| 1 | = Cluster C Review = |
| 2 | |
| 3 | ''July 8, 2009'' |
| 4 | |
| 5 | == Attendees == |
| 6 | |
| 7 | * Chip Elliot |
| 8 | * Aaron Falk |
| 9 | * Yao Luo |
| 10 | * Larry Lannom |
| 11 | * Paul Barford |
| 12 | * Evan Zhang |
| 13 | * Heidi Picher Dempsey |
| 14 | * Christopher Small (GPO) |
| 15 | * Jim Griffioen |
| 16 | * Justin Cappos |
| 17 | * Rob Ricci |
| 18 | * Vic Thomas |
| 19 | * Nick Feamster |
| 20 | |
| 21 | == Intro ''(Aaron and Chip)'' == |
| 22 | |
| 23 | Goals for today's meeting. |
| 24 | |
| 25 | '''Paul Barford''' This is a woefully underfunded project and has |
| 26 | amazing promise. What about some of the $3B stimulus money? |
| 27 | |
| 28 | == ProtoGENI ''(Rob Ricci)'' == |
| 29 | |
| 30 | Atlanta is the next most important PoP for Rob. |
| 31 | |
| 32 | Implementations for CH, CM, Slice Authority (all are fairly |
| 33 | functional); RSpec for creating slices with topology (within site |
| 34 | VLANs, between sites tunnels). Functional federation; can get slices |
| 35 | that span seven sites, with complex topology. Authentication and |
| 36 | authorization system. Backbone deployment progressing. |
| 37 | |
| 38 | Have worked with about 1/2 of all spiral 1 projects. |
| 39 | |
| 40 | * Some important pieces not in the control framework, e.g. cross-CM |
| 41 | coordination, relinquish tickets. |
| 42 | |
| 43 | * Lots of projects to coordinate with, not all integrations are as |
| 44 | tight as they would like. |
| 45 | |
| 46 | * Lots of bureaucratic hassles; University of Utah not used to this |
| 47 | kind of contract. |
| 48 | |
| 49 | The University has gotten very interested in the fact that they are |
| 50 | installing a separate (not controlled by university IT) backbone link; |
| 51 | concerned about security, privacy, and liability issues. Rob has |
| 52 | pointed them at Larry Peterson's document on doing this. GPO is |
| 53 | interested in their help in getting recommendations on how to do this. |
| 54 | |
| 55 | Simple clearinghouse up and running. Several registries, Trust anchor |
| 56 | for federation. Seven federates: Utah x 2 , Kentucky, Wisc, UML, BBN, |
| 57 | CMU. |
| 58 | |
| 59 | Backbone will be all layer 2 or below, VLAN based. Have been working |
| 60 | with Internet2, HP, IU, GPO to make this happen. Will provide transit for |
| 61 | SPP. Campus/regional out connections soon at Utah, Wisc, MAX, |
| 62 | GpENI. Early exporation with UKY, GATech, ISI. Talking with Internet2 about |
| 63 | DCN. |
| 64 | |
| 65 | At Univ Utah talking with UEN (Utah Educational Network); UEN is |
| 66 | building a metro fiber ring around SLC, hits campus and the Level 3 |
| 67 | PoP. Plan is to get 10Gbps wave on this fiber. |
| 68 | |
| 69 | '''Paul Barford''' Our CIO (Ron Kramer) represents WiscNet (regional entity in |
| 70 | Wisconsin). Wisc runs BOREAS-Net (Broadband Optical Research, Education and |
| 71 | Sciences Network) that spans three states, etc. He is very interested in GENI; |
| 72 | we could relatively easily put together a consortium of CIOs to put together |
| 73 | regional deployment -- not just Internet2 and NLR -- get a GENI wave across all of |
| 74 | the regional deployments. It will be work, but there are resources available |
| 75 | and the will to do it with this group, and GENI could dramatically expand the |
| 76 | footprint with little resources. |
| 77 | |
| 78 | '''Chip Elliot''' Maybe early fall we should put together a meeting. |
| 79 | |
| 80 | '''Heidi Picher Dempsey''' There is a working group in Quilt for GENI |
| 81 | that is already thinking about some of these issues. |
| 82 | |
| 83 | '''Aaron Falk''' Good insight that campus CIOs are influentional in |
| 84 | the regions, we should pursue this. |
| 85 | |
| 86 | I'm the primary PoC for all of the inter-project connections. |
| 87 | |
| 88 | Talking with Internet2 to get tail circuits (using DCN) to other |
| 89 | campuses. Internet2 views DCN connections as ephemeral (policy, not |
| 90 | technical), which may not be a good match for GENI. We might need to |
