Changes between Initial Version and Version 1 of Gec8ClusterDAgenda

Timestamp:

07/14/10 12:14:53 (14 years ago)

Author:

hmussman@bbn.com

Comment:

--

Gec8ClusterDAgenda

@@ +1 @@
+[[PageOutline]]
+
+
+= Cluster-D Meetings at GEC8, Agenda and Notes  =
+
+==  Tuesday, July 22, 12noon - 4:30pm   == 
+
+==  Introductions (12 noon) ==
+
+==  Review of current status near the end of Spiral 2, and plans for Spiral 3  ==
+
+=== Spiral 2 and Spiral 3 Goals (12:10pm, Mussman)  ===
+
+The following goals have been established by the GPO for Spiral 2.  [[BR]]
+
+ 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. [[BR]]
+ Goal 2:  Identity management  [[BR]]
+ Goal 3:  Improved integration of data & control planes, within clusters  [[BR]]
+ Goal 4:  Instrumentation        [[BR]]
+ Goal 5:  Interoperability, permitting clusters to access the widest number of aggregate  [[BR]]
+
+For Spiral 3, these goals are reaffirmed, and particularly: [[BR]]
+
+ Goal 1:  Live experiments [[BR]]
+ Goal 2:  Identity management  [[BR]]
+ Goal 5:  Interoperability, permitting clusters to access the widest number of aggregate, including use of GENI AM API  [[BR]]
+
+=== ORCA Software Releases for Spirals 2 and 3  (Baldine, 12:20pm) ===
+
+==== 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. [[BR]]
+ ORCA Rel 1.3                   10/1/2009[[BR]]
+
+ ORCA Rel 2.0 (was 2.1)         2/1/10[[BR]]
+ ORCA Rel 2.1 (was 2.2)         6/1/10[[BR]]
+
+ ORCA Rel 2.2 (was 2.3)         10/1/10 [[BR]]
+ 
+ More in Spiral 3?[[BR]]
+ 
+ 
+
+
+==== Status ====
+
+Current progress on each feature, including: [[BR]]
+
+ Known issues [[BR]]
+
+ XML-RPC interface [[BR]]
+
+ Shibboleth interop [[BR]]
+
+
+Additional features under consideration for Spiral 3: [[BR]]
+
+
+
+
+
+=== ORCA/BEN ===
+
+Ilia Baldine [[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/GEC7-orca-ben-report.pdf slides]
+
+ORCA software releases, and features: [[BR]]
+
+Bella  2.0 [[BR]]
+Available now  (was 2.1, planned for Feb 1) [[BR]]
++  New substrate API obsoletes COD [[BR]]
++  Lease cancellation [[BR]]
++  Multiple redeem predecessor support [[BR]]
++  Eucalyptus support [[BR]]
++  New broker policies [[BR]]
++  XML-­‐RPC ProtoGENI-­‐like SM interface [[BR]]
++  Global actor registry XML-­‐RPC function [[BR]]
++  Extended NDL support [[BR]]
++  NLR Sherpa driver [[BR]]
+
+Bella 2.1 [[BR]]
+Panned for ?  (was 2.2, planned for June 1) [[BR]]
++  Code hardening, more unit tests, better documentation [[BR]]
++  Broker querying coupled with registry [[BR]]
++  Juniper EX3200 support [[BR]]
++  Multi‐point VLAN slice support (BEN and NLR) [[BR]]
++  Possible: [[BR]]
+– Support for newer Eucalyptus release [[BR]]
+– Internet2 IDC support [[BR]]
+– Gush integration (in cooperation with Gush team) [[BR]]
+
+===  ViSE ===
+
+David Irwin[[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/SliceUmassDuke.mov  UMass-Duke demo movie][[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/CASA_Firefox.mov  slice setup movie]
+
+
+=== DICloud  ===
+
+Michael Zink  [[BR]] 
+slides? [[BR]] 
+
+=== KANSEI  ===
+
+Mukundan Sridharan [[BR]] 
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/GEC7-KanseiStatus-Presentation.pdf  slides][[BR]] 
+
+Progress in Spiral 2:[[BR]] 
+1. Created v1.1 KanseiGenie Installer[[BR]] 
+2. Ported KanseiGenie to NetEye (Wayne State), and provided support to porting efforts at OKGems (Oklahoma State)[[BR]] 
+3. Extended KanseiGenie Researcher Portal to include baseline federated experimentation[[BR]] 
+4. Completed development to make the switch form Soap to SecureSoap.  New version will be rolled out after GEC7.[[BR]] 
+5. Started planning v2.0 of software installer package, including baseline federation[[BR]] 
