP.html for more information.
> 
> Thanks for your timely contribution,
> Mr. Lee and co-authors.
> I will certainly take time to review.
> 
> Al
> bmwg co-chair
> 
> > -----Original Message-----
> > From: bmwg [mailto:[e-mail]] On Behalf Of WB Lee
> > Sent: Thursday, March 7, 2019 11:12 PM
> > To: [e-mail]
> > Subject: [bmwg] A new draft aimed at bmwg
> >
> >
> > Hi All,
> >
> > We have been researched the POD testing in Containerized
> Infrastructures.
> > We submitted the new draft, "Considerations for Benchmarking Network
> > Performance in Containerized Infrastructures".
> >
> > https://urldefense.proofpoint.com/v2/url?u=https-
> > 3A__www.ietf.org_id_draft-2Ddcn-2Dbmwg-2Dcontainerized-2Dinfra-
> > 2D00.txt&d=DwICAg&c=LFYZ-
> >
> o9_HUMeMTSQicvjIg&r=OfsSu8kTIltVyD1oL72cBw&m=GqCJ6KxJEotosmsZYGIRkgzSnPHw6
> > 9gsd-qljl5VsKM&s=8TVO-ODX_CHjQHvGlCffATtz0hQeX6HUeog7Q8mGajc&e=
> >
> > Abstract
> >
> >    This draft describes benchmarking considerations for a containerized
> >    infrastructure.  In a containerized infrastructure, Virtualized
> >    Network Functions(VNFs) are deployed on operating-system-level
> >    virtualization platform by abstracting the user namespace as opposed
> >    to virtualization using a hypervisor.  Leveraging this, the system
> >    configurations and networking scenarios for VNF benchmarking will be
> >    partially changed by way of resource allocation and network port
> >    binding between a physical host and VNFs.  In this draft we compare
> >    the state of the art in container networking architecture with
> >    networking on VM-based virtualized systems, and provide several test
> >    scenarios for network performance in containerized infrastructure.
> >
> > We welcome any comments.
> >
> > Regards,
> > Wangbong Lee
> >
> > _______________________________________________
> > bmwg mailing list
> > [e-mail]
> > https://urldefense.proofpoint.com/v2/url?u=https-
> > 3A__www.ietf.org_mailman_listinfo_bmwg&d=DwICAg&c=LFYZ-
> >
> o9_HUMeMTSQicvjIg&r=OfsSu8kTIltVyD1oL72cBw&m=GqCJ6KxJEotosmsZYGIRkgzSnPHw6
> > 9gsd-qljl5VsKM&s=Bw0LUXS-NxJHsEwXDVqXkyG-vuE7YWyeQMbEHEWsGfA&e=
> 
> _______________________________________________
> bmwg mailing list
> [e-mail]
> https://urldefense.proofpoint.com/v2/url?u=https-
> 3A__www.ietf.org_mailman_listinfo_bmwg&d=DwICAg&c=LFYZ-
> o9_HUMeMTSQicvjIg&r=OfsSu8kTIltVyD1oL72cBw&m=qXXXuSDIome9R0009wamlXyeacskd
> HvEajvNIkJnZi0&s=5A7xn7LyNn_hq7KAWSW9fFJj4i1SrXVeH8uoZ4DlPZU&e=