Changes between Version 9 and Version 10 of Gec5ClusterBAgenda
- Timestamp:
- 07/22/09 15:40:19 (15 years ago)
Legend:
- Unmodified
- Added
- Removed
- Modified
-
Gec5ClusterBAgenda
v9 v10 40 40 41 41 ---- 42 42 43 ''The following notes were taken by Christopher Small, GPO System Engineer'' 43 44 … … 45 46 46 47 == To Do List == 47 48 * Larry Peterson 48 49 === Larry Peterson === 49 50 50 51 Plan for this meeting is to bring each other up to speed on what we're doing. … … 56 57 Three ways to federate. 57 58 58 SFA only -- WSDL interface specification, you implement it. This is59 * SFA only -- WSDL interface specification, you implement it. This is 59 60 what Enterprise GENI has done. Generated this based on the running 60 61 code, but plan to go forward using the WSDL spec as the base and we'll … … 64 65 Enterprise GENI is using this. 65 66 66 SFA + geniwrapper -- includes security. Write your own back end to67 * SFA + geniwrapper -- includes security. Write your own back end to 67 68 geniwrapper. If you've got a preexisting testbed, this is probably the 68 69 fastest way to go; use the security machinery and interface spec, and 69 70 have it call into your system. 70 71 71 SFA + geniwrapper + MyPLC -- includes PlanetLab O&M machinery72 * SFA + geniwrapper + MyPLC -- includes PlanetLab O&M machinery 72 73 73 74 RSpecs are the place where we've always put things that we don't know … … 95 96 Did demo of PlanetLab with MAX nodes. 96 97 97 * James Sterbenz 98 === James Sterbenz === 98 99 99 100 GpENI demo, showed that Lancaster University UK, ETH-Zurich are … … 101 102 their own PLC at KSU. Separare PLC, separate VINI, not federated. 102 103 103 * Jeannie Albrecht 104 === Jeannie Albrecht === 104 105 105 106 XML description of the experiment. Command line program reads the xml, 106 107 sets up and runs the experiment. 107 108 108 * Chris Tracy 109 === Chris Tracy === 109 110 110 111 Describes the recursive aggregate manager model they have running, … … 115 116 ----- 116 117 117 = Tuesday 1:00-3:00118 119 * Larry Peterson 118 == Tuesday 1:00-3:00 == 119 120 === Larry Peterson === 120 121 121 122 We need to work on attracting real users … … 125 126 Let's advertise our new cluster B aggregates to those users 126 127 127 * James Sterbenz 128 === James Sterbenz === 128 129 129 130 Are we just adding users to have users? Are you suggesting we 130 131 advertise our resources as standard PlanetLab nodes? 131 132 132 * Larry Peterson 133 === Larry Peterson === 133 134 134 135 Control Framework todo … … 148 149 manager. 149 150 150 * Guido Appenzeller 151 === Guido Appenzeller === 151 152 152 153 Eventually we'd like to have something that abstracts some of the 153 154 details away, a web interface, to make this all easier to deal with. 154 155 155 * Larry Peterson 156 === Larry Peterson === 156 157 157 158 Chris Tracy put up a recursive aggregate manager, this may be what you … … 170 171 (or set of flat) 171 172 172 * Guido Appenzeller 173 === Guido Appenzeller === 173 174 174 175 Don't we need some kind of pluggable user interface to manage 175 176 heterogeneous aggregates? 176 177 177 * Larry Peterson 178 === Larry Peterson === 178 179 179 180 That's what Chris Tracy demonstrated, this is what we need. The SFI … … 185 186 We've just broken the clearinghouse model. 186 187 187 * Peter O'Neil 188 === Peter O'Neil === 188 189 189 190 From a usability standpoint this will be chaos. Need a single portal. 190 191 191 * Larry Peterson 192 === Larry Peterson === 192 193 193 194 There's not just one portal to the internet. … … 196 197 thing, or is it a loose federation of entities? 197 198 198 * Guido Appenzeller 199 === Guido Appenzeller === 199 200 200 201 Seems to fit into GENI goals to have one centralized clearinghouse. 201 202 202 * Larry Peterson 203 === Larry Peterson === 203 204 204 205 What do you mean by GENI? 205 206 206 * Guido Appenzeller 207 === Guido Appenzeller === 207 208 208 209 The internet has DNS, it tells me where I can look for things 209 210 210 * Larry Peterson 211 === Larry Peterson === 211 212 212 213 we can have a root registry and if we want to kill a slice we can … … 227 228 Policies will only be as sophisticated as the mechanisms. 228 229 229 * Guido Appenzeller 230 === Guido Appenzeller === 230 231 231 232 We only talk with known entities. We talk to an SM / clearinghouse, 232 233 not a user. A researcher hand-editing XML sounds ambitious. 233 234 234 * Larry Peterson 235 === Larry Peterson === 235 236 236 237 Writing an SM that spoke to a hererogeneous collection of aggregates … … 250 251 Can somebody explain this to me? 251 252 252 * James Sterbenz 253 === James Sterbenz === 253 254 254 255 Because the GPO told us to. 255 256 256 * James Kempf (?) 257 === James Kempf (?) === 257 258 258 259 You might want dedicated bandwidth 259 260 260 * Larry Peterson 261 === Larry Peterson === 261 262 262 263 IP encapsulation is only 28 bytes, it's running over layer 2. 263 264 264 * Guido Appenzeller 265 === Guido Appenzeller === 265 266 266 267 We want to give people hop-by-hop layer 2 so they can put a router in … … 268 269 them, not have an IP header. 269 270 270 * Larry Peterson 271 === Larry Peterson === 271 272 272 273 My argument is cost effectiveness. This will be a large investment of 273 274 time and effort with minimal value. 274 275 275 * Jon Turner 276 === Jon Turner === 276 277 277 278 Both are legitimate; you can get to more places with IP, layer 2 VLANs 278 279 will let you get closer to the bare network. 279 280 280 * Guido Appenzeller 281 === Guido Appenzeller === 281 282 282 283 Every slice must support IP fallback 283 284 284 * Larry Peterson 285 === Larry Peterson === 285 286 286 287 We haven't solved the problem because we still need to demux between … … 294 295 fairly low on user-level tools. 295 296 296 * Jeannie Albrecht 297 === Jeannie Albrecht === 297 298 298 299 Once you have a stable version of SFI, I can try to make things 299 300 prettier. But it doesn't make sense until SFI stabilizes. 300 301 301 * Larry Peterson 302 === Larry Peterson === 302 303 303 304 Look at the VINI rspec. … … 308 309 Goal for SFI: readable/editable RSpecs e.g. using XSGR 309 310 310 * Chris Tracy 311 === Chris Tracy === 311 312 312 313 We have some tools that might be usable 313 314 314 * Larry Peterson 315 === Larry Peterson === 315 316 316 317 Show of hands -- who has time for a GUI? 317 318 318 * Guido Appenzeller 319 === Guido Appenzeller === 319 320 320 321 What we see as the hard part of the GUI is how to help the user if the … … 322 323 accepted. 323 324 324 * Larry Peterson 325 === Larry Peterson === 325 326 326 327 Will try for pretty text plus policy feedback. 327
