wiki:ClusterD

Version 72 (modified by hmussman@bbn.com, 15 years ago) (diff)

--

Cluster D Integration

Projects

ORCA/BEN and ORCA Augmentation
DOME
ViSE
KanseiSensorNet (KANSEI)
Embedded Real-Time Measurements (ERM)
Data Intensive Cloud Control (DICLOUD)
OKGems: Cyber-Physical System (OKGEMS)
Integrated Measurement Framework (IMF)
iGENI Distributed Network Research Infrastructure (IGENI)
LEARN Network (LEARN)
BBN ORCA Node

Meetings and Demos

ORCA-fest May 28, 2009, web site
Cluster D Conference Call on June 11, 2009
Cluster D Review Meeting July 2, 2009
"ORCA/BEN Integration Demo with Stitching of VLANs between BEN and NLR", July 7, 2009
"ORCA/BEN Integration Demo July 7, 2009, web site
Cluster D Meetings at GEC5 July 20 and 21, 2009

References and Documents

GENI - ORCA web site, including ORCA code and ticket system
Subscribe to Cluster D mailing list (cluster-d@renci.org), for PIs and senior personnel, to discuss administrative cluster-wide issues
Subscribe to CLuster D mailing list (geni-orca-users@renci.org), primarily for developers, to discuss technical issues about the implementation, etc.
Prototype DOME controller and handler modules for ORCA, 2009
Spiral 1 Plan Document, on 7/02/09
Spiral 2 Roadmap slides, with task ratings, on 7/21/09
Cluster D Roadmap Document, on 10/15/09
Cluster D Roadmap Drawings, on 10/15/09

Roadmap for Spiral 2

Goals

The following goals have been established by the GPO for Spiral 2.

Goal 1: Live experiments, the central goal of Spiral 2. Support significant numbers of research experiments in the end-to-end prototype systems. The GPO expects live experimentation to begin near the end of Spiral 1, which will intensify through Spiral 2 as we begin continuous operation of the prototype systems. This will begin to give us all substantial (early) operational experience, as these experiments will help us all understand the prototypes' strengths and weakness, which will drive our Spiral 3 goals.

Goal 2: Identity management

Goal 3: Improved integration of data & control planes, within clusters

Goal 4: Instrumentation

Goal 5: Interoperability, permitting clusters to access the widest number of aggregate

Key Dates

Begin 9/1/09 or 10/1/09, depending on the project
ORCA Rel 1.3 10/1/2009
GEC6 11/16/09
ORCA Rel 2.1 2/1/10
GEC7 3/16/10
ORCA Rel 2.2 6/1/10
GEC8 7/20/10
End 8/31/10 or 9/30/10, depending on the project
ORCA Rel 2.3 10/1/10

ORCA Software Releases

Roadmap

The ORCA software releases are currently scheduled for every 4 mo. , and include work done by both Solicitation 1 (1582) and Solicitation 2 (1700) ORCA projects.

ORCA Rel 1.3 10/1/2009
ORCA Rel 2.1 2/1/10
ORCA Rel 2.2 6/1/10
ORCA Rel 2.3 10/1/10

The following roadmap describes features and capabilities of ORCA software releases.



Figure 1-6: Slice controller API on the service manager fig 1-6

Figure 1-5: Identity provider verifies login fig 1-5

Status

1) For latest status, go to GENI - ORCA, web site maintained by RENCI, including ORCA code and ticketing system or contact Ilia Baldine Ilia Baldine at RENCI

Issues

1) Revoking / Canceling Leases
With ORCA, is there a method for leases to be canceled, or for outstanding requests to be revoked? This is required, or at least very highly desired, for DOME. Some scenarios:

  • An experiment is started on the buses and the researcher

begins getting erroneous results, or the experiment starts generating too much data (this happens). We'd like to be able to prematurely terminate a lease to stop the experiment, even if the allocated resources are not returned to the pool.

  • An experiment is scheduled for the future, but then the

researcher discovers he or she wants to change the experiment or alter the time.
Obviously, there are many other scenarios that make the functionality desirable.

Cluster D Operating Capabilities and Clearinghouse

Roadmap

The following roadmap describes features and capabilities of the Cluster D operating environment, including a production-level Cluster D clearinghouse.



Figure 1-1: Cluster D Configuration at start of Spiral 2 fig 1-1

Figure 1-2: Message flows between actors fig 1-2

Figure 1-3: Parallel local and clearinghouse brokers fig 1-3

Figure 1-4: Experiment from remote or local service manager, and clearnghouse broker fig 1-4

Status

For access to Cluster D Clearinghouse, go to Cluster D Clearinghouse at RENCI ; get login from Ilia Baldine

1) Clearinghouse at RENCI: Yes?

2a) BEN and VM brokers in clearinghouse: Yes?

2b) DOME broker in clearinghouse: Yes?

2c) ViSE broker in clearinghouse: Yes?

2d) Kansei broker in clearinghouse: Yes?

2e) BBN broker in clearinghouse: No?

Figure 1-1: Cluster D Configuration at start of Spiral 2 fig 1-1

Issues

Cluster D VLAN Capabilities

Roadmap

The following roadmap describes Cluster D VLAN capabilities.


Status

1) "All S1 aggregates connect to Internet2 (or NLR) with link capable of Layer 2 connection" Due 10/1/09

1a) BEN and VMs connected to NLR; see BEN Connectivity

1b) DOME connected to Internet2, or NLR; see DOME Connectivity

1c) ViSE connected to Internet2, or NLR; see ViSE

1d) Kansei not connected?; see Kansei Connectivity

1e) BBN VMs connected to NLR?; see ?

2) "Cluster D completes plan for the setup of VLANs between testbeds, to be carried by Internet 2 (or NLR) backbone network between the testbeds" Due 11/16/09, GEC6

See: clusterdvlan Cluster D VLAN connection plan?

Issues ====-

Cluster D Experiment Capabilities

Roadmap

The following roadmap describes Cluster D experiment capabilities.


Figure 1-3: Parallel local and clearinghouse brokers
fig 1-3

Figure 1-4: Experiment from remote or local service manager, and clearnghouse broker
fig 1-4

Figure 1-7: Experiment with two aggregates
fig 1-7

Status

1) "All S1 aggregates ready for experiments by outside researchers" Due 10/1/09

1a) BEN experiments by ERM project?

1b) DOME experiments by Undergraduate networking class in fall 2009?

1c) ViSE experiments using rapidly deployable node?

1d) Kansei experiments by internal researchers?

1e) BBN VMs experiments by?

2) "All S1 aggregates ready for experiments by researchers using remote SMs and CH brokers." Due 11/16/09, GEC6

Issues

2a) For each S1 aggregate, how can remote portal and SM be built?

2b) DOME and Federated Scheduling, i.e., towards an experiment that utilizes resources from multiple testbeds

Attachments (23)