| 1 | = 4) GENI I&M Priority Design Topics = |
| 2 | |
| 3 | |
| 4 | == Description == |
| 5 | |
| 6 | These are the GENI I&M priority design topics that must be resolved to allow the current GENI I&M projects to proceed, particularly the large Sol 3 project (GIMI and GEMINI) [[BR]] |
| 7 | |
| 8 | These topics were identified and reviewed at the GEC12 I&M meeting. [[BR]] |
| 9 | |
| 10 | |
| 11 | == Goals == |
| 12 | |
| 13 | Resolve between now (GEC12) and GEC13. [[BR]] |
| 14 | |
| 15 | Simplify where possible! [[BR]] |
| 16 | |
| 17 | == Approach == |
| 18 | |
| 19 | Form teams for each topic, and identify a LEAD for each topic. [[BR] |
| 20 | |
| 21 | Teams should meet by confernce call, or in person, and complete a report for presentation at GEC13. [[BR]] |
| 22 | |
| 23 | |
| 24 | == 4.1) List of All Identified Topics == |
| 25 | |
| 26 | List: [[BR]] |
| 27 | |
| 28 | a) I&M Use Cases for Experiments [[BR]] |
| 29 | |
| 30 | b) I&M Use Cases for Infrastructure Measurement [[BR]] |
| 31 | |
| 32 | c) Platform Targets [[BR]] |
| 33 | |
| 34 | d) Experiment and Measurement Orchestration Tools [[BR]] |
| 35 | |
| 36 | e) Interfaces/Protocols Between Tools and Services [[BR]] |
| 37 | |
| 38 | f) Descriptor Schema and Registry Service [[BR]] |
| 39 | |
| 40 | g) Object Names and Registry Service [[BR]] |
| 41 | |
| 42 | h) XML Messaging Service [[BR]] |
| 43 | |
| 44 | i) Lookup Service [[BR]] |
| 45 | |
| 46 | j) Portal Service [[BR]] |
| 47 | |
| 48 | k) User Workspace and Archive Services [[BR]] |
| 49 | |
| 50 | l) Operational Services [[BR]] |
| 51 | |
| 52 | m) Services Shared by GIMI and GEMINI projects [[BR]] |
| 53 | |
| 54 | n) Support for Experimenters [[BR]] |
| 55 | |
| 56 | o) Support for Operators [[BR]] |
| 57 | |
| 58 | |
| 59 | These have been combined and summarized and shown in the sections below. [[BR]] |
| 60 | |
| 61 | |
| 62 | === 4.2) I&M Use Cases for Experiments, and Support for Experimenters === |
| 63 | I&M Use Cases for Experiments [[BR]] |
| 64 | Support for Experimenters [[BR]] |
| 65 | |
| 66 | 1) Definition of topic: [[BR]] |
| 67 | |
| 68 | a) How do we work towards Max Ott's vision for experiment support? [http://groups.geni.net/geni/attachment/wiki/GEC12InstMeasWorkingSession/Will%20RSpec%20help%20us%20to%20do%20better%20science%20-%20GEC11%20-%20July%202011.pdf slides] [[BR]] |
| 69 | |
| 70 | b) Nominal experiment use case: [[BR]] |
| 71 | 1) Setup and run experiment using Experiment Control Tools, Experiment Orchestration Service and Lab Portal. [[BR]] |
| 72 | 2) Setup and run measurements using services within slice, and persistent services. Option: get measurement data from persistent MP. [[BR]] |
| 73 | 3) Gather MD, and observe during experiment; formulate MDOD. [[BR]] |
| 74 | 4) Archive MD with MDOD. [[BR]] |
| 75 | 5) Share archived MD with others, per policy included within MDOD. [[BR]] |
| 76 | 6) Pull MD out of archive, to analyze and/or visualize. [[BR]] |
| 77 | |
| 78 | c) What support must be provided for Experimenters? how? [[BR]] |
| 79 | |
| 80 | 2) Who: Max Ott (NICTA); Jim Griffioen (U Kentucky); Prasad Calyam (Ohio Super Center); Mike Zink (UMass Amherst) [[BR]] |
| 81 | |
| 82 | |
| 83 | |
| 84 | === 4.3) I&M Use Cases for Infrastructure Measurement, and Support for Operators === |
| 85 | I&M Use Cases for Infrastructure Measurement [[BR]] |
| 86 | Support for Operators [[BR]] |
| 87 | |
| 88 | 1) Definition of topic: [[BR]] |
| 89 | |
| 90 | a) Delineate what infrastructure monitoring is and is not covered by this approach. [[BR]] |
| 91 | 1) Yes: monitoring of clusters/racks [[BR]] |
| 92 | 2) Yes: passive measurements of transport switches, etc. [[BR]] |