+
+Milestones Completed:[[BR]] 
+KANSEI: S2.d Extend researcher portal for use with Kansei and NetEye (Due 3/16/10)[[BR]] 
+KANSEI: S2.e Bring up NetEye testbed at WSU (Due 03/16/10)[[BR]] 
+KANSEI: S2.f Connection plan from Kansei and NetEye to Internet2 (Due 03/16/10)[[BR]] 
+KANSEI: S2.h Demo experiment using Kansei and NetEye (partially done; Due 07/20/10)[[BR]] 
+
+=== OKGems  ===
+
+PIs: Xiaolin (Andy) Li, Nazanin Rahnavard, Weihua Sheng, Venkatesh Sarangan, Johnson Thomas[[BR]] 
+Students: Han Zhao, Xinxin Liu, Praveen Khethavath, Li Yu, Nhat Nguyen [[BR]] 
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/OKGems-GEC7-Final.pptx  slides][[BR]] 
+
+Hardware status:[[BR]] 
+Dedicated lab space [[BR]] 
+About 100 sensor nodes (mainly Telosb and Micaz)[[BR]] 
+About 20 robots [[BR]] 
+
+ORCA status:[[BR]] 
+Installed and configured ORCA (Anacortes 1.2) in local environment successfully. [[BR]] 
+Built one ORCA local master and one node agent for initial testing.[[BR]] 
+
+KanseiGenie status:[[BR]] 
+Set up a small set of lab sensor nodes into the KanseiGenie framework. [[BR]] 
+Customized the KanseiGenie web portal.[[BR]] 
+
+Robotics VLab status:[[BR]] 
+A client session, control a robotic arm for arranging an alphabetical pattern[[BR]] 
+
+Year 1 Plan 1:[[BR]] 
+Design and develop drivers for our resource aggregates (PCs and sensors) using ORCA Bella 2.0.[[BR]] 
+Set up VLAN to stitch distributed resources together.[[BR]] 
+Deploy field sensor nodes and integrate them into the OKGems framework.[[BR]] 
+Further design OKGems portal[[BR]] 
+Give a demo in GEC 8[[BR]] 
+
+Year 1 Plan 2:[[BR]] 
+Modify existing plugins and develop new plugins for ORCA, i.e., plugins of scheduling policies.[[BR]] 
+Develop spatial-temporal slivers of sensor nodes/networks through TinyHive virtual machine and TinyBee mobile agent. (to be fully completed in Year Two)[[BR]] 
+Define programmable abstractions in finer granularity [[BR]] 
+
+Year 1 Plan 3:[[BR]] 
+Draft a system design to include robots[[BR]] 
+Define initial programmable abstractions of actuators and effectors[[BR]] 
+Remotely control robots over networks (LAN & WAN)  [[BR]]   
+
+
+
+=== Bella  2.0 Release Features ===
+
+Aydan Aydan Yumerefendi [[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/NewOrcaFeatures-Aydan.pdf  slides] 
+
++  New Substrate API[[BR]]
+-  Simplify integration of new substrates[[BR]]
+-  Simplified IConcreteSet implementation[[BR]]
+.  UnitSet represents a collection of Unit-s[[BR]]
+.  Unit designed to represent any resource unit, e.g, VM, VLAN, Testbed, etc.[[BR]]
+-  IShirakoPlugin extension: ISubstrate[[BR]]
+.  Implemented as Substrate and AuthoritySubstrate[[BR]]
+.  Defines clearly the extension points for a new substrate[[BR]]
+.  transferIn, modify, transferOut[[BR]]
+-  New authority Resource Control classes[[BR]]
+.  UnitsControl[[BR]]
+.  SimpleVMControl[[BR]]
+.  VMControl[[BR]]
+.  VLANControl[[BR]]
+-  First steps to simplify and generalize the web portal[[BR]]
+.  New project: webapp2[[BR]]
+
++  New drivers and handlers: NLR, Eucalyptus[[BR]]
+-  NLR[[BR]]
+.  Provisions dynamically a path over NLR[[BR]]
+-  Eucalyptus[[BR]]
+.  Provisions dynamically a VM from an Eucalyptus cluster[[BR]]
+.  Can also provision from Amazon EC2 (same API)[[BR]]
+.  Supports assigning VLANs to VMs[[BR]]
+.  Requires a patch to Eucalyptus 1.5[[BR]]
+
++  Simplified actor configuration file format[[BR]]
+-  Define an actor with a few lines of XML[[BR]]
+   
++  Lease cancellation (vacate) support[[BR]]
+-  Brokers can allocate resources more efficiently[[BR]]
+
++  New controllers: inter-domain, xmlrpc (protogeni)[[BR]]