| 91 | convince them to let us have long-lived DCN connections. |
| 92 | |
| 93 | '''Chip Elliot''' How has working with Internet2 been? |
| 94 | |
| 95 | '''Justin Cappos''' UW has been connected since 1999. We aren't |
| 96 | thinking about layer 2 VLANs. |
| 97 | |
| 98 | '''Chip Elliot''' We'd love to have your campus GENI-enabled. Same |
| 99 | with UML. We'll gladly help. |
| 100 | |
| 101 | '''Aaron Falk''' GENI's utility goes up as the number of connected |
| 102 | users goes up; you can be advisors and advocates to campuses. |
| 103 | |
| 104 | '''Larry Lannom''' What is the simplest test for a campus being "GENI |
| 105 | enabled"? |
| 106 | |
| 107 | '''Chip Elliot''' If it can be controlled via ProtoGENI, it's GENI |
| 108 | enabled. It'd be great if you stood up a small ProtoGENI or Emulab |
| 109 | it'd be great. |
| 110 | |
| 111 | '''Rob Ricci''' We have shipped nodes to some of o the other campuses |
| 112 | in this cluster, and controlled from Utah -- little or no local |
| 113 | administration. |
| 114 | |
| 115 | '''Larry Lannom''' We're not on MAX or Internet2, we could be, given funds. |
| 116 | |
| 117 | '''Paul Barford''' One of the success stories of PlanetLab was their |
| 118 | very simple method of joining PlanetLab, simple setup, a couple of |
| 119 | machines on the internet -- if you give us two machines at your site, |
| 120 | you can be part of this great infrastructure. |
| 121 | |
| 122 | '''Justin Cappos''' This is part of our story -- you download this |
| 123 | software, you're part of Million-node GENI. |
| 124 | |
| 125 | '''Rob Ricci''' I have some money to buy some machines -- there are |
| 126 | some at Wisc -- and the reason we haven't gone further is that |
| 127 | coordination on the sending side takes a lot of effort. But we could |
| 128 | do more of it. |
| 129 | |
| 130 | '''Chip Elliot''' Let's work together on this. |
| 131 | |
| 132 | If DCN works out, all you have to do is turn it on. |
| 133 | |
| 134 | '''Heidi Picher Dempsey''' From what I've heard the software isn't |
| 135 | really there yet. |
| 136 | |
| 137 | '''Jim Griffioen''' The reason that it might be attractive to us is |
| 138 | that Internet2 comes to the regional, but not all the way to us. DCN |
| 139 | may be the easiest way for us to get a VLAN. |
| 140 | |
| 141 | '''Paul Barford''' We're planning on doing a lot of things with VLANs, |
| 142 | but VLANs are not a panacea. All kinds of weird things happen when you |
| 143 | start multiplexing over VLANs, stuff you can't imagine. The most |
| 144 | attractive option for me for GENI would be to have separate lambdas |
| 145 | for everyone. It's never been clear in my mind where this is all going |
| 146 | to go. |
| 147 | |
| 148 | '''Chip Elliot''' The GPO view is to go to VLANs on Spiral 1, but it's |
| 149 | not obvious that that's what we want to be using in Spiral 3. |
| 150 | |
| 151 | '''Heidi Picher Dempsey''' It's amazing how much push-back we're |
| 152 | getting trying to get people to use anything but IP! |
| 153 | |
| 154 | '''Paul Barford''' I'd really like to have vendors here at the table |
| 155 | telling us what is available, what their stuff does, so we can explore |
| 156 | more options. |
| 157 | |
| 158 | '''Chip Elliot''' That's a good idea, we're heading there I think. |
| 159 | |
| 160 | ProtoGENI is (theoretically) available to a few thousand |
| 161 | experimenters, but they don't know it. I've been holding off on |
| 162 | recruiting users waiting to hear on our round 2 proposal, largely for |
| 163 | interface issues; GENI is less than what Emulab users are used to. |
| 164 | |
| 165 | The control framework API and tools, as we've defined it, is not |
| 166 | enough for what users need. We need more services and basic setup |
| 167 | stuff for researchers to be successful. |
| 168 | |
| 169 | '''Paul Barford''' Our experience with WAIL is that without good |
| 170 | tools, the learning curve is too long for people to pick it up. Need a |