| 93 | 3) Yes: active measurements of IP networks, of Layer 2 and OpenFlow paths [[BR]] |
| 94 | |
| 95 | |
| 96 | b) Nominal infrastructure measurement process: [[BR]] |
| 97 | 1) Setup persistent or on-demand infrastructure measurement slice. [[BR]] |
| 98 | 2) Make passive measurements or make active measurements. [[BR]] |
| 99 | 3) Gather MD, and observe as it is gathered; formulate MDOD. [[BR]] |
| 100 | 4) Store MD in collector, describe with MDOD, and register MDOD so that MD can be shared. [[BR]] |
| 101 | 5) Typically share MD with Aggregate Operator, GMOC and/or Experimenters, per policy written into MDOD. [[BR]] |
| 102 | 6) Pull MD out of collector, analyze and visualize. [[BR]] |
| 103 | 7) Archive MD with MDOD. [[BR]] |
| 104 | 8) Share archived MD with others, per policy included within MDOD. [[BR]] |
| 105 | 9) Pull MD out of archive, to analyze and/or visualize. [[BR]] |
| 106 | |
| 107 | c) What support must be provided for Operator? how? [[BR]] |
| 108 | |
| 109 | 2) Who: Martin Swany (Indiana U); Eric Boyd (Internet2); Jason Zurawski (Internet2); Chris Small (Indiana U); Ilia Baldine, for ExoGENI racks (RENCI); ?, for InstaGENI racks (?)[[BR]] |
| 110 | |
| 111 | === 4.4) GENI Environment === |
| 112 | Platform Targets [[BR]] |
| 113 | Experiment and Measurement Orchestration Tools [[BR]] |
| 114 | Interfaces/Protocols Between Tools and Services [[BR]] |
| 115 | |
| 116 | 1) Definition of topic: [[BR]] |
| 117 | |
| 118 | a) Platform targets: [[BR]] |
| 119 | 1) protoGENI servers/VMs [[BR]] |
| 120 | 2) ORCA servers/VMs [[BR]] |
| 121 | 3) WiMAX sites [[BR]] |
| 122 | 4) ExoGENI (RENCI) racks; like ORCA [[BR]] |
| 123 | 5) TransCloud (HP) racks; like protoGENI? [[BR]] |
| 124 | 6) PlanetLab |
| 125 | |
| 126 | b) Experiment and Measurement Orchestration tools: [[BR]] |
| 127 | 1) FLACK [[BR]] |
| 128 | 2) GUSH (with Omni) [[BR]] |
| 129 | 3) SFACE + RAVEN [[BR]] |
| 130 | |
| 131 | c) Interfaces/protocols between tools and services: [[BR]] |
| 132 | 1) protoGENI API, with protoGENI RSpec v2, for resource assignment [[BR]] |
| 133 | 2) GENI AM API, with protoGENI RSpec v2, for resource assignment [[BR]] |
| 134 | 3) ns for topology [[BR]] |
| 135 | 4) ssh for image loading [[BR]] |
| 136 | 5) OMF messages for orchestration [[BR]] |
| 137 | |
| 138 | 2) Who: Martin Swany (IU); Jim Griffioen (U Kentucky); Ilia Baldine, for ExoGENI racks (RENCI); ?, for InstaGENI racks (?); David Irwin (UMass Amherst); Mike Zink (UMass Amherst); Christoph Dwertmann (NICTA); Tom Mitchell (GPO); Aaron Helsinger (GPO) [[BR]] |
| 139 | |
| 140 | |
| 141 | |
| 142 | === 4.5) Descriptors, Objects and Registries, and Lookup Service === |
| 143 | Descriptor Schema and Registry Service [[BR]] |
| 144 | Object Names and Registry Service [[BR]] |
| 145 | Lookup Service [[BR]] |
| 146 | |
| 147 | 1) Definition of topic: [[BR]] |
| 148 | |
| 149 | a) MDOD schema, for XML file. [http://groups.geni.net/geni/wiki/GEC11InstMeasWorkingSession#a2MeasurementDataObjectDescriptorMDOD References] [[BR]] |
| 150 | |
| 151 | b) MDOD registry. Use DOR registry? Include in iRODS? Consider IF-MAP server? [[BR]] |
| 152 | |
| 153 | c) Object (e.g., I&M service) names. [[BR]] |
| 154 | |
| 155 | d) Object registry, with names and public key. [[BR]] |
| 156 | |
| 157 | e) Home and global Lookup Service, as defined by perfSONAR [[BR]] |
| 158 | |
| 159 | f) Map MDOD into metadata that is registered [[BR]] |
| 160 | |
| 161 | g) Can Lookup Service be used for finding portals? [[BR]] |
| 162 | |
| 163 | h) Also Topology Service? following UNIS? [[BR]] |
| 164 | |
| 165 | 2) Who: Jason Zurawski (Internet2); Ezra Kissel (U Delaware); Eric Boyd (Internet2); Scott Jensen (Indiana U); Larry Lannom (CNRI); LEAD Giridhar Manepalli (CNRI); Deniz Gurken (UH); Aaron Helsinger (GPO) [[BR]] |
