[HamWAN PSDR] Do we have any cloud / distributed storage experts?

Eric Wolak ag6ie at wolak.net
Tue Feb 28 20:11:28 PST 2017


Hey Bart, 

I'm new here, but I figure now's as good a chance as any to contribute.
I have quite a bit of experience with distributed systems, and I'd love
to help. Let's talk reconciliation semantics!

On Tue, Feb 28, 2017, at 01:19 PM, Bart Kus wrote:
> Well, I can follow up with Subhro off-list.  I have more detailed 
> questions than that.  Was just looking for volunteers to help out with 
> design.
> 
> --Bart
> 
> 
> On 2/28/2017 1:18 PM, Nigel Vander Houwen wrote:
> > We’re looking for recommendations on what to use, and how to set it up. As Bart mentioned, the critical aspects of this are N-way redundancy so that any given fragment of the network if it gets separated can still operate, and then graceful handling of re-syncing after the network rejoins.
> >
> > Nigel
> >
> >> On Feb 28, 2017, at 13:15, Subhro Kar <subhro at 80386.org> wrote:
> >>
> >> I have some experience with distributed file systems. What do you need to know?
> >> Subhro
> >>
> >> Sent from my iPhone
> >>
> >>> On Feb 28, 2017, at 12:52, Bart Kus <me at bartk.us> wrote:
> >>>
> >>> Hello,
> >>>
> >>> Lately I've been working on network automation stuff for HamWAN, and I'm at the point where storage systems have to be dealt with.  I'm wondering if anyone in the group has knowledge to provide some design feedback in distributed storage systems.  Not from a "how to use it in AWS" perspective, but rather how to design them ground-up.
> >>>
> >>> Ideally we want a couple abstractions available network-wide:
> >>>
> >>> 1) A global DB (NoSQL should be fine, and Cassandra is looking possible for this)
> >>> 2) A global FS (GlusterFS is in very early stages of being evaluated for this)
> >>>
> >>> Each of these has to have N-way redundancy where N is the # of sites we have, with each instance being capable of continuous operation even if it finds itself on its own network island for a while. Conflict resolution / merging when the network heals is a must.
> >>>
> >>> --Bart
> >>>
> >>> _______________________________________________
> >>> PSDR mailing list
> >>> PSDR at hamwan.org
> >>> http://mail.hamwan.net/mailman/listinfo/psdr
> >> _______________________________________________
> >> PSDR mailing list
> >> PSDR at hamwan.org
> >> http://mail.hamwan.net/mailman/listinfo/psdr
> > _______________________________________________
> > PSDR mailing list
> > PSDR at hamwan.org
> > http://mail.hamwan.net/mailman/listinfo/psdr
> 
> _______________________________________________
> PSDR mailing list
> PSDR at hamwan.org
> http://mail.hamwan.net/mailman/listinfo/psdr


More information about the PSDR mailing list