| 171 | front-end that makes the experience for users simpler. Just got |
| 172 | Cybertrust funding for a tool for Emulab that helps you repeat a |
| 173 | configuration, publish an experiment, provides pre-configured |
| 174 | environments for different types of experiments (e.g. honeynet, |
| 175 | botnet, ...). If there is any way to do this in GENI so a student or a |
| 176 | researcher could sit down at a dashboard and push some buttons and get |
| 177 | an experiment running, it'll really help get GENI off the ground. |
| 178 | |
| 179 | '''Aaron Falk''' We're completely in sync with this. It wasn't in |
| 180 | spiral 1, we're thinking about it for spiral 2. |
| 181 | |
| 182 | '''Jim Griffioen''' The interface is difficult to use, and little |
| 183 | documentation, so it's pretty difficult to talk about instrumentation. |
| 184 | |
| 185 | '''Jim Griffioen''' I see the backbone evolving as the PlanetLab |
| 186 | backbone did. A lot of campuses are on Internet2, but a lot are on the |
| 187 | regular internet. The backbone is connectivity no matter how you get |
| 188 | it. A lot of us will try to get on the wave links, if we can. |
| 189 | |
| 190 | Once our switches get deployed in Internet2 PoPs, we'll manage them |
| 191 | the same way we do in Emulab, we won't overprovision them. |
| 192 | |
| 193 | '''Chip Elliot''' Right now we have a lambda from NLR and a lambda |
| 194 | from Internet2, and in the next few weeks we'll start attaching |
| 195 | equipment on it. Some people want to run repeatable experiments, and |
| 196 | they'll migrate off the internet onto this. |
| 197 | |
| 198 | The only real difference between Emulab and DETER is that DETER doesn't |
| 199 | provide publicly routable addresses; nodes have to be behind a NAT. |
| 200 | |
| 201 | '''Paul Barford''' Containment is a seurious issue. The community needs |
| 202 | to think about what it means to run larger scale security |
| 203 | experiments. Arguably, the community doesn't really know how to do |
| 204 | repeatable experiments under laboratory conditions of this kind. We |
| 205 | need to think carefully about what is required to bootstrap |
| 206 | experiments, and how to run them. |
| 207 | |
| 208 | '''Heidi Picher Dempsey''' There is this tension between "you can |
| 209 | connect anywhere via the internet, but you don't know how you're |
| 210 | connected", and "here is a fixed, repeatable network topology, but |
| 211 | it's hard to get onto it." |
| 212 | |
| 213 | === Milestones === |
| 214 | |
| 215 | Integrate, install 3 nodes in Internet2 -- waiting for switches, not arrived, |
| 216 | this date will slip. |
| 217 | |
| 218 | Next two milestones are blocked on this one, but should fall out as soon as |
| 219 | the switches are up. |
| 220 | |
| 221 | Support experiments on ProtoGENI cluster -- we're already there, except for |
| 222 | not running on the Internet2 backbone. And last milestone is also well underway. |
| 223 | |
| 224 | == Instrumentation Tools ''(Jim Griffioen)'' == |
| 225 | |
| 226 | About 70 machines available, will be more useful after we get OpenVZ (a |
| 227 | virtualization environment) on them. They are connected to ProtoGENI. |
| 228 | |
| 229 | Some of the software we're running doesn't understand virtual interfaces |
| 230 | (e.g. tcpdump). We need more experience with that. |
| 231 | |
| 232 | We've had some conversations with the GMOC about what we're measuring and |
| 233 | how. We're trying to grab experiment-specific data; they are interested in |
| 234 | global information. |
| 235 | |
| 236 | One problem with the API is that it's designed around users, not groups. If |
| 237 | I'm teaching a class, I have students, I have a TA, want some sharing, some |
| 238 | restrictions. |
| 239 | |
| 240 | '''Chip Elliot''' CF working group should talk about this. |
| 241 | |
| 242 | I have a paper -- Heidi has a pointer to it. |
| 243 | |
| 244 | UK now has a 10Gbps Internet2 link, will be putting in a second "research" |
| 245 | link to Internet2 (which will be shared with other departments). No VLAN |
