Initial investigation result of federation issues

[Sangjin Jeong]

Hello folks,

We have been investigating the federation issues recently and got some
initial results.
We plan to prepare a documentation, but it would be helpful to solicit
reviews/comments
from the WG before drafting the document. We have requested a
presentation slot in the
control framework WG during GEC6, but the agenda is already set and it
is not sure
if there will be some time to present this initial result.

So, we would like to solicit comments from the mailing list.
The slides can be found from http://cclab.cnu.ac.kr/~yhchoi/etri/federation.pdf

Any comment/review would be appreciated.

Regards,
Sangjin

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg

[Robert P Ricci]

On Slide 11, you mention different possible levels of federation.

There's a federation going within the US for sites based on the Emulab
software. So far, we have very simple policies for giving access to
members of that federation, but it seems likely to me that the policy
issues might need to get more complicated when federating over
international boundaries. Do you have any thoughts on the types of
policies you would want to enforce on, say, US-based researchers, that
might be different from policies for local Korean researchers?

Thus spake Sangjin Jeong on Wed, Nov 11, 2009 at 08:26:03PM +0900:

> Hello folks,
>
> We have been investigating the federation issues recently and got some
> initial results.
> We plan to prepare a documentation, but it would be helpful to solicit
> reviews/comments
> from the WG before drafting the document. We have requested a
> presentation slot in the
> control framework WG during GEC6, but the agenda is already set and it
> is not sure
> if there will be some time to present this initial result.
>
> So, we would like to solicit comments from the mailing list.
> The slides can be found from http://cclab.cnu.ac.kr/~yhchoi/etri/federation.pdf
>
> Any comment/review would be appreciated.
>
> Regards,
> Sangjin
>
> _______________________________________________
> control-wg mailing list
> control-wg@...
> http://lists.geni.net/mailman/listinfo/control-wg

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg

[Schwab, Stephen-2]

Sangjin, Rob --
Following this line of thought further -- perhaps the interesting thing
that Sangjin and his colleagues might be able to share are the policies
that they have already been using on their own testbed -- or the
policies they would *like* to apply, both within their own local
testbed, across various Korean testbeds that might be viewed as a
Korean-GENI, and at the larger-scale of GENI-FIRE-Korea-Japan...

To be very concrete, how can Sangjin's testbed connect to the current
protoGENI/Internet2 infrastructure? Is there anything special we might
have to do regarding allocation policy for the network path between an
Internet2 PoP and the wide-area research network Sangjin's testbed
connects to?

Thanks,

--Steve



-----Original Message-----
From: control-wg-bounces@... [mailto:control-wg-bounces@...]
On Behalf Of Robert P Ricci
Sent: Thursday, November 12, 2009 8:34 AM
To: Sangjin Jeong
Cc: control-wg@...
Subject: Re: [cwg] Initial investigation result of federation issues

On Slide 11, you mention different possible levels of federation.

There's a federation going within the US for sites based on the Emulab
software. So far, we have very simple policies for giving access to
members of that federation, but it seems likely to me that the policy
issues might need to get more complicated when federating over
international boundaries. Do you have any thoughts on the types of
policies you would want to enforce on, say, US-based researchers, that
might be different from policies for local Korean researchers?

Thus spake Sangjin Jeong on Wed, Nov 11, 2009 at 08:26:03PM +0900:

> Hello folks,
>
> We have been investigating the federation issues recently and got some
> initial results.
> We plan to prepare a documentation, but it would be helpful to solicit
> reviews/comments
> from the WG before drafting the document. We have requested a
> presentation slot in the
> control framework WG during GEC6, but the agenda is already set and it
> is not sure
> if there will be some time to present this initial result.
>
> So, we would like to solicit comments from the mailing list.
> The slides can be found from

http://cclab.cnu.ac.kr/~yhchoi/etri/federation.pdf

>
> Any comment/review would be appreciated.
>
> Regards,
> Sangjin
>
> _______________________________________________
> control-wg mailing list
> control-wg@...
> http://lists.geni.net/mailman/listinfo/control-wg

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg

[Larry Peterson-2]

PlanetLab has had to address federation policies issues with our
international partners. Supporting different policies is a barrier-to-entry
for many people that want to contribute resources to the global effort.
To this end, we have developed a policy engine and an interface that
lets admins define policies for their aggregates. I'm attaching a draft
paper that describes our approach. Comments are most welcome.
(You can also find the code at svn.planet-lab.org.)

Larry

On Wed, Nov 11, 2009 at 6:26 AM, Sangjin Jeong <sjjeong@...> wrote:

Hello folks,

We have been investigating the federation issues recently and got some
initial results.
We plan to prepare a documentation, but it would be helpful to solicit
reviews/comments
from the WG before drafting the document. We have requested a
presentation slot in the
control framework WG during GEC6, but the agenda is already set and it
is not sure
if there will be some time to present this initial result.

So, we would like to solicit comments from the mailing list.
The slides can be found from http://cclab.cnu.ac.kr/~yhchoi/etri/federation.pdf

Any comment/review would be appreciated.

Regards,
Sangjin

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg

[Sangjin Jeong]

Hi Rob,

On Thu, Nov 12, 2009 at 10:34 PM, Robert P Ricci <ricci@...> wrote:
> On Slide 11, you mention different possible levels of federation.
>
> There's a federation going within the US for sites based on the Emulab
> software. So far, we have very simple policies for giving access to
> members of that federation, but it seems likely to me that the policy
> issues might need to get more complicated when federating over
> international boundaries. Do you have any thoughts on the types of
> policies you would want to enforce on, say, US-based researchers, that
> might be different from policies for local Korean researchers?

In the early stage of testbeds federation, it will be likely to enforce very
simple policy, ex) allow/disallow foreign users to access local testbed.
However, at the later stage, the administrators of each testbed may
tend to apply local policy and to give a restriction to foreign users.
These restrictions would be determined by policy negotiations such as
SLAs between testbeds.
It is necessary to preform more investigation in order to decide what kinds
of policy to be enforced, IMHO.

Regards,
Sangjin

>
> Thus spake Sangjin Jeong on Wed, Nov 11, 2009 at 08:26:03PM +0900:
>> Hello folks,
>>
>> We have been investigating the federation issues recently and got some
>> initial results.
>> We plan to prepare a documentation, but it would be helpful to solicit
>> reviews/comments
>> from the WG before drafting the document. We have requested a
>> presentation slot in the
>> control framework WG during GEC6, but the agenda is already set and it
>> is not sure
>> if there will be some time to present this initial result.
>>
>> So, we would like to solicit comments from the mailing list.
>> The slides can be found from http://cclab.cnu.ac.kr/~yhchoi/etri/federation.pdf
>>
>> Any comment/review would be appreciated.
>>
>> Regards,
>> Sangjin
>>
>> _______________________________________________
>> control-wg mailing list
>> control-wg@...
>> http://lists.geni.net/mailman/listinfo/control-wg
>

_______________________________________________
control-wg mailing list
control-wg@...
http://lists.geni.net/mailman/listinfo/control-wg