Changes between Version 13 and Version 14 of 2ndInstMeasWork


Ignore:
Timestamp:
07/01/10 11:52:05 (9 years ago)
Author:
jtaylor@bbn.com
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • 2ndInstMeasWork

    v13 v14  
    1818==  Attendees at workshop  ==
    1919
    20 (attended workshop:  yes or no)
    21         Paul Barford - University of Wisconsin – Madison  (no)
    22         Bruce Maggs – Duke University and Akamai  (yes)
    23         Harry Mussman – BBN/GPO  (yes)
    24         Vic Thomas - BBN/GPO  (yes)
    25         Evan Zhang – BBN/GPO  (yes)
     20(Attended workshop:  yes or no)
     21  Paul Barford - University of Wisconsin – Madison  (no)[[BR]]
     22  Bruce Maggs – Duke University and Akamai  (yes)[[BR]]
     23  Harry Mussman – BBN/GPO  (yes)[[BR]]
     24  Vic Thomas - BBN/GPO  (yes)[[BR]]
     25  Evan Zhang – BBN/GPO  (yes)[[BR]]
    2626OML (ORBIT Measurement Library) OMF (ORBIT Management Framework)
    27         Max Ott – NICTA  (yes, by phone)
    28         Ivan Seskar – Rutgers WINLAB  (yes)
     27  Max Ott – NICTA  (yes, by phone)[[BR]]
     28  Ivan Seskar – Rutgers WINLAB  (yes)[[BR]]
    2929Instrumentation Tools 
    30         Jim Griffioen - Univ Kentucky  (yes)
     30  Jim Griffioen - Univ Kentucky  (yes)
    3131perfSONAR 
    32         Matt Zekauskas - Internet2 (no)
    33         Jason Zurawski – Internet2  (yes)
    34         Martin Swany - Univ Delaware (yes)
    35         Guilherme Fernandes – Univ Delaware  (yes)
    36         Ezra Kissel – Univ Delaware  (yes)
    37 Scalable Sensing Service (S3) 
    38         Sonia Fahmy – Purdue  (yes)
    39         Puneet Sharma - HP Labs  (yes)
    40 OnTimeMeasure for network measurements
    41         Prasad Calyam - Ohio Supercomputing Ctr  (yes)
    42 GENI Meta-Operations Center and NetKArma
    43         Jon-Paul Herron - Indiana Univ
    44         Camilo Viecco - Indiana Univ   (yes)   
    45         Chris Small - Indiana Univ  (yes)
    46         Beth Plale - Indiana Univ  (yes)
    47 Virtual Machine Introspection (VMI) 
    48         Brian Hay – Univ Alaska  (yes)
     32  Matt Zekauskas - Internet2 (no)[[BR]]
     33  Jason Zurawski – Internet2  (yes)[[BR]]
     34  Martin Swany - Univ Delaware (yes)[[BR]]
     35  Guilherme Fernandes – Univ Delaware  (yes)[[BR]]
     36  Ezra Kissel – Univ Delaware  (yes)[[BR]]
     37Scalable Sensing Service (S3)
     38  Sonia Fahmy – Purdue  (yes)[[BR]]
     39  Puneet Sharma - HP Labs  (yes)[[BR]]
     40!OnTimeMeasure for network measurements
     41  Prasad Calyam - Ohio Supercomputing Ctr  (yes)
     42GENI Meta-Operations Center and !NetKarma
     43  Jon-Paul Herron - Indiana Univ[[BR]]
     44  Camilo Viecco - Indiana Univ   (yes)[[BR]]   
     45  Chris Small - Indiana Univ  (yes)[[BR]]
     46  Beth Plale - Indiana Univ  (no)[[BR]]
     47Virtual Machine Introspection (VMI)
     48  Brian Hay – Univ Alaska  (yes)[[BR]]
    4949Data-Intensive Cloud Control for GENI
    50         Michael Zink  - UMass Amherst  (yes)
     50  Michael Zink  - UMass Amherst  (yes)[[BR]]
    5151Experiment Management Service – Digital Object Registry
    52         Jim French - CNRI    (yes)
    53         Giridhar Manepalli - CNRI   (yes)
    54         Larry Lannom – CNRI  (no)
     52  Jim French - CNRI (yes)[[BR]]
     53  Giridhar Manepalli - CNRI (yes)[[BR]]
     54  Larry Lannom – CNRI (no)[[BR]]
    5555
    5656
     
    6060
    6161[http://groups.geni.net/geni/attachment/wiki/2ndInstMeasWork/062510%20InstMeasWorkshopAgenda.pdf  Announcement, v1.2 (062510)] [http://groups.geni.net/geni/attachment/wiki/2ndInstMeasWork/062510%20InstMeasWorkshopAgenda.doc  word file][[BR]]
    62 [  Figures, v1.2 (062510)] [[BR]]
     62[Figures, v1.2 (062510)] [[BR]]
    6363
    6464
     
    7171
    7272Team members:
    73         Paul Barford - University of Wisconsin – Madison  (no)
    74         Jim Griffioen - Univ Kentucky  (yes)
    75         *  Prasad Calyam - Ohio Supercomputing Ctr  (yes)
    76         Camilo Viecco - Indiana Univ   (yes)
    77         Brian Hay – Univ Alaska  (yes)
    78         * agreed to organize first discussion and writing
     73  Paul Barford - University of Wisconsin – Madison  (no)[[BR]]
     74  Jim Griffioen - Univ Kentucky  (yes)[[BR]]
     75  Prasad Calyam* - Ohio Supercomputing Ctr  (yes)[[BR]]
     76  Camilo Viecco - Indiana Univ   (yes)[[BR]]
     77  Brian Hay – Univ Alaska  (yes)[[BR]]
     78 
     79  *agreed to organize first writing and discussion
    7980
    8081Identify all user groups, and provide basic use cases:
    81         1) GMOC operations group (NOC)
    82         2) Cluster and aggregate operations groups (NOCs)
    83         3) Archive service providers
    84         4) Experiment researchers
    85         5) Experiment (opt-in) users  (see http://groups.geni.net/geni/attachment/wiki/041409NYCOptInWGAgenda/071509%20%20GENI-SE-OI-Overview-01.4.pdf for  listing of opt-in issues, such as privacy)
    86         6)  Researchers that use measurement data archived by other researchers (DatCat model)
     82  1. GMOC operations group (NOC)
     83  1. Cluster and aggregate operations groups (NOCs)
     84  1. Archive service providers
     85  1. Experiment researchers
     86  1. Experiment (opt-in) users  (see http://groups.geni.net/geni/attachment/wiki/041409NYCOptInWGAgenda/071509%20%20GENI-SE-OI-Overview-01.4.pdf for  listing of opt-in issues, such as privacy)
     87  1. Researchers that use measurement data archived by other researchers (!DatCat model)
    8788
    8889=== Topic 2  GENI I&M Services ===
    8990
     91Team members:[[BR]]
     92  Harry Mussman* – BBN/GPO  (yes)[[BR]]
     93  Evan Zhang – BBN/GPO[[BR]]
     94  Giridhar Manepalli - CNRI   (yes)[[BR]]
     95  Chris Small - Indiana Univ  (yes)[[BR]]
     96  Beth Plale - Indiana Univ  (yes)[[BR]]
     97
     98  *agreed to organize first writing and discussion
     99
     100Summarize current view
     101
     102Identify different types of services:[[BR]]
     103  Completely dedicated to an experiment[[BR]]
     104  Common portion, plus parts associated with different experiments[[BR]]
     105  Common service, with data provided to multiple experiments[[BR]]
     106
     107Need:  Basic definition of an archive service
     108
     109
     110=== Topic 3  GENI I&M Resources ===
     111
    90112Team members:
    91         *  Harry Mussman – BBN/GPO  (yes)
    92         Evan Zhang – BBN/GPO
    93         Giridhar Manepalli - CNRI   (yes)
    94         Chris Small - Indiana Univ  (yes)       
    95         Beth Plale - Indiana Univ  (yes)
    96         * agreed to organize first discussion and writing
    97 
    98 Summarize current view
    99 
    100 Identify different types of services:
    101         Completely dedicated to an experiment
    102         Common portion, plus parts associated with different experiments
    103         Common service, with data provided to multiple experiments
    104 
    105 Need:  Basic definition of an archive service
    106 
    107 
    108 === Topic 3  GENI I&M Resources ===
     113  Vic Thomas - BBN/GPO  (yes)[[BR]]
     114  Jim Griffioen* - Univ Kentucky (yes)[[BR]]
     115  Martin Swany - Univ Delaware (yes)[[BR]]
     116  Camilo Viecco - Indiana Univ (yes)[[BR]]     
     117  Brian Hay – Univ Alaska  (yes)[[BR]]
     118  Giridhar Manepalli - CNRI (yes)[[BR]]
     119 
     120  *agreed to organize first writing and discussion
     121
     122Significant question uncovered at workshop:[[BR]]
     123Jim on 6/25 via email:  We should involve Rob Ricci in the discussion.