| 246 | support planned at this time (although it might be possible over the |
| 247 | new link). Right now using GRE tunnels. Have discussed implementing |
| 248 | DCN with Internet2 and KRON. |
| 249 | |
| 250 | '''Aaron Falk''' Dynamic VLANs are not a requirement; having static |
| 251 | VLANs you can run non-IP layer 2 experiments, which would be great. |
| 252 | |
| 253 | Lots of researchers have access via ProtoGENI -- although most of them |
| 254 | don't know it. Not ready for class use yet, still gets rebooted and |
| 255 | reinstalled every couple of weeks. (Would like to use it for operating |
| 256 | systems/distributed systems course in spiral 2.) Want to use it with |
| 257 | our FIND research. |
| 258 | |
| 259 | '''Chip Elliot''' It would be great to give tutorials to researchers |
| 260 | (how to use it), professors (how to teach classes on it). |
| 261 | |
| 262 | '''Paul Barford''' Been teaching a network systems course in WAIL for |
| 263 | years, it could be generalized to work on GENI. It's been very |
| 264 | popular. Anyone who wants it is welcome to it. Of the 50 people who |
| 265 | have taken this class, 20 have been hired by Cisco. |
| 266 | |
| 267 | === Milestones === |
| 268 | |
| 269 | 2nd milestone: Jim went the extra mile to update his software and be |
| 270 | more interactive with ProtoGENI, integrate more often. So although the |
| 271 | milestone was a little bit late, the effort was above and beyond what |
| 272 | was agreed to, so the minor slippage is not an issue. |
| 273 | |
| 274 | == DTunnels ''(Nick Feamster)'' == |
| 275 | |
| 276 | Design Requirements for BGPMux |
| 277 | |
| 278 | Sesison transparency (BGP looks like adirect conect), session |
| 279 | stability (no transient behavior visible to upstream nets), isolation |
| 280 | (individual networks able to set own policies, foward independently), |
| 281 | scalability (support many networks). |
| 282 | |
| 283 | The last two are still in progress. The first two are being used in |
| 284 | the prototype now. |
| 285 | |
| 286 | Reviewed BGP mux functions in slightly more detail---we've seen this |
| 287 | before, so I'm not making notes on it. |
| 288 | |
| 289 | Year 1 progress: |
| 290 | |
| 291 | * DTunnels: Kernel patches to create Ethernet GRE tunnels Interface for |
| 292 | specifying topology in XML and instantiating topology in OpenVS |
| 293 | (ProtoGENI nodes will run) Still need to discuss RSPECs more to make |
| 294 | BGPMux RSPEC fit in with ProtoGENI RSPECs |
| 295 | |
| 296 | * BGP Mux: Design and implementatio of control plane Deployment in three |
| 297 | locations: GT, Wisc, PSG. Will demo data plane at GEC5 |
| 298 | |
| 299 | '''Paul Barford''' Do you used differnt AS numbers for each site? |
| 300 | |
| 301 | No, we are using one AS for all sites. Challenges were a combination |
| 302 | of logistical for things like AS number, and hacking for GRETunnels |
| 303 | and IPTables mapping. |
| 304 | |
| 305 | Integration Progress in Spiral 1: |
| 306 | |
| 307 | * DTunnels: Instantiate topology from an XML spec that resembls and RSPEC initial |
| 308 | discussions about how to make RSPEC and XML spec have more in common |
| 309 | |
| 310 | * BGP Mux: Install Mux nodes on same subnets as ProtoGENI nodes. Will get two |
| 311 | PG nodes assigned to BGPMux and then try running the BGPMux on them. |
| 312 | Would like GEC5 to demo this. |
| 313 | |
| 314 | No RSpec integration yet; need an AM |
| 315 | |
| 316 | * I2 progress: I2 is accepting BGP advertisements from Georgia Tech, PSGNet (not |
| 317 | Wisconsin) Soeme filtering on commodity links due to rwois/IRR |
| 318 | problems |
| 319 | |
| 320 | DTunnels I2 connections doesn't really apply, but we do have the |
| 321 | option of previsioning between deployed sites. Just using GRETunnels |
| 322 | now. |
| 323 | |
| 324 | First Experiment with Jen Rexford at Princeton, NameCast. CNS |
| 325 | resolvers at multiple sites and advertised on a common IP prefix. |
