[scientific-sig] IRC meeting 1100 UTC: Edge computing and Scientific use cases
Bogdan Dobrelya
bdobreli at redhat.com
Wed Mar 6 16:47:28 UTC 2019
[added edge-computing at lists.openstack.org]
JFYI: some background for ICFC 2019 event:
> =================================================================================
> IEEE International Conference on Fog Computing (ICFC 2019)
> June 24-26, 2019
> Prague, Czech Republic
> http://conferences.computer.org/ICFC/2019/
> Co-located with the IEEE International Conference on Cloud Engineering
> (IC2E 2019)
> ==================================================================================
>
> Important Dates
>
> --------------------
>
> Paper registration and abstract: Nov 1st, 2018
> Full paper submission due: Nov 8th, 2018
> Notification of paper acceptance: Jan. 20th, 2019
> Workshop and tutorial proposals due: Nov 11, 2018
> Notification of proposal acceptance: Nov 18, 2018
So I have a draft of a positioning paper [0]... what missed that event :)
tl;dr: It is known that causal consistency is the best suitable for
high-latency, high-scale and highly dynamic nature of membership in
clusters... which is exactly what multiple autonomous control planes
represent in edge clouds for post-MVP phases.
Even though the paper didn't make it for ICFC 2019 acceptance as I
naively planned (not), hopefully we as community could find some use of
it in future. Like collaboratively improving its contents, or making it
readable by humans, then pushing it for other places for wider audience.
It's just free for any use, please go for it! :)
Or maybe we could prototyping something for OpenStack edge post-MVP
cases out of SwiftCloud [1] just like the folks did for CockroachDB [2]?
Note, while that github repo does not look super alive and driven by
many contributions, that work looks really promising, you can read more
for the theoretical fundamentals of it in "SwiftCloud: Fault-Tolerant
Geo-Replication Integrated all the Way to the Client Machine" and
"SwiftCloud: Write fast, read in the past: Causal consistency for
client-side applications". That really buys to me to put it all
client-sided at the edge and make it doing all geo-replication dances.
Ofc there is more for prototyping. For example, building some
Bolt-on'ish shim middleware on top of any popular eventual consistent
storage, like Riak (not sure there is opensource implementations for
Bolt-on yet). Or trying something at api-to-api level for high-level
synchronization with "delayed replication" of queued operations.
PS. @Pierre and @Adrien I gave it a quick glimpse into some of the
sources you linked for the related work section of the positioning
paper, thank you a lot for that and x1000 for authoring such things.
So the draft now almost hits the recommended limit of a 6 pages, so I
think there should be pretty much it for a positioning paper.
[0]
https://github.com/bogdando/papers-ieee/blob/master/ICFC-2019/LaTeX/position_paper_1570506394.pdf
[1] https://github.com/SyncFree/SwiftCloud
[2]
http://beyondtheclouds.github.io/blog/openstack/cockroachdb/2018/06/04/evaluation-of-openstack-multi-region-keystone-deployments.html
On 27.02.2019 14:33, Adrien Lebre wrote:
> Hi,
>
> Getting a look to this post could also make sense:
> http://beyondtheclouds.github.io/blog/openstack/cockroachdb/2018/06/04/evaluation-of-openstack-multi-region-keystone-deployments.html
>
> Best regards,
> Adrien
> PS: Thanks Pierre.
> --
> Prof. IMT Atlantique / Inria / LS2N
> Head of the STACK Research Group
> www.emn.fr/x-info/alebre08 <http://www.emn.fr/x-info/alebre08>
>
>> On 27 Feb 2019, at 14:18, Pierre Riteau <pierre at stackhpc.com
>> <mailto:pierre at stackhpc.com>> wrote:
>>
>> Hello Bogdan,
>>
>> In the context of data management in edge computing, you may be
>> interested by research done by the Discovery Initiative:
>> http://beyondtheclouds.github.io/publications.html
>> I am adding Adrien Lèbre to the thread, he may be able to point you to
>> the publications most relevant to your whitepaper draft.
>>
>> Best wishes,
>> Pierre Riteau
>>
>> On Wed, 27 Feb 2019 at 10:21, Bogdan Dobrelya <bdobreli at redhat.com> wrote:
>>>
>>> On 27.02.2019 10:51, Stig Telfer wrote:
>>>> Hi All -
>>>>
>>>> We have an IRC meeting at 1100 UTC (about an hour’s time) in channel
>>>> #openstack-meeting. Everyone is welcome.
>>>>
>>>> Today’s agenda is here:
>>>>
>>>> https://wiki.openstack.org/wiki/Scientific_SIG#IRC_Meeting_February_27th_2019
>>>>
>>>> Today’s main event is that we have Ildikó Vancsa joining us to talk
>>>> about edge computing use cases and find overlap with Scientific SIG
>>>> use cases.
>>>>
>>>> Cheers,
>>>> Stig
>>>>
>>>>
>>>
>>> Edge computing is a *great* topic for lot of future research work indeed.
>>> Thank you for this useful announcement, appreciated!
>>> I'll attend at least the first half of it, and in case I'll miss the
>>> agenda item that I placed at the end of the list, please be kind to
>>> review that research request with my absence as well. Thanks!
>>>
>>>
>>> --
>>> Best regards,
>>> Bogdan Dobrelya,
>>> Irc #bogdando
>>>
>
--
Best regards,
Bogdan Dobrelya,
Irc #bogdando
More information about the openstack-discuss
mailing list