     124
     125What are resources:
     126  1. Hosts, VMs, etc.
     127  1. Network connectivity
     128  1. Software, e.g., I&M software that can be included in an experiment
     129  1. I&M services
     130  1. I&M data flows and file transfers
     131  1. I&M data files stored in archives
     132
     133How are each of these discovered, specified, authorized and assigned:
     134  a. Always by mechanisms provided by the CF?
     135  b. With CF plus additional mechanisms? 
     136  Consider example of LS in perfSONAR  [[BR]]
     137  Consider example of data file stored in archive, owned by an experimenter
     138
     139Goals:[[BR]]
     140  Need to define and then compare these options[[BR]]
     141  Need to understand interop with CF for each option[[BR]]
     142  Does CF setup secondary authorization mechanisms in some cases?  If so, how?[[BR]]
     143
     144Does each item have:[[BR]]
     145  Unique and persistent name?[[BR]]
     146  Unique and persistent identifier?[[BR]]
     147  Need to carefully consider this for all of GENI[[BR]]
     148
     149For each item, consider how to:[[BR]]
     150  Create[[BR]]
     151  Name[[BR]]
     152  Register and discover[[BR]]
     153  Authorize and assign[[BR]]
     154
     155For each item, consider:[[BR]]
     156  Ownership[[BR]]
     157  What sort of policies the owner may want to apply[[BR]]
     158
     159
     160
     161=== Topic 4  GENI I&M Measurement Plane and Interfaces ===
    109162
    110163Team members:
    111         Vic Thomas - BBN/GPO  (yes)
    112         *  Jim Griffioen - Univ Kentucky  (yes)
    113         Martin Swany - Univ Delaware (yes)
    114         Camilo Viecco - Indiana Univ   (yes)   
    115         Brian Hay – Univ Alaska  (yes)
    116         Giridhar Manepalli - CNRI   (yes)
    117         * agreed to organize first discussion and writing
    118 
    119 Significant question uncovered at workshop!
    120 Jim on 6/25 via email:  We should involve Rob Ricci in the discussion.
    121 
    122 What are resources:
    123         1)  Hosts, VMs, etc.
    124         2)  Network connectivity
    125         3)  Software, e.g., I&M software that can be included in an experiment
    126         3)  I&M services
    127         4)  I&M data flows and file transfers
    128         5)  I&M data files stored in archives
    129 
    130 How are each of these discovered, specified, authorized and assigned:
    131         a)  Always by mechanisms provided by the CF?
    132         b)  With CF plus additional mechanisms? 
