Meeting Introduction: Integrating GENI Networks with Control Frameworks

1. Meeting Introduction:Integrating GENI Networks with Control Frameworks Aaron Falk GENI Project Office June 25, 2009

2. Who’s Here? • Rob Ricci, ProtoGENI • Jon Duerig, ProtoGENI • Guido Appenzeller, Enterprise GENI • Rod Sherwood, Enterprise GENI • Larry Peterson, PlanetLab • Ilia Baldine, ORCA/BEN • Yufeng Xin, ORCA/BEN • Yogesh Mundada, DTunnels • Nick Feamster, DTunnels • Ted Faber, TIED • Ivan Seskar, ORBIT/WiMAX • Chris Tracy, GMOC • Camilo Viecco, GMOC • Harry Mussman, GPO • Chris Small, GPO • Chip Elliott, GPO • Aaron Falk, GPO By phone: • Max Ott, ORBIT • James Sterbenz, GpENI • Heidi Picher-Dempsey, GPO

3. Key Goals for GENI Spiral 1Drive down critical technical risks in GENI’s concept Create my slice GENI Clearinghouse Goal #1 Fund multiple, competing teams to develop GENI Clearinghouse technology, encourage strong competition within the first few spirals Goal #2 Demonstrate end-to-end slices across representative samples of the major substrates / technologies envisioned in GENI Components Components Components Aggregate A Computer Cluster Aggregate B Backbone Net Aggregate C Metro Wireless

4. Goals for this Meeting • For each cluster answer: • How does a network device or aggregate represent/reserve resources? • How do network slivers join to form an end-to-end slice? • Produce concrete examples, plans

5. Network Configuration Use Case (slightly updated excerpt from GEC2 presentation)

6. Sliver Creation: Computation Resources (1 of 3) Compute Cluster Storage GbE 10 GbE Measurement Aggregate Manager Network Interface GID Slice & User Registry • I want to reserve all the pieces to • build my experiment, I will start with • a CPU cluster, reserving the following • 300 Processors • 10 TB Local storage (multiple disks) • 1 GbE and 10 GbE link capacity from network interface to CPU’s and 10TB drive • CPU utilization measurement w/ data transfer rates 1. Researcher submits credentials and request to aggregate manager (AM) for resources, that includes a reservation time 2. CM checks policies against credentialsand accepts reservation by returning signed RSpec to User (called a ticket). NSF GENI clearinghouse Resource Discovery Service CPU Cluster 3a. AM sends schedule update with reservation information (resources and dates) R Regional Research CPU Cluster R 3b. AM sends copy of ticket to Slice Registry (who tracks resources in each slice). Optical Backbone Optical Edge Storage Server A similar set of actions are performed for the other CPU cluster and storage array R Metro Wireless Access At this stage, the researcher has a right to use specific resources (i.e., establish slivers) from several of the aggregates. However, these resources are not active and have not been composed into a coherent experiment.

7. Sliver Creation: Aggregate Interconnection (2 of 3) GID Now that my edge processor resources are reserved, I need to establish data link connections across the multiple network domains. But, how do I set-up connections between two aggregates? Processing Center Compute Cluster Storage 1 2 Measurement 3 The reservation specifies network bandwidth between the 3 components and Optical Backbone. The Processing Center AM handles establishing requested internal network configuration. Connectivity to OB already exists. When the AM issues a ticket, the signed Rspec will contain interconnection parameters of each network reservation, e.g., port number, VLAN ID, source Ethernet address, so the OB AM can associate network resources in PC to those in OB. GbE Aggregate Manager 10 GbE GIMS Aggregate Manager This is repeated for the Regional Research, Storage Server , Optical Edge (and Metro Wireless) networks. Optical Transport Optical Edge/ Metro Wireless Regional Research Optical Backbone Measurement At this stage, the researcher has obtained payload mapping information from each of the aggregate managers. No connections have been established Storage Server

8. Sliver Creation: Networking Resources (3 of 3) GID Slice & User Registry • Now that aggregate interconnection • mapping is established, I want to reserve • the following resources: • Optical multipoint topology from storage server to PC,OE, RR network interfaces • FPGA framers on linecards mapped to 10GbE payload at each aggregate interface • 1 GbE link capacity between Processing center and other aggregate interfaces • BER and OSNR measurements on all links • Measurement data transfer rates Processing Center GIMS Optical Backbone Optical Transport Optical Edge Regional Research 10GbE Aggregate Manager Measurement 1. Researcher submits credentials, aggregate interconnection maps and request to aggregate manager (AM) for resources, that includes a reservation time Storage Server 2. AM checks policies against credentials and accepts reservation by returning signed RSpec to User 3a. AM maps payloads to interfaces and provides information in ticket returned to user R CPU Cluster NSF GENI clearinghouse R Regional Research R CPU Cluster Resource Discovery Service 3b. AM sends status update with reservation information Optical Backbone R R 3c. AM sends copy of ticket to Slice Registry R Optical Edge Storage Server At this stage, a complete slice reservation exists. However, until that reservation is exercised (i.e., tickets redeemed), active slivers cannot be programmed. R Metro Wireless Access

9. Questions • Is a standard method of describing network “coordinates” required? • Does it go into the RSpec? • Are there now constraints on the order in which networks can be added to a slice? • How does work with multiple networks in series? • How are ordering constraints handled in the control framework? • How are tunnels (i.e., non-adjacent resources) handled?

10. Today’s Agenda (ready to bash…)

11. Ground rules for today • Spiral 1 focus (i.e., by this fall) • Different answers are OK for different clusters • Shared responsibility between control framework & aggregate projects is required to solve this • This is an engineering, not research, discussion