service instance and temporal aspects of services
Hi all, I could not hear the later part of the previous session clearly, so maybe I miss the point, but: - if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path. - temporal aspects may exist not only for the connection service but also for other (to be defined) services. Tomohiro
Dear Tomohiro, We decided to take out the temporal aspects from the generic architecture and move it into the Connection Service discussion. The reason being that the NSI protocol message exchanges did not require a time synchronization. They mostly depend on state machines and relative timing. Inder On Jun 20, 2010, at 1:32 PM, Tomohiro Kudoh wrote:
Hi all,
I could not hear the later part of the previous session clearly, so maybe I miss the point, but:
- if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path.
- temporal aspects may exist not only for the connection service but also for other (to be defined) services.
Tomohiro
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
Hi Inder, But in the future, we may have services like point-to-multipoint (i.e. virtual network) service, which will have temporal aspects too. So, I believe "temporal aspects" is generic to (not all but) multiple services, and putting it to the generic architecture part is better for future extension of NSI. Tomohiro On Sun, 20 Jun 2010 13:41:48 -0700 Inder Monga <imonga@es.net> wrote:
Dear Tomohiro,
We decided to take out the temporal aspects from the generic architecture and move it into the Connection Service discussion. The reason being that the NSI protocol message exchanges did not require a time synchronization. They mostly depend on state machines and relative timing.
Inder
On Jun 20, 2010, at 1:32 PM, Tomohiro Kudoh wrote:
Hi all,
I could not hear the later part of the previous session clearly, so maybe I miss the point, but:
- if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path.
- temporal aspects may exist not only for the connection service but also for other (to be defined) services.
Tomohiro
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
Could you describe the point-to-multipoint use-case a bit specifically where it relates to temporal aspect? Is it joins and removes? Inder On Jun 20, 2010, at 1:51 PM, Tomohiro Kudoh wrote:
Hi Inder,
But in the future, we may have services like point-to-multipoint (i.e. virtual network) service, which will have temporal aspects too.
So, I believe "temporal aspects" is generic to (not all but) multiple services, and putting it to the generic architecture part is better for future extension of NSI.
Tomohiro
On Sun, 20 Jun 2010 13:41:48 -0700 Inder Monga <imonga@es.net> wrote:
Dear Tomohiro,
We decided to take out the temporal aspects from the generic architecture and move it into the Connection Service discussion. The reason being that the NSI protocol message exchanges did not require a time synchronization. They mostly depend on state machines and relative timing.
Inder
On Jun 20, 2010, at 1:32 PM, Tomohiro Kudoh wrote:
Hi all,
I could not hear the later part of the previous session clearly, so maybe I miss the point, but:
- if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path.
- temporal aspects may exist not only for the connection service but also for other (to be defined) services.
Tomohiro
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
It is a scheduling service of a network which includes multiple edge points and switching function, (that is, provide a virtual network). May be such services can be first classified into reservation (or scheduling) service class, and the connection service is a sub-class of the reservation service class. And the temporal aspects is a matter of the reservation service class. Tomohiro On Sun, 20 Jun 2010 13:55:48 -0700 Inder Monga <imonga@es.net> wrote:
Could you describe the point-to-multipoint use-case a bit specifically where it relates to temporal aspect? Is it joins and removes?
Inder
On Jun 20, 2010, at 1:51 PM, Tomohiro Kudoh wrote:
Hi Inder,
But in the future, we may have services like point-to-multipoint (i.e. virtual network) service, which will have temporal aspects too.
So, I believe "temporal aspects" is generic to (not all but) multiple services, and putting it to the generic architecture part is better for future extension of NSI.
Tomohiro
On Sun, 20 Jun 2010 13:41:48 -0700 Inder Monga <imonga@es.net> wrote:
Dear Tomohiro,
We decided to take out the temporal aspects from the generic architecture and move it into the Connection Service discussion. The reason being that the NSI protocol message exchanges did not require a time synchronization. They mostly depend on state machines and relative timing.
Inder
On Jun 20, 2010, at 1:32 PM, Tomohiro Kudoh wrote:
Hi all,
I could not hear the later part of the previous session clearly, so maybe I miss the point, but:
- if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path.
- temporal aspects may exist not only for the connection service but also for other (to be defined) services.
