[oe-sync] New PoC description

Tue Oct 13 12:13:26 EDT 2015

Excellent. Thanks, I will modify poc proposal and send out.
Thank you very much!

-----Original Message-----
From: Mahadev Satyanarayanan [mailto:satya at cs.cmu.edu] 
Sent: Tuesday, October 13, 2015 8:36 AM
To: oe-sync at lists.andrew.cmu.edu
Cc: Young, Valerie J <valerie.j.young at intel.com>; Schuster, Rolf, Vodafone Group <Rolf.Schuster at vodafone.com>; Klas, Guenter, Vodafone Group <Guenter.Klas at vodafone.com>; Neal, Adrian, Vodafone Group <Adrian.Neal at vodafone.com>
Subject: Re: [oe-sync] New PoC description

Hi Valerie,
 Here is the Word version. 
             -- Satya

On Tuesday 13 October 2015 11:19:37 Young, Valerie J wrote:
> This is great work, folks.  The proposal is a great proposal, however, it needs a strong tie into or relative to MEC, for the MEC community to be interested.   We also have to utilize the PoC template (Word doc) for the MEC proposal as per the standard.  It would be helpful to share the word version of the doc, and I can cut and paste to have a revised PoC version.
> 
> >From MEC perspective, the computer vision software needs to have the MEC server at the edge of the network and this is how we should articulate the tie to MEC, correct?  MEC hardware is available (standard server) with virtualized environment and the routing for the individual consumer's cloudlet is available in the NEV Sdk.  Thus once the consumer requests the app, the consumer request will get routed.  Perhaps I am missing something with cloudlets.  However, from the above perspective, the work that would need to be done is the work already being done by the CMU team, except for the augmented reality app Satya is describing, which I was hoping his students were already implementing.  :)
> 
> For MEC, Openstack is a framework for management.  The cloudlet utilizes openstack and rest api's to get deployed on the MEC server.  This is my understanding, am I missing something here as I am not involved in the technical discussions?
> 
> Because MEC is becoming a standard for edge computing, my assumption is we would want cloudlets to influence the MEC standard, so cloudlets are part of the standard, correct?  
> 
> Should we have a meeting to discuss?
> 
> Thanks,
> Best regards,
> valerie
> 
> 
> 
> -----Original Message-----
> From: oe-sync-bounces+valerie.j.young=intel.com at lists.andrew.cmu.edu [mailto:oe-sync-bounces+valerie.j.young=intel.com at lists.andrew.cmu.edu] On Behalf Of Schuster, Rolf, Vodafone Group
> Sent: Tuesday, October 13, 2015 5:38 AM
> To: Klas, Guenter, Vodafone Group <Guenter.Klas at vodafone.com>; Mahadev Satyanarayanan <satya at cs.cmu.edu>; Weekly open-edge sync <oe-sync at lists.andrew.cmu.edu>; Neal, Adrian, Vodafone Group <Adrian.Neal at vodafone.com>
> Subject: Re: [oe-sync] New PoC description
> 
> I am fine with the PoC proposal. The key thing for me is to demonstrate the improvement of customer experience through Edge Computing live - rather that with videos. And be able to compare it directly with a backend cloud solution. Also it is good that we can have multiple instances of the demo and show it at fairs, conferences and meeting rooms anywhere on the planet...
> 
> Yes, the timeline is quite long, however, I guess Satya has to squeeze that into the schedule of his students...
> 
> @Valerie, Yun Chao and Adrian, are you still able to bring that PoC proposal on the MEC PoC list? Just as another channel to market? This would be great!
> 
> All the best
> Rolf.
> 
> -----Original Message-----
> From: oe-sync-bounces+rolf.schuster=vodafone.com at lists.andrew.cmu.edu [mailto:oe-sync-bounces+rolf.schuster=vodafone.com at lists.andrew.cmu.edu] On Behalf Of Klas, Guenter, Vodafone Group
> Sent: Dienstag, 13. Oktober 2015 12:33
> To: Mahadev Satyanarayanan; Weekly open-edge sync
> Subject: Re: [oe-sync] New PoC description
> 
> Hi all,
> 
> My take: if you look at the requirements section, I think that makes good sense. In order to maximise the value from a PoC, being able to show it to an audience without special device requirements and special network coverage and having portability of the PoC is mind are good points. Also, being able to compare the 2 situations (computation at edge versus fully in cloud) should help any audience to more easily understand the true benefit. Fine with those aspects.
> How this can fit into the ETSI MEC framework: Valerie, can you share your thoughts on this?
> Br
> Guenter
> 
> -----Original Message-----
> From: oe-sync-bounces+guenter.klas=vodafone.com at lists.andrew.cmu.edu [mailto:oe-sync-bounces+guenter.klas=vodafone.com at lists.andrew.cmu.edu] On Behalf Of Mahadev Satyanarayanan
> Sent: 12 October 2015 18:27
> To: Weekly open-edge sync
> Subject: [oe-sync] New PoC description
> 
> Hi everyone,
>  Because of my travel to MEC, Vodafone Newbury, SOSP, etc.  over the past few weeks, I have not been participating in the discussion of the draft 
> PoC that Valerie had sent out.    Late last week, I finally had a chance
> to read the draft PoC and some of the comments in the discussion
> threads.   Here are my thoughts:
> 
> 1. Using augmented reality (AR) as the basic application to show off the value of low latency and cloudlets is an excellent idea.  Kudos to Valerie
> for proposing this.      
> 
> 2. The specific AR approach being proposed seems to require significant 
> site customization (e.g. adding CMU-specific content, etc.).   This means
> that it is not easily portable to another site.  You will need site-specific content for the demo to be meaningful at the new site.
> 
> 3. The timeline seems too aggressive.   Virtually all the implementation
> work will be done by the CMU team, and I know in depth how is already
> on everyone's plate.   Generating custom content takes time, even if it
> is already available and only has to be repackaged.
> 
> Using Valerie's PoC as a starting point, I have created a description of a
> new PoC.   In some ways this more aggressive (functionally) than the
> original PoC.    At the same time, it is less agressive timewise and content-wise.
> 
> Please review and share your comments.   One clarification:  I am proposing
> this as  a PoC for the Open Edge Computing initiative.    Whether this is also
> a PoC for MEC is something that Vodafone, Huawei and Intel should decide.
> I don't know if this timeline is acceptable for the MEC PoC.  Also, I recall that a decision about MEC PoCs has to be made very soon.  Not sure
> if this is already too late for consideration.   If it is attractive as a MEC PoC
> also, then I will delighted that we can get a two-fer.
> 
>   Cheers
>                      -- Satya
> _______________________________________________
> oe-sync mailing list
> oe-sync at lists.andrew.cmu.edu
> https://lists.andrew.cmu.edu/mailman/listinfo/oe-sync
> _______________________________________________
> oe-sync mailing list
> oe-sync at lists.andrew.cmu.edu
> https://lists.andrew.cmu.edu/mailman/listinfo/oe-sync
> 