| 166 | |
| 167 | |
| 168 | |
| 169 | === 4.6) Messaging Service === |
| 170 | XML Messaging Service [[BR]] |
| 171 | |
| 172 | 1) Definition of topic: [[BR]] |
| 173 | |
| 174 | a) XMPP server, in public IP space. Consider AMQP server? Consider IF-MAP server? [[BR]] |
| 175 | |
| 176 | b) Entities connect, and are authenticated. [[BR]] |
| 177 | |
| 178 | c) An entity may start a pub/sub node. [[BR]] |
| 179 | |
| 180 | d) When an entity subscribes, a message is sent to publisher requesting authorization. [[BR]] |
| 181 | |
| 182 | e) Use for OMF messages. [[BR]] |
| 183 | |
| 184 | f) Use for "event record messages". [[BR]] |
| 185 | |
| 186 | 2) Who: Rudra Dutta (NCSU); Shu Huang (RENCI); Chirstoph Dwertmann (NICTA); Martin Swany (IU); Prasad Calyam (OSC); Deniz Gurken (UH) [[BR]] |
| 187 | |
| 188 | |
| 189 | === 4.7) Lookup Service === |
| 190 | Lookup Service [[BR]] |
| 191 | |
| 192 | Note: Merged into 4.5) [[BR]] |
| 193 | |
| 194 | |
| 195 | |
| 196 | === 4.8) Portal Service === |
| 197 | Portal Service [[BR]] |
| 198 | |
| 199 | 1) Definition of topic: [[BR]] |
| 200 | |
| 201 | a) Towards Max Ott's vision for experiment support [[BR]] |
| 202 | |
| 203 | b) Consider INSTOOLS portal service [[BR]] |
| 204 | |
| 205 | c) Define portal service for GIMI tools, with multiple functions. [[BR]] |
| 206 | |
| 207 | d) How are partitions for different users managed? [[BR]] |
| 208 | |
| 209 | e) Can these portal services be merged? |
| 210 | |
| 211 | f) Are these portal services really a combination of gateway and user workspace services?[[BR]] |
| 212 | |
| 213 | 2) Who: Jim Griffioen or ? (U Kentucky); Christoph Dwertmann (NICTA); Jason Zurawski, for NetKarma (Internet2) [[BR]] |
| 214 | |
| 215 | |
| 216 | === 4.9) User Workspace and Archive Services === |
| 217 | User Workspace and Archive Services [[BR]] |
| 218 | |
| 219 | 1) Definition of topic: [[BR]] |
| 220 | |
| 221 | a) Consider CNRI prototype user workspace and archive service [[BR]] |
| 222 | |
| 223 | b) Consider U Kentucky archive service [[BR]] |
| 224 | |
| 225 | c) Consider iRODS service [[BR]] |
| 226 | |
| 227 | d) How to move data to service: move file; move SQL DB; or? [[BR]] |
| 228 | |
| 229 | e) How are partitions for different users managed? [[BR]] |
| 230 | |
| 231 | f) How is data in MDOD used to set archive sharing policy? what is required in MDOD? [[BR]] |
| 232 | |
| 233 | 2) Who: Giridhar Manepalli (CNRI); Shu Huang (RENCI) ; Larry Lannom (CNRI); Wesley Wu (U Kentucky) [[BR]] |
| 234 | |
| 235 | |
| 236 | === 4.10 Operational Services === |
| 237 | Persistent Operational Services [[BR]] |
| 238 | Services Shared by GIMI and GEMINI [[BR]] |
| 239 | |
| 240 | 1) Definition of topic: [[BR]] |
| 241 | |
| 242 | a) What persistent operational services must be supported: [[BR]] |
| 243 | 1) Descriptor registry. DOR? iRODS? IF-MAP? [[BR]] |
| 244 | 2) Object registry. DOR? [[BR]] |
| 245 | 3) XML messaging service [[BR]] |
| 246 | 4) Home and global Lookup Services; also Topology Service? UNIS? [[BR]] |
| 247 | 5) INSTOOLS portal service [[BR]] |
| 248 | 6) GIMI portal service [[BR]] |
| 249 | 7) U Kentucky archive service [[BR]] |
| 250 | 8) CNRI prototype user workspace and archive service [[BR]] |
| 251 | 9) iRODs archive service [[BR]] |
| 252 | |
| 253 | b) What services can be shared by GIMI and GEMINI tools? [[BR]] |
| 254 | |
| 255 | c) How do these relate to other GENI services/functions, including clearinghouse? |
| 256 | |
| 257 | 2) Who: Mike Zink (UMass Amherst); Martin Swany (Inidana U); Jim Griffioen (U Kentucky); Giridhar Manepalli (CNRI); Aaron Helsinger (GPO) [[BR]] |