Date
1 - 7 of 7
Model-driven SAL and static mapping
|
Ken Gray <kgray@...>
Hey, folks. I had a curiosity question coming out of today's meeting.
One of my personal desires would be for the early daylight controller to engage the R&E community - which is an openflow thing. I'd like to explore if there was a simple way to map the NOX API set to OpenDaylight so that existing programs could just "run" ...as a bootstrap. Is there any merit and method to do this? |
|
|
|
Bhushan Kanekar (bkanekar) <bkanekar@...>
Without knowing the details of NOX API sets and how Appa use them and devil is on the details, but one can have a API translater App which takes in NOX API calls and maps them to ODL API and fulfills them by calling correspondening ODL API. Perhaps that App can be something a Univ can take up as a project.
toggle quoted message
Show quoted text
Thanks, Bhushan On Sep 26, 2013, at 11:39 AM, "Ken Gray" <kgray@...> wrote:
Hey, folks. I had a curiosity question coming out of today's meeting. |
|
|
|
Chris Wright <chrisw@...>
* Ken Gray (kgray@...) wrote:
Hey, folks. I had a curiosity question coming out of today's meeting.I like the idea of attracting existing apps, I don't know how much work is involved. IIRC, nox apps are either shared libraries (seems hard as you have to mock nox runtime+REST calls to ODL) or python (maybe easier if you "just" build py bindings to direct to ODL). Curious to hear what you find out. |
|
|
|
David Meyer <dmm@...>
On Thu, Sep 26, 2013 at 11:39 AM, Ken Gray <kgray@...> wrote:
Hey, folks. I had a curiosity question coming out of today's meeting.The idea is good (attract existing apps, etc). NOX seems to be a bit of a dying breed, however. So I'm not sure what the band-for-the-buck would be (are new applications being written to NOX?). Another thing is that NOX has limited expressiveness (see e.g. http://www.cs.princeton.edu/~jrex/talks/presto11.pptx), but that might be an advantage if you want to write a translator (or something like it); for example NOX programs won't be built of compositions of other NOX programs (NOX as a programming language doesn't really support composition), perhaps easing translation. --dmm
|
|
|
|
Ken Gray <kgray@...>
It's a fair question how much there is to salvage here or whether we need
toggle quoted message
Show quoted text
to do so at all - which is why getting feedback from R&E would be nice. My gut says that if your predecessor had started some work in NOX and you were riffing on it, you wouldn't want to rewrite to a new API - but I don't know how prevalent that is ... On the flip side, there are a couple of "production" R&E SDN nets out there, right ... Ophelia and INET2. May be a good idea to see if they're interested in porting? I'm supposed to speak at a BigTen users gig in Chicago next week and the InCentre guys are there giving an OF tutorial using ... Wait or it ... OpenDaylight. So, maybe there's some willingness. On 9/26/13 3:44 PM, "David Meyer" <dmm@...> wrote:
On Thu, Sep 26, 2013 at 11:39 AM, Ken Gray <kgray@...> wrote:Hey, folks. I had a curiosity question coming out of today's meeting.The idea is good (attract existing apps, etc). NOX seems to be a bit |
|
|
|
David Meyer <dmm@...>
On Thu, Sep 26, 2013 at 1:19 PM, Ken Gray <kgray@...> wrote:
It's a fair question how much there is to salvage here or whether we needThere's also pox (http://www.noxrepo.org/pox/about-pox/) --dmm
|
|
|
|
Colin Dixon <ckd@...>
I think POX is likely a better target since it's easier for us to write/create Python bindings for it's API calls and map them to OpenDaylight REST/MD-SAL calls than it would be to do the same thing with compiled C/C++. Just my two cents. |
|
|