+ 
++  No major new features in the next release[[BR]]
+-  Primary focus:[[BR]]
+.  Harden and productize existing codebase[[BR]]
+.  Simplify installation and setup[[BR]]
+.  Documentation[[BR]]
+-  If time permits:[[BR]]
+.  SOAP management API[[BR]]
+.  Integrate Orca in command-line tools and other portals[[BR]]
+.  Distributed Orca without Tomcat[[BR]]
+
+=== Actor Registry and XML-­RPC Controller  ===
+
+Anirban Mandal[[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/GEC7-anirban.pdf  slides][[BR]]
+
++  ORCA Actor Registry[[BR]]
+-  Snapshot of available actors from various ORCA containers[[BR]]
+—  Users can use this for resource discovery and other queries[[BR]]
+-  Users can visit http://geni.renci.org/actors.php to view current set of actors registered with the registry[[BR]]
+-  Users can also programmatically query the registry using an XML-RPC interface[[BR]]
+
++  XML-RPC Controller[[BR]]
+-  Goal: Provide an XML-RPC interface to ORCA for experiment control tools and external users[[BR]]
+-  Tools/users would use this interface to[[BR]]
+.  discover available resources[[BR]]
+.  instantiate and destroy experiment slivers[[BR]]
+-  Initial version of ProtoGENI-like XML-RPC interface available on an ORCA SM[[BR]]
+-  Generic slice manager acts on behalf of experiment control tools/user[[BR]]
+-  Details of control framework hidden from user[[BR]]
+-  These interfaces provide a minimal set of features[[BR]]
+.  DiscoverResources(), CreateSliver(), DeleteSliver()[[BR]]
+-  Extended set of features (available later) would include[[BR]]
+.  GetTicket(), RedeemTicket(), UpdateTicket(), UpdateSliver()[[BR]]
+.  Support for resource representation using NDL[[BR]]
+
+== New NDL features in ORCA ==
+
+Yufeng Xin [attachment:gec7-yxin.pptx presentation]
+
+==  Day 2: March 17, 1pm - 3pm, Agenda and Notes ==
+
+
+=== ERM:  Embedded real-time substrate measurements for real-time cross-layer experiments === 
+
+Keren Bergman, Caroline Lai, Michael Wang [[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/GEC7_ClusterD_ERM.pptx  slides][[BR]]
+
+
++  Year 1: Unified measurement Framework architecture[[BR]]
++  Now:  get measurements from Polatis, Infinera, etc.[[BR]]
++  Next:  experiment use case for GEC8[[BR]]
+-  Control:[[BR]]
+.  Stand-aline programmable attentuator[[BR]]
+.  Polatis swithch[[BR]]
+
+
+
+
+=== LEARN Measurement Handler ===
+
+Debjyoti Majumder[[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/Learn%20%202%20%20MEASUREMENT_HANDLER.pptx  slides] LEARN measurement handler[[BR]]
+
++  Measurement handler:[[BR]]
+-  receives XML-RPC requests[[BR]]
+-  logs in to Polatis/Infinera via TL1[[BR]]
+-  gets measurements[[BR]]
+-  logs out
+
+=== LEARN Network Integration ===
+
+Shade EL-Hadik [[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/Learn%201%20%20ORCA_%20ppt.pptx  slides]  LEARN network
+
++  Create an isolated ORCA instance in the lab with a Cisco 3750 switch[[BR]]
+-  The ORCA Master is now equipped with handlers.[[BR]]
+-  The node agent has drivers installed.[[BR]]
+-  Drivers for accessing the Cisco 6509 switch are available, which we need to modify for Cisco 3750. [[BR]]
+-  Need to test if the handlers of the ORCA master are suitable for the existing drivers of the node agent. [[BR]]
+-  If successful, the Cisco 3750 switch needs to be accessed by node agent for our lab [[BR]]
+-  Finally to set up a VLAN using the switch, node agent, and the ORCA Master.[[BR]]
++  Goal:  Integrate the LEARN, regional optical network, located in Texas, into the ORCA control framework (GENI Cluster D) by GENI Engineering Conference 8 (July 2010)[[BR]]
+
+
+=== IMF ===
+
+Rudra Dutta 