| 326 | Service advertises and withdraws BGP routes to control how traffic |
| 327 | reaches the service. Deployment in progress at Georgia Tech. This |
| 328 | experiment was designed for VINI, but there's no VINI at Wisconsin, so |
| 329 | need to work on ProtoGENI. |
| 330 | |
| 331 | Next experiment will be somethign that requires virtual |
| 332 | networks/tunneling behind the mux. |
| 333 | |
| 334 | Have to be careful about things like advertising BGPMux as a transit |
| 335 | net and causing trouble in the rest of the network when this is |
| 336 | deployed and used by many people, not just friends and family. |
| 337 | |
| 338 | * Spiral 2 plans: |
| 339 | * More experiments |
| 340 | * Integrate AM with ProtoGENI |
| 341 | * Automate BGP Mux setup from RSpec |
| 342 | * Integrate with other control frameworks |
| 343 | |
| 344 | Some discussion about status of BGPMux milestones. Some are being |
| 345 | changed, so wiki status will be revised when Aaron and Nick close that |
| 346 | out. |
| 347 | |
| 348 | == Measurement System ''(Paul Barford)'' == |
| 349 | |
| 350 | GIMS aggregate manager controls the instrumentation plane. Basic model |
| 351 | is that there are multiple sensors, managed by the GIMS aggregate |
| 352 | manager. Researchers select which sensors they want to use in their |
| 353 | experiment, which kinds of packets to gather in the experiment. |
| 354 | |
| 355 | Sensor can use a high performance packet capture card (cost about $4K |
| 356 | for the card, provide a GPS timestamp). There is a new Intel NIC that |
| 357 | comes on the motherboard of new machines; it splits streams across |
| 358 | multiple cores, may be able to do the work we need for substantially |
| 359 | less. Not generally programmable, but does what we need. Total system |
| 360 | cost under $2K per sensor node. Old NICs start dropping packets at |
| 361 | 300-400 pps. |
| 362 | |
| 363 | Will demo at GEC5. |
| 364 | |
| 365 | High management overhead + low publication opportunity makes the |
| 366 | program a challenge. |
| 367 | |
| 368 | There is a fluidity to the GENI architecture which makes designing for |
| 369 | it difficult. |
| 370 | |
| 371 | We did a redesign of the GIMS architecture and the GIMS sensor |
| 372 | architecture. |
| 373 | |
| 374 | Upgraded WAIL to GENI. WAIL users can access GIMS, GENI, and GENI |
| 375 | users access to WAIL. GIMS UI extension for PRotoGENI underway, will |
| 376 | demo at GEC5. |
| 377 | |
| 378 | No WAN connections to date. NLR will support GIMS deployment, have not |
| 379 | officially solicited Internet2. |
| 380 | |
| 381 | Spiral 2, want to expand suite, move from prototype to v1.0, and |
| 382 | deploy GIMS on GENI (with minimal scope). |
| 383 | |
| 384 | Daryl Veitch is working with him on timing. High-perf packet capture |
| 385 | cards come with GPS capabiity. |
| 386 | |
| 387 | Chips says GENI could engineer a tier 1 stratum. Million node GENI |
| 388 | would like to use it if it existed. Someone would have to propose it |
| 389 | and get it funded of course. Paul noted that he and Daryl are |
| 390 | skeptical about how well current stratum 1s are actually synchronized. |
| 391 | |
| 392 | Chip said we will have to address the lack of ethernet frame info in |
| 393 | this project, since GENI is being built based on eth VLANs for |
| 394 | prototypes. |
| 395 | |
| 396 | Paul said his work is a complement to Jim's. Jim might be able to use |
| 397 | pub/sub interface to GIMS. Difference has to do with Paul owning the |
| 398 | sensor and Jim not. May need some kind of shim to make Jim's stuff |
| 399 | able to use it. |
| 400 | |
| 401 | Other measurement infrastructures that have been deployed: NEMEY, |
| 402 | Surveyor, laely move towards end-host application measurements, but |
| 403 | Paul is trying to cover the in-network measurements. |
| 404 | |
| 405 | Chip and Paul discussed the need for better funding for building |
| 406 | testbeds. Need to convince NSF that it needs better funding. Should |
