AutoGOLE + NSI demo at Supercomputing
Here are some slides how I would like to set up the demos for NSI v2 and Automated GOLE. We can discuss on NSI call ... Jerry
Hi, It is unfortunate that opennsa does not support v2 yet. This means that several important GOLEs can not participate in the v2 demo. But it also means that my scheduling and querying tools do not work. These tools would need to be reimplemented for v2 by someone else if we want the automated earth and AIST monitor to work. A simpler solution would be to see if we can make a v1 - v2 proxy, so that we can do a complete demonstration. I understand it is not desirable to have v1 proxying, and this should be disabled after SC. Jeroen. On 31 okt. 2012, at 03:27, Jerry Sobieski <jerry@nordu.net> wrote:
Here are some slides how I would like to set up the demos for NSI v2 and Automated GOLE.
We can discuss on NSI call ...
Jerry <Supercomputing2012-SLC AutomatedGOLE Update.pptx> <Supercomputing2012-SLC AutomatedGOLE Update.pptx.pdf>
Hi Jeroen, Jerry and all, It will be not easy to do two separate demonstrations. It will overload those who have developed v2 implementation, and it will be hard to understand what we are doing for the audiences. AIST decided to develop an aggregator which can talk to both v1 uPA and v2 uPA. With this aggregator, we will be able to have a unified demo. Jeroen, will you be able to develop a QA which will talk to AIST aggregator using v2 query? Tomohiro On Wed, 31 Oct 2012 09:05:19 +0100 Jeroen van der Ham <vdham@uva.nl> wrote:
Hi,
It is unfortunate that opennsa does not support v2 yet. This means that several important GOLEs can not participate in the v2 demo. But it also means that my scheduling and querying tools do not work. These tools would need to be reimplemented for v2 by someone else if we want the automated earth and AIST monitor to work.
A simpler solution would be to see if we can make a v1 - v2 proxy, so that we can do a complete demonstration.
I understand it is not desirable to have v1 proxying, and this should be disabled after SC.
Jeroen.
On 31 okt. 2012, at 03:27, Jerry Sobieski <jerry@nordu.net> wrote:
Here are some slides how I would like to set up the demos for NSI v2 and Automated GOLE.
We can discuss on NSI call ...
Jerry <Supercomputing2012-SLC AutomatedGOLE Update.pptx> <Supercomputing2012-SLC AutomatedGOLE Update.pptx.pdf>
Hi, On 31 okt. 2012, at 09:17, Tomohiro Kudoh <t.kudoh@aist.go.jp> wrote:
Jeroen, will you be able to develop a QA which will talk to AIST aggregator using v2 query?
I would need to have a look at the code. Perhaps Henrik can use his limited time to implement the query in v2 opennsa? Jeroen.
On Wed, 31 Oct 2012, Jeroen van der Ham wrote:
Jeroen, will you be able to develop a QA which will talk to AIST aggregator using v2 query?
I would need to have a look at the code.
Perhaps Henrik can use his limited time to implement the query in v2 opennsa?
If I get the time, but I cannot promish anything for SC :-|. Best regards, Henrik Henrik Thostrup Jensen <htj at nordu.net> Software Developer, NORDUnet
Hi Jerry, G-lambda-K also supports NSIv2. So, KDDILabs and JGN-X domains join NSIv2 demo. Takahiro On 2012/10/31 11:27, Jerry Sobieski wrote:
Here are some slides how I would like to set up the demos for NSI v2 and Automated GOLE.
We can discuss on NSI call ...
Jerry
-- Takahiro Miyamoto <tk-miyamoto@kddi.com> KDDI Corporation TEL: +81-80-5945-9004
OK it sounds like a few updates can be made: --- AIST / G-LAMBDA-A will act as an go between v1 and v2. So, all multidomain requests will be sent to AIST and it will do a one level tree segmentation and issue requests to each NSA as if they are a uPA. Question: Does G-LAMBDA-A process v2 NML topology? or v1 topology? or both? --- KDDI / G-LAMBDA-K will be added to the v2 sites --- Question to Jeroen: Is the NSI version announced in the NSA topology object? IMO, tt should be. This will make the job of v1/v2 differentiation much easier... Also, the uPA/Aggregator/uRA capabilities should also be announced in the NSA object. ?? --- Question to KISTI/ Jeonghoon: Will DynamicKL be v1 or v2? --- Question to OSCARS folks: Will OSCARS be able to cover, say, WIX in washington? (Brocade driver). This would be a good simple entrance for OSCARS into the AutoGOLE/NSI environment.. Jerry
participants (5)
-
Henrik Thostrup Jensen
-
Jeroen van der Ham
-
Jerry Sobieski
-
Takahiro Miyamoto
-
Tomohiro Kudoh