    133         Consider example of LS in perfSONAR 
    134         Consider example of data file stored in archive, owned by an experimenter
    135 
    136 Goals:
    137         Need to define and then compare these options
    138         Need to understand interop with CF for each option
    139         Does CF setup secondary authorization mechanisms in some cases?  If so, how?
    140 
    141 Does each item have:
    142         Unique and persistent name?
    143         Unique and persistent identifier?
    144         Need to carefully consider this for all of GENI
    145 
    146 For each item, consider how to:
    147         Create
    148         Name
    149         Register and discover
    150         Authorize and assign
    151 
    152 For each item, consider:
    153         Ownership
    154         What sort of policies the owner may want to apply
    155 
    156 
    157 
    158 
    159 
    160 
     164  Harry Mussman* – BBN/GPO  (yes)[[BR]]
     165  Ezra Kissel – Univ Delaware  (yes)[[BR]]
     166  Chris Small - Indiana Univ  (yes)[[BR]]
     167
     168  *agreed to organize first writing and discussion
     169
     170Consider:
     171  IP network[[BR]]
     172  Layer 2 (VLAN) connections[[BR]]
     173
     174Discuss:[[BR]]
     175  Which protocols are active[[BR]]
     176  Access to resources in aggregates, even when resources are in private address space, via GWs or proxies[[BR]]
     177  How to provide authentication and authorization[[BR]]
     178  How to provide QoS to protect measurement traffic[[BR]]
     179  How to provide QoS to protect other traffic when measurement traffic is large.[[BR]]
     180  Reserve bandwidth?[[BR]]
     181
     182Martin on 6/28:  Consider XSP (extensible session protocol) to provide transport layer GW functions.
     183
     184
     185=== Topic 5  GENI I&M  Interfaces and Protocols (APIs):  Manage Services  ===
     186
     187Team members:
     188  Vic Thomas - BBN/GPO  (yes)[[BR]]
     189  Ivan Seskar – Rutgers WINLAB  (yes)[[BR]]
     190  Max Ott – NICTA  (yes, by phone)[[BR]]
     191  Sonia Fahmy* – Purdue  (yes)[[BR]]
     192  Giridhar Manepalli - CNRI   (yes)[[BR]]
     193 
     194  *agreed to organize first discussion and writing
     195
     196Define an approach based on OMF/OML and S3:
     197  HTTP(S)[[BR]]
     198  REST vs SOAP[[BR]]
     199  Authorization by credentials or ?  If credentials, how to revoke?[[BR]]
     200  Pass XML fragments[[BR]]
     201  Define basic API[[BR]]
     202
     203===  Topic 6  GENI I&M  Interfaces and Protocols (APIs):  Data Flows and Data File Transfers ===
     204
     205Team members:[[BR]]
     206  Harry Mussman* – BBN/GPO  (yes)[[BR]]
     207  Ivan Seskar – Rutgers WINLAB  (yes)[[BR]]
     208  Max Ott – NICTA  (yes, by phone)[[BR]]
     209  Ezra Kissel – Univ Delaware  (yes)[[BR]]
     210  Prasad Calyam - Ohio Supercomputing Ctr  (yes)[[BR]]
     211  Michael Zink  - UMass Amherst  (yes)[[BR]]
     212
     213  *agreed to organize first writing and discussion
     214
     215Consider data flows and data file transfers between all services
     216
     217Define range of options:
     218
     219What:[[BR]]
     220  Data flows[[BR]]
     221  Data files transfers[[BR]]
    161222       
    162 === Topic 4  GENI I&M Measurement Plane and Interfaces ===
    163 
    164 Team members:
    165         *  Harry Mussman – BBN/GPO  (yes)