| 407 | we have the meeting of a dozen blue-ribbon people who are |
| 408 | participating in testbeds and make recommendations to NSF. If a dozen |
| 409 | of leading people in the US are saying it is needed, maybe they will |
| 410 | listen. |
| 411 | |
| 412 | Larry points out that the digital library people had the same problem |
| 413 | 10 years ago and they failed. The GRID people have "money running out |
| 414 | their ears" for this kind of thing. |
| 415 | |
| 416 | Jason said what about MRI funding? Can only submit 3 proposals per |
| 417 | campus. |
| 418 | |
| 419 | GENI architecture is complicated and fluid. Coming up with measurement |
| 420 | AM was a new approach ad harder than the original idea. Redesigns |
| 421 | will still be required in the future. |
| 422 | |
| 423 | WAIL has been upgraded to ProtoGENI API. |
| 424 | |
| 425 | NLR deployment was original plan, haven't talked to I2 about deployment yet. |
| 426 | |
| 427 | Can build interfaces to other projects outside the cluster. Sensors |
| 428 | and AM are designed to be general and extensible. |
| 429 | |
| 430 | Working closely with BGP mux project. Paul will try to come up with |
| 431 | experiments to measure what BGP mux is doing to integrate more |
| 432 | closely. |
| 433 | |
| 434 | SEER wrapper from DETER (Schwab) might be extended to be useful with |
| 435 | GIMS for experimenter. SEER is built on emulab APIs. Could port to |
| 436 | ProtoGENI API instead. |
| 437 | |
| 438 | Probably Y2 deployments are likely to be off Gigabit links nearer to |
| 439 | campus edges than cores. |
| 440 | |
| 441 | |
| 442 | === Milestones === |
| 443 | |
| 444 | * Develop set of specifications. It's kind of a general description or |
| 445 | narrative. |
| 446 | |
| 447 | * Develop test suite. Harpoon-based suite, complete. |
| 448 | |
| 449 | * Develop prototype that can be deployed and demonstrated in WAIL. Done, will |
| 450 | demo at GEC5. |
| 451 | |
| 452 | Looks like basically all done. The specifications need some refinement. |
| 453 | |
| 454 | == Programmable Edge Node (PEN) ''(Yan Luo)'' == |
| 455 | |
| 456 | In year 1, acquired and assembled a PEN multi-core server with |
| 457 | network-processor-based acceleration card (Netronome NFE-i8000). |
| 458 | Acquired switch, PCs. Integrated with Emulab/ProtoGENI. Use Netronome |
| 459 | Flow Manager to establish up to 256 virtual NICs. |
| 460 | |
| 461 | PEN modified the DB to give it the illustion that it has several |
| 462 | physical nodes, even though they are actually virtual nodes. The rest |
| 463 | of the work could be done on the client side. |
| 464 | |
| 465 | Integration was a challenge. Emulab source base is huge, and not all |
| 466 | that completely documented. |
| 467 | |
| 468 | '''Rob Ricci''' A big (250 page) document is coming. |
| 469 | |
| 470 | === Spiral 2 goals === |
| 471 | |
| 472 | Want deployable PEN node ready (regression test, bill of materials, user |
| 473 | guide, wiki) by 10/01/09. Work with University of Utah to deploy two PENs in |
| 474 | Internet2 backbone. Enhance the measurement and diagnosis capabilities of NP |
| 475 | card. Support external research on PEN. |
| 476 | |
| 477 | '''Heidi Picher Dempsey''' Does it have to be in the Internet2 PoPs? It's expensive. |
| 478 | |
| 479 | '''Chip Elliot''' This seems like an interesting match for BGPMux; you |
| 480 | should talk to Nick Feamster. |
| 481 | |
| 482 | '''Chip Elliot''' Do you expect researchers to program the network cards? |
| 483 | |
| 484 | '''Yao Luo''' We'll program the network cards, not researchers. |
| 485 | |
| 486 | === Milestones === |
| 487 | |
| 488 | All of the milestones were done on time, one was done far in advance. |
| 489 | |
| 490 | == Million Node GENI ''(Justin Cappos)'' == |
| 491 | |
| 492 | 2 programmers and 15 undergrads working on this project over the summer |
| 493 | |
| 494 | Built Repy (VM), Node Manager, Seash (shell) experiment manger |
| 495 | (similar to PLUSH) |
| 496 | |
| 497 | Prototyped SeattleGENI (component manager/clearinghouse) This tracks |