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/imf_gec7_cluster-d.pdf slides] [[BR]]
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/GEC7_Measurment_Poster_v5.pdf  poster], including IMF, ERM and LEARN
+
++  IMF Overview  (see poster)[[BR]]
+-  IMF provides an abstraction of measurement capabilities available within the GENIsubstrate by allowing experimenters to measure properties within the substrate, such as physical layer attributes (e.g.optical or RF power) and performance parameters (e.g.BER, packet loss,CPU usage) through a unified measurement interface.[[BR]]
+
++  IMF Architecture (see poster)[[BR]]
+-  An experiment slice is created by instantiating slivers of resources from multiples sites and interconnecting them.[[BR]]
+-  The data plane of a slice reflects the topology desired by user.[[BR]]
+-  The measurement plane of a slice represents the measurement functions available within the slice.[[BR]]
+-  A slice also has substrate environment control capability so that the precise environment of a slice may be controlled and manipulated.[[BR]]
+
++  PubSub module included in IMF[[BR]]
+-  Knows about what it wants, polls MH[[BR]]
+-  Publishes results[[BR]]
+
++  MH module, included in IMF[[BR]]
+-  Simple driver, no state [[BR]]
+ 
++  Use Case[[BR]]
+-  Run SILO applications on ORCA-BEN.[[BR]]
+-  SILO: Services Integration, controL, and Optimization.
+
++  Next [[BR]]
+-  Run SILO – IMF for a complete closedloop feedback control demonstration[[BR]]
+-  Create SCM to mirror measurement capability[[BR]]
+-  Measure – control – monitor[[BR]]
+-  Down the road – better presenting capabilities to experimenter who only consumes[[BR]]
+-  Also examine API for in-slice consumer programmers[[BR]]
+-  Integrate with measurement ontology ?
+  
+
+=== DOME  === 
+
+Brian Lynn, Thijs de Vries  [[BR]] 
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/DOME-ClusterD-Overview.pdf  slides] [[BR]] 
+
+Since Sept, 2009, have several users running experiments on DOME [[BR]] 
++  UMass graduate students [[BR]] 
++  Research project with University of Arkansas [[BR]] 
++  NSF GENI Research Experience for Undergraduate (REU) program [[BR]] 
++  U of Washington’s Million Node GENI runs on DOME when no other experiments are scheduled [[BR]] 
+
+=== iGENI ===
+
+Joe Mambretti, Jim Chen   [[BR]] 
+[http://groups.geni.net/geni/attachment/wiki/Gec7ClusterDAgenda/iGENI%20Presentation%20GEC%20March%202010.ppt  slides]
+
++  "GENI Cluster D Network", or GCDNet [[BR]] 
++  iGENI as an aggregate [[BR]] 
+
++  Now [[BR]] 
+-  Installed 6509 at NW in Chicago [[BR]] 
+-  Initial Site to Site Implementation Demonstrated at GEC 7 [[BR]] 
+-  Connected UMass (via un-switched VLAN) to RENCI (via switched VLAN) [[BR]] 
+
++  Next
+-  Prototype/Model for Other Instantiations  [[BR]] 
+-  Planning Has Begun for Other Cluster-D Sites [[BR]] 
+-  Next Demonstration in July 2010 at GEC 8 [[BR]] 
+-  Planning for Cluster-D November Wash DC Showcase [[BR]] 
+
++  Goal[[BR]] 
+-  Bridge NLR and I2 soon, for a good reason[[BR]] 
+
+
+=== Plan for GEC8 demos/experiments ===
+ 
++  Questions:[[BR]] 
+-  Exactly where?[[BR]] 
+-  How much bandwidth?
+
++  ORCA-BEN[[BR]] 
+-  Multi‐point slice demo with several Cluster‐D members[[BR]] 
+-  Assist other projects with experiment demonstrations
+
++  OKGems:  robots?
+
++  Kansei:  Federation, with visualization?
+
+=== Plan for GEC9 demos/experiments ===
+ 
++  Questions:[[BR]] 
+-  Exactly where?[[BR]] 
+-  How much bandwidth?
+
++  Can Cluster D provide a featured demo?[[BR]] 
+-  What could it be?
+
+
+
+
+==  Ongoing Cluster D Interactions ==
+
+Monthly PI calls[[BR]]
+Calls on specific topics[[BR]]
+Resource description[[BR]]
+
+
+
+