Tomohiro
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
So the hesitation was to make maintaining NTP like time in the NSA's as mandatory. Each service, like a connection service, might have a requirement to implement NTP to provide a reservation svc. Iner On Jun 20, 2010, at 2:00 PM, Tomohiro Kudoh wrote:
It is a scheduling service of a network which includes multiple edge points and switching function, (that is, provide a virtual network).
May be such services can be first classified into reservation (or scheduling) service class, and the connection service is a sub-class of the reservation service class. And the temporal aspects is a matter of the reservation service class.
Tomohiro
On Sun, 20 Jun 2010 13:55:48 -0700 Inder Monga <imonga@es.net> wrote:
Could you describe the point-to-multipoint use-case a bit specifically where it relates to temporal aspect? Is it joins and removes?
Inder
On Jun 20, 2010, at 1:51 PM, Tomohiro Kudoh wrote:
Hi Inder,
But in the future, we may have services like point-to-multipoint (i.e. virtual network) service, which will have temporal aspects too.
So, I believe "temporal aspects" is generic to (not all but) multiple services, and putting it to the generic architecture part is better for future extension of NSI.
Tomohiro
On Sun, 20 Jun 2010 13:41:48 -0700 Inder Monga <imonga@es.net> wrote:
Dear Tomohiro,
We decided to take out the temporal aspects from the generic architecture and move it into the Connection Service discussion. The reason being that the NSI protocol message exchanges did not require a time synchronization. They mostly depend on state machines and relative timing.
Inder
On Jun 20, 2010, at 1:32 PM, Tomohiro Kudoh wrote:
Hi all,
I could not hear the later part of the previous session clearly, so maybe I miss the point, but:
- if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path.
- temporal aspects may exist not only for the connection service but also for other (to be defined) services.
Tomohiro
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
ok. understood and agreed. so, I think we need to define the reservation (or scheduling) service class in the future (if not now). Which may also be useful for provisioning a cloud type hybrid environment. Tomohiro On Sun, 20 Jun 2010 14:30:20 -0700 Inder Monga <imonga@es.net> wrote:
So the hesitation was to make maintaining NTP like time in the NSA's as mandatory. Each service, like a connection service, might have a requirement to implement NTP to provide a reservation svc.
Iner
On Jun 20, 2010, at 2:00 PM, Tomohiro Kudoh wrote:
It is a scheduling service of a network which includes multiple edge points and switching function, (that is, provide a virtual network).
May be such services can be first classified into reservation (or scheduling) service class, and the connection service is a sub-class of the reservation service class. And the temporal aspects is a matter of the reservation service class.
Tomohiro
On Sun, 20 Jun 2010 13:55:48 -0700 Inder Monga <imonga@es.net> wrote:
Could you describe the point-to-multipoint use-case a bit specifically where it relates to temporal aspect? Is it joins and removes?
Inder
On Jun 20, 2010, at 1:51 PM, Tomohiro Kudoh wrote:
Hi Inder,
But in the future, we may have services like point-to-multipoint (i.e. virtual network) service, which will have temporal aspects too.
So, I believe "temporal aspects" is generic to (not all but) multiple services, and putting it to the generic architecture part is better for future extension of NSI.
Tomohiro
On Sun, 20 Jun 2010 13:41:48 -0700 Inder Monga <imonga@es.net> wrote:
Dear Tomohiro,
We decided to take out the temporal aspects from the generic architecture and move it into the Connection Service discussion. The reason being that the NSI protocol message exchanges did not require a time synchronization. They mostly depend on state machines and relative timing.
Inder
On Jun 20, 2010, at 1:32 PM, Tomohiro Kudoh wrote:
Hi all,
I could not hear the later part of the previous session clearly, so maybe I miss the point, but:
- if request path is hierarchical, service instances also have hierarchical relationship. That is, a service instance is created for each provider along the path.
- temporal aspects may exist not only for the connection service but also for other (to be defined) services.
Tomohiro
_______________________________________________ nsi-wg mailing list nsi-wg@ogf.org http://www.ogf.org/mailman/listinfo/nsi-wg
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
--- Inder Monga http://100gbs.lbl.gov imonga@es.net http://www.es.net (510) 499 8065 (c) (510) 486 6531 (o)
participants (2)
-
Inder Monga -
Tomohiro Kudoh