| 498 | your "donations" of systems and gives you credit for them. When you |
| 499 | donate more, you get more allocated to you. (Basically VMs). |
| 500 | |
| 501 | Collaborated: PLab, PgENI, ProtoGENI, DOR |
| 502 | |
| 503 | Publications at SIGCSE, PN-ASEE/WCERT, NW-DCSD workshops, 12 talks |
| 504 | |
| 505 | This project is heavily geared towards education. What teachers want |
| 506 | to do is either run on a WAN or on a LAN. LAN is easier to debug. |
| 507 | Obviously researchers will have different requirements that we might |
| 508 | be able to add. |
| 509 | |
| 510 | Paul B. mentioned CONDOR GRID computing environment. Wondered whether |
| 511 | million node GENI could get access to Condor nodes (there are over |
| 512 | 100,000). Justin says yes, they've thought about it, and also |
| 513 | similarly with the @Home project. The issue is the trust model |
| 514 | because Million Node GENI will allow ANYONE to have an account, but GRID and other |
| 515 | projects have more restrictions. |
| 516 | |
| 517 | VM is based on Python, and is custom to Million Node GENI. |
| 518 | |
| 519 | Chip asks what kind of resources: Disk, processing, memory, network |
| 520 | bandwidth, hardware random number generator...15-20 resources are |
| 521 | restricted by Million Node GENI VM now. Million Node GENI measures based on Justin's Mac laptop |
| 522 | to decide how much to limit resources. |
| 523 | |
| 524 | Expected security model for Million Node GENI is to restrict traffic to within the |
| 525 | "testbed." VMs in a tesbed talk to each other, but nothing is an exit |
| 526 | node elsewhere. There may be other models too, but this seems to be |
| 527 | the most likely. |
| 528 | |
| 529 | Testbed this year went up over 1000 nodes. This counts anything that's |
| 530 | ever been up. |
| 531 | |
| 532 | Had to adapt to firewalls. Not looking for more end users. |
| 533 | |
| 534 | Time goes backwards on modern machines frequently |
| 535 | |
| 536 | ProtoGENI clearinghouse integration. Demo version will run at GEC5. |
| 537 | Seattle runnning on ProtoGENI. Includes AutoGrader, gets Emulab |
| 538 | resources, deploys Seattle, deploys student code and then makes grade. |
| 539 | This is supposed to be repeatable. |
| 540 | |
| 541 | Have been discussing whether it makes sense for Million Node GENI to |
| 542 | use ProtoGENI control framework. Million Node GENI belives there is a |
| 543 | problem with that. Million Node GENI doesn't have a notion of a lot |
| 544 | of things that are calls in ProtoGENI (for example booting a node). |
| 545 | |
| 546 | Might want to implement the interface but not the semantics. Might |
| 547 | want to build absolute bare minimum interface to ProtoGENI for what |
| 548 | Million Node GENI wants to do. |
| 549 | |
| 550 | Chip asked about where integration is going with Million Node GENI. |
| 551 | Million Node GENI says they want people to be able to use Million Node |
| 552 | GENI through a GENI control framework, but not the other way around. |
| 553 | To combine Million Node GENI with something else, you would use a tap |
| 554 | on Million Node GENI that lets a random internet node to act like a |
| 555 | Million Node GENI node even thought it doesn't show up as a general |
| 556 | purpose Million Node GENI node. |
| 557 | |
| 558 | Rob said he'd like to be able to have the notiion of a slice in |
| 559 | Million Node GENI. Justin said you can (but it souds like there are |
| 560 | complications with this). |
| 561 | |
| 562 | Chip asked about whether it runs on handsets. Justin said they've |
| 563 | done proof of concept on Nokia 0800, jailbroken Iphones, One laptop |
| 564 | per child notebook. Couldn't do Windows mobile despite significant |
| 565 | effort. |
| 566 | |
| 567 | Chip asked about privacy concerns, especially with handsets and |
| 568 | location information. Can route traffic through TOR and have this |
| 569 | information be hidden. If the information isn't hidden, experiments |
