Version 10 (modified by Aaron Falk, 15 years ago) (diff)


Network RSpec Mini-workshop


Paraphrasing Larry's framework from 1, the goal of this meeting is to agree on the basic semantics and syntax of a minimal statement needed to reserve connectivity resources. I.e., this is not about discovery, capability expression, or configuration, although clearly these functions can potentially be added in the future. Also, the goal is to provide projects that need it something they can work with in this spiral. We can revise and extend later.


The GPO will prepare an initial proposal to serve as a basis for discussion. Several folks have expressed thoughts on this topic, including Rob, Larry, Ted, Max, and Chris. My plan is to prepare the following:

  • discussion of the requirements
  • analysis of the proposals
  • recommendation of a proposal to use as a starting point
  • identification of the missing pieces (i.e., work that needs to be done)

Anyone who has thoughts or cycles to contribute is welcome to collaborate, just let me know. Individual proposals are welcome for discussion as well.


9:00 Introduction Aaron Falk
Morning How do networks represent resources?
Enterprise GENI view Rob Sherwood
ProtoGENI CF view Rob Ricci
10:30 BREAK
PlanetLab CF view Larry Peterson
NDL for RSpecs Max Ott/Ivan Seskar
Noon - 1pm LUNCH
Afternoon How do networks slivers join to make slices?
DRAGON for PlanetLab networks Chris Tracy
2:30pm BREAK
Planning Aaron Falk
5pm Close


When: Thursday, June 25th, 9am - 5pm CDT
Where: Chicago O'Hare Hilton, Boardroom 2048
Who: Primarily projects with spiral 1 milestones related to this topic; GPO
Invited Projects: ProtoGENI, Dtunnels, PlanetLab, EnterpriseGENI, Mid-Atlantic, GpENI, SPP, ORCA/BEN, Regional Opt-In, Orbit, TIED, Internet2/GMOC
Confirmed attending:

Aaron Falk, GPO
Harry Mussman, GPO
Chris Small, GPO
Rob Ricci, ProtoGENI
Guido Appenzeller, Enterprise GENI
Rob Sherwood, Enterprise GENI
Camilo Viecco, GMOC
Larry Peterson, PlanetLab
Jon Duerig, ProtoGENI
Ilia Baldine, ORCA/BEN
Yufeng Xin, ORCA/BEN
Ivan Seskar, ORBIT
Yogesh Mundada, DTunnels
Nick Feamster, DTunnels
Ted Faber, TIED
Chris Tracy, MAX

Travel expenses will be paid by the GPO
Voice conference may be available


Meeting Introduction & Goals, Aaron Falk, GENI Project Office



Following up from post-GEC4 discussions, I'd like to propose a gathering of interested parties to agree on a Spiral 1 approach for network resource description. The DTunnels project is hitting this topic now and others will (or should) very soon. GENI prototyping projects are providing network resources using varied abstractions, some as topologies and others as individual links; both may be L1, L2 (true or tunneled), or L3 (true or tunneled). While each of these network abstractions have differences in their description, they share a common need of expressing connectivity. A key GENI capability is to interconnect networks so it is useful to encourage (but not require) early commonality in how network resources are described.

Note that there is a clear emphasis on wired networks here. Wireless networks share these issues but in Spiral 1 let's try to nail the low-hanging fruit.

Details are below. Rob Ricci has generously agreed to host this meeting, so we'll be in Salt Lake City. (Thanks, Rob!) Please complete the scheduling Doodle right away, as we need to set to the date quickly to identify the hotel and finalize the logistics. I do hope you can make it.

Best regards,


Attachments (7)