    166         Ezra Kissel – Univ Delaware  (yes)
    167         Chris Small - Indiana Univ  (yes)
    168         * agreed to organize first discussion and writing
    169 
    170 Consider:
    171         IP network
    172         Layer 2 (VLAN) connections
    173 
    174 Discuss
    175         Which protocols are active
    176         Access to resources in aggregates, even when resources are in private address space, via GWs or proxies
    177         How to provide authentication and authorization
    178         How to provide QoS to protect measurement traffic
    179         How to provide QoS to protect other traffic when measurement traffic is large. 
    180         Reserve bandwidth?
    181 
    182 Martin on 6/28:  Consider XSP (extensible session protocol) to provide transport layer GW functions.
    183 
    184 
    185 === Topic 5  GENI I&M  Interfaces and Protocols (APIs):  Manage Services  ===
    186 
    187         Vic Thomas - BBN/GPO  (yes)
    188         Ivan Seskar – Rutgers WINLAB  (yes)
    189         Max Ott – NICTA  (yes, by phone)
    190         *  Sonia Fahmy – Purdue  (yes)
    191         Giridhar Manepalli - CNRI   (yes)
    192         * agreed to organize first discussion and writing
    193 
    194 Define an approach based on OMF/OML and S3:
    195         HTTP(S)
    196         REST vs SOAP
    197         Authorization by credentials or ?  If credentials, how to revoke?
    198         Pass XML fragments
    199         Define basic API
    200 
    201 ===  Topic 6  GENI I&M  Interfaces and Protocols (APIs):  Data Flows and Data File Transfers ===
    202 
    203         *  Harry Mussman – BBN/GPO  (yes)
    204         Ivan Seskar – Rutgers WINLAB  (yes)
    205         Max Ott – NICTA  (yes, by phone)
    206         Ezra Kissel – Univ Delaware  (yes)
    207         Prasad Calyam - Ohio Supercomputing Ctr  (yes)
    208         Michael Zink  - UMass Amherst  (yes)
    209         * agreed to organize first discussion and writing
    210 
    211 Consider data flows and data file transfers between all services
    212 
    213 Define range of options:
    214 
    215 What:
    216         Data flows
    217         Data files transfers
    218        
    219 Type:
    220         Pull
    221         Push
    222         Pub/Sub
    223 
    224 Protocol:
    225         SNMP
    226         SCP
    227         FTP and gridFTP
    228         HTTP
    229         XMPP
    230         TCP
    231         SCTP
    232 
    233 Consider:
    234         Naming
    235         Discovery
    236         Connectivity
    237         Authentication and authorization mechanisms
    238 
    239 Map to current projects, giving examples:
     223Type:[[BR]]
     224  Pull[[BR]]
     225  Push[[BR]]
     226  Pub/Sub[[BR]]
     227
     228Protocol:[[BR]]
     229  SNMP[[BR]]
     230  SCP[[BR]]
     231  FTP and gridFTP[[BR]]
     232  HTTP[[BR]]
     233  XMPP[[BR]]
     234  TCP[[BR]]
     235  SCTP[[BR]]
     236
     237Consider:[[BR]]
     238  Naming[[BR]]
     239  Discovery[[BR]]
     240  Connectivity [[BR]]
     241  Authentication and authorization mechanisms[[BR]]
     242
     243Map to current projects, giving examples:[[BR]]
     244Consider:  Minimum set required for GENI[[BR]]
     245
     246
     247=== Topic 7  GENI I&M  Interfaces and Protocols (APIs):  Service Registration and Discovery ===
     248
     249Team members:[[BR]]
     250  Jason Zurawski* – Internet2  (yes)[[BR]]
     251  Prasad Calyam - Ohio Supercomputing Ctr  (yes)[[BR]]
     252  *agreed to organize first writing and discussion[[BR]]
     253
     254Consider approach used in perfSONAR
     255
     256Summarize for:[[BR]]
     257  Services with data flows[[BR]]
     258  Also sources of file transfers?[[BR]]
     259  Also GUIs?[[BR]]
     260
     261
     262=== Topic 8  GENI I&M  Interfaces and Protocols (APIs):  GUIs ===
     263
     264Team members:[[BR]]
     265  Jeremy Reed - Univ Kentucky  (yes)[[BR]]