| 570 | can take advantage of the location information. This is a different |
| 571 | model than the cloud projects because the node will be ON your own |
| 572 | network. |
| 573 | |
| 574 | Million Node GENI will work on adding some UW nodes to GENI via I2. |
| 575 | |
| 576 | Aaron said it would be good to know whether nodes were on GENI-enabled |
| 577 | infrastructure. |
| 578 | |
| 579 | Currently used in 6 classes. Suport from NW-DCSD. Want more adoption |
| 580 | this fall. Easier for students to do experiments in this environment |
| 581 | than in others. |
| 582 | |
| 583 | * Spiral 2 plans |
| 584 | |
| 585 | * Increasing support for researchers/developers |
| 586 | * Installer/end user interface |
| 587 | * Repy v0.2 node manager (performance improvements, resource reassignment, measurement) |
| 588 | * Services |
| 589 | * Spec for end-host Clearinghouse API |
| 590 | * Prototype end-host Clearinghouse |
| 591 | * mash ups |
| 592 | * identiy management |
| 593 | * Collaborate with O&M and security |
| 594 | * GENI outreach |
| 595 | |
| 596 | Vic asked about how ProtoGENI interface works. There's a slice |
| 597 | defined for Seattle. Emulab resources can be "owned" by Seattle on |
| 598 | behalf of Million Node GENI users. |
| 599 | |
| 600 | On status Vic asked about status for tutorial (could be called done in |
| 601 | Justin's opinion). Also discussed RSPECs--still working on |
| 602 | definitions of that. |
| 603 | |
| 604 | Aaron pointed out that GENI reallly needs to be able to glue slices |
| 605 | into parts of Million Node GENI to make it useful---can't just |
| 606 | |
| 607 | Million Node GENI would like to have people deploy on Windows machines |
| 608 | to help them with debugging. HPD to follow up on the possibility of |
| 609 | deploying some Windows boxes on our BBN Emulab site. |
| 610 | |
| 611 | HPD asked if GpENI install of Million Node GENI used anything but just |
| 612 | access the GpENI Plab node. No. But it was a good node to add |
| 613 | because it has ssh access, which most nodes don't. |
| 614 | |
| 615 | === Milestones === |
| 616 | |
| 617 | == Digital Object Registry ''(Larry Lannom)'' == |
| 618 | |
| 619 | Explained the background of the Handle System and the Digital Object |
| 620 | Registry. |
| 621 | |
| 622 | Has grown up in a DoD environment in the past 3 years or so. |
| 623 | |
| 624 | Larry put up a chart of terms for GFC (GENI Federated Clearinghouse of |
| 625 | CNRI), ProtoGENI, and Million Node GENI |
| 626 | |
| 627 | Rob commented that all CMs in ProtoGENi are also AMs. There are are |
| 628 | no CMs that only do single components. |
| 629 | |
| 630 | * Spiral 2 goals |
| 631 | * Clearinghouse: |
| 632 | * defined a normalized and interoperable GENI clearinghouse specificatoin |
| 633 | * provide our sw to new and existing projects |
| 634 | * federate individual clearinghouse s into the GFC allowing researchers to discover resources across GENI. |
| 635 | * Security: |
| 636 | * Integrate and make available the propsoed PKI solution (from CNRI). Has been running for at least a decade. |
| 637 | * Identifiers |
| 638 | |
| 639 | Scholarly publishers are current most frequent users of the Handle system. |
| 640 | |
| 641 | Can do interfaces into repositories. |
| 642 | |
| 643 | == GPO Spiral 2 Overview and Discussion ''(Chip Elliot)'' == |
| 644 | |
| 645 | '''Chip Elliot''' We should have a workshop for people involved in building and running |
| 646 | testbeds and infrastructure and send some recommendations to NSF explaining |
| 647 | what is needed for research infrastructure. |
| 648 | |
| 649 | '''Chip Elliot''' There is a clear disconnect. They like the idea of infrastructure, but |
| 650 | don't like the cost. |
| 651 | |
| 652 | GPO to put Shib/InCommon info from Docushare onto wiki for convenience of GENI projects investigating it. Harry could do this (also overlaps with Ketly). |
| 653 | |
| 654 | Paul said the last NSF nets call included a one-page supplemental about how the research would be done in GENI. |