Changes between Version 9 and Version 10 of Gec5ClusterBAgenda


Ignore:
Timestamp:
07/22/09 15:40:19 (10 years ago)
Author:
Christopher Small
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Gec5ClusterBAgenda

    v9 v10  
    4040
    4141----
     42
    4243''The following notes were taken by Christopher Small, GPO System Engineer''
    4344
     
    4546
    4647== To Do List ==
    47 
    48 * Larry Peterson
     48 
     49=== Larry Peterson ===
    4950
    5051Plan for this meeting is to bring each other up to speed on what we're doing.
     
    5657Three ways to federate.
    5758
    58 SFA only -- WSDL interface specification, you implement it. This is
     59 * SFA only -- WSDL interface specification, you implement it. This is
    5960what Enterprise GENI has done. Generated this based on the running
    6061code, but plan to go forward using the WSDL spec as the base and we'll
     
    6465Enterprise GENI is using this.
    6566
    66 SFA + geniwrapper -- includes security. Write your own back end to
     67 * SFA + geniwrapper -- includes security. Write your own back end to
    6768geniwrapper. If you've got a preexisting testbed, this is probably the
    6869fastest way to go; use the security machinery and interface spec, and
    6970have it call into your system.
    7071
    71 SFA + geniwrapper + MyPLC -- includes PlanetLab O&M machinery
     72 * SFA + geniwrapper + MyPLC -- includes PlanetLab O&M machinery
    7273
    7374RSpecs are the place where we've always put things that we don't know
     
    9596Did demo of PlanetLab with MAX nodes.
    9697
    97 * James Sterbenz
     98=== James Sterbenz ===
    9899
    99100GpENI demo, showed that Lancaster University UK, ETH-Zurich are
     
    101102their own PLC at KSU. Separare PLC, separate VINI, not federated.
    102103
    103 * Jeannie Albrecht
     104=== Jeannie Albrecht ===
    104105
    105106XML description of the experiment. Command line program reads the xml,
    106107sets up and runs the experiment.
    107108
    108 * Chris Tracy
     109=== Chris Tracy ===
    109110
    110111Describes the recursive aggregate manager model they have running,
     
    115116-----
    116117
    117 = Tuesday 1:00-3:00
    118 
    119 * Larry Peterson
     118== Tuesday 1:00-3:00 ==
     119
     120=== Larry Peterson ===
    120121
    121122We need to work on attracting real users
     
    125126Let's advertise our new cluster B aggregates to those users
    126127
    127 * James Sterbenz
     128=== James Sterbenz ===
    128129
    129130Are we just adding users to have users? Are you suggesting we
    130131advertise our resources as standard PlanetLab nodes?
    131132
    132 * Larry Peterson
     133=== Larry Peterson ===
    133134
    134135Control Framework todo
     
    148149manager.
    149150
    150 * Guido Appenzeller
     151=== Guido Appenzeller ===
    151152
    152153Eventually we'd like to have something that abstracts some of the
    153154details away, a web interface, to make this all easier to deal with.
    154155
    155 * Larry Peterson
     156=== Larry Peterson ===
    156157
    157158Chris Tracy put up a recursive aggregate manager, this may be what you
     
    170171(or set of flat)
    171172
    172 * Guido Appenzeller
     173=== Guido Appenzeller ===
    173174
    174175Don't we need some kind of pluggable user interface to manage
    175176heterogeneous aggregates?
    176177
    177 * Larry Peterson
     178=== Larry Peterson ===
    178179
    179180That's what Chris Tracy demonstrated, this is what we need. The SFI
     
    185186We've just broken the clearinghouse model.
    186187
    187 * Peter O'Neil
     188=== Peter O'Neil ===
    188189
    189190From a usability standpoint this will be chaos. Need a single portal.
    190191
    191 * Larry Peterson
     192=== Larry Peterson ===
    192193
    193194There's not just one portal to the internet.
     
    196197thing, or is it a loose federation of entities?
    197198
    198 * Guido Appenzeller
     199=== Guido Appenzeller ===
    199200
    200201Seems to fit into GENI goals to have one centralized clearinghouse.
    201202
    202 * Larry Peterson
     203=== Larry Peterson ===
    203204
    204205What do you mean by GENI?
    205206
    206 * Guido Appenzeller
     207=== Guido Appenzeller  ===
    207208
    208209The internet has DNS, it tells me where I can look for things
    209210
    210 * Larry Peterson
     211=== Larry Peterson ===
    211212
    212213we can have a root registry and if we want to kill a slice we can
     
    227228Policies will only be as sophisticated as the mechanisms.
    228229
    229 * Guido Appenzeller
     230=== Guido Appenzeller  ===
    230231
    231232We only talk with known entities. We talk to an SM / clearinghouse,
    232233not a user. A researcher hand-editing XML sounds ambitious.
    233234
    234 * Larry Peterson
     235=== Larry Peterson ===
    235236
    236237Writing an SM that spoke to a hererogeneous collection of aggregates
     
    250251Can somebody explain this to me?
    251252
    252 * James Sterbenz
     253=== James Sterbenz ===
    253254
    254255Because the GPO told us to.
    255256
    256 * James Kempf (?)
     257=== James Kempf (?) ===
    257258
    258259You might want dedicated bandwidth
    259260
    260 * Larry Peterson
     261=== Larry Peterson ===
    261262
    262263IP encapsulation is only 28 bytes, it's running over layer 2.
    263264
    264 * Guido Appenzeller
     265=== Guido Appenzeller ===
    265266
    266267We want to give people hop-by-hop layer 2 so they can put a router in
     
    268269them, not have an IP header.
    269270
    270 * Larry Peterson
     271=== Larry Peterson ===
    271272
    272273My argument is cost effectiveness. This will be a large investment of
    273274time and effort with minimal value.
    274275
    275 * Jon Turner
     276=== Jon Turner ===
    276277
    277278Both are legitimate; you can get to more places with IP, layer 2 VLANs
    278279will let you get closer to the bare network.
    279280
    280 * Guido Appenzeller
     281=== Guido Appenzeller ===
    281282
    282283Every slice must support IP fallback
    283284
    284 * Larry Peterson
     285=== Larry Peterson ===
    285286
    286287We haven't solved the problem because we still need to demux between
     
    294295fairly low on user-level tools.
    295296
    296 * Jeannie Albrecht
     297=== Jeannie Albrecht ===
    297298
    298299Once you have a stable version of SFI, I can try to make things
    299300prettier. But it doesn't make sense until SFI stabilizes.
    300301
    301 * Larry Peterson
     302=== Larry Peterson ===
    302303
    303304Look at the VINI rspec.
     
    308309Goal for SFI: readable/editable RSpecs e.g. using XSGR
    309310
    310 * Chris Tracy
     311=== Chris Tracy ===
    311312
    312313We have some tools that might be usable
    313314
    314 * Larry Peterson
     315=== Larry Peterson ===
    315316
    316317Show of hands -- who has time for a GUI?
    317318
    318 * Guido Appenzeller
     319=== Guido Appenzeller ===
    319320
    320321What we see as the hard part of the GUI is how to help the user if the
     
    322323accepted.
    323324
    324 * Larry Peterson
     325=== Larry Peterson ===
    325326
    326327Will try for pretty text plus policy feedback.
    327