     266  Guilherme Fernandes* – Univ Delaware  (yes)[[BR]]
     267  Puneet Sharma - HP Labs  (yes)[[BR]]
     268  *Agreed to organize team[[BR]]
     269
     270Define overall goals for GENI GUIs
     271
     272Types of GUIs:[[BR]]
     273  Control experiments[[BR]]
     274  Display I&M results[[BR]]
     275  Report status[[BR]]
     276  View archive service[[BR]]
     277
     278Consider portal, for access to multiple GUIs
     279
     280Consider need for authentication and authroization
     281
     282=== Topic 9  GENI Measurement Data Schema ===
     283
     284Team members:[[BR]]
     285  Bruce Maggs – Duke University and Akamai  (yes)[[BR]]
     286  Max Ott – NICTA  (yes, by phone)[[BR]]
     287  Ivan Seskar – Rutgers WINLAB  (yes)[[BR]]
     288  Martin Swany* - Univ Delaware (yes)[[BR]]
     289  Camilo Viecco - Indiana Univ (yes)[[BR]]
     290  Michael Zink  - UMass Amherst (yes)[[BR]]
     291  Jim French - CNRI (yes)[[BR]]
     292 
     293  *agreed to organize first discussion and writing
     294
     295
     296Consider:[[BR]]
     297  Measurement data schema[[BR]]
     298  Metadata schema[[BR]]
     299  Metadata contents[[BR]]
     300
     301
     302Consider measurement data schema and/or metadata schema from:[[BR]]
     303  perfSONAR[[BR]]
     304  GMOC-provided[[BR]]
     305  Current OML[[BR]]
     306  Proposed using IPFIX[[BR]]
     307  NetCDF (as used by DI Cloud)[[BR]]
     308
    240309Consider:  Minimum set required for GENI
    241310
    242 
    243 
    244 
    245 === Topic 7  GENI I&M  Interfaces and Protocols (APIs):  Service Registration and Discovery ===
    246 
    247 Team members:
    248         *  Jason Zurawski – Internet2  (yes)
    249         Prasad Calyam - Ohio Supercomputing Ctr  (yes)
    250         * agreed to organize first discussion and writing
    251 
    252 Consider approach used in perfSONAR
    253 
    254 Summarize for:
    255         Services with data flows
    256         Also sources of file transfers?
    257         Also GUIs?
    258 
    259 
    260 === Topic 8  GENI I&M  Interfaces and Protocols (APIs):  GUIs ===
    261 
    262 Team members:
    263         Jeremy Reed - Univ Kentucky  (yes)
    264         *  Guilherme Fernandes – Univ Delaware  (yes)
    265         Puneet Sharma - HP Labs  (yes)
    266         * Agreed to organize team
    267 Define overall goals for GENI GUIs
    268 
    269 Types of GUIs:
    270         Control experiments
    271         Display I&M results
    272         Report status
    273         View archive service
    274 
    275 Consider portal, for access to multiple GUIs
    276 
    277 Consider need for authentication and authroization
    278 
    279 === Topic 9  GENI Measurement Data Schema ===
    280 
    281 Team members:
    282         Bruce Maggs – Duke University and Akamai  (yes)
    283         Max Ott – NICTA  (yes, by phone)
    284         Ivan Seskar – Rutgers WINLAB  (yes)
    285         *  Martin Swany - Univ Delaware (yes)
    286         Camilo Viecco - Indiana Univ   (yes)   
    287         Michael Zink  - UMass Amherst  (yes)
    288         Jim French - CNRI    (yes)
    289         * agreed to organize first discussion and writing
    290 
    291 
    292 Consider:
    293         Measurement data schema
    294         Metadata schema
    295         Metadata contents
    296 
    297 
    298 Consider measurement data schema and/or metadata schema from:
    299         perfSONAR
    300         GMOC-provided
    301         Current OML
    302         Proposed using IPFIX
    303         NetCDF (as used by DI Cloud)
    304 
    305 Consider:  Minimum set required for GENI
    306 
    307 Provide overall template for GENI metadata, considering above.
     311Provide overall template for GENI metadata, considering above.[[BR]]
    308312Which items in GENI metadata template are:
    309         Required?
    310         Invariant?
     313  Required?[[BR]]
     314  Invariant?[[BR]]
    311315 
    312316