IARC (International Arctic Research Center Data Archive) and DataONE - conversations

10 April 2014 (1p Eastern/9a Alaska)

Attending:  Jim Long (IARC), Bruce Wilson (UT/ORNL/DataONE), Laura Moyers (UT/DataONE)

IARC's website:  http://climate.iarc.uaf.edu/geonetwork/srv/en/main.home

Helpful information related to joining DataONE as a Member Node:
 * http://www.dataone.org/member-node-deployment-process
 * http://mule1.dataone.org/OperationDocs/member_node_deployment/mn_checklist.html (a little more detailed version of the above)
 * http://www.dataone.org/member_node-deployment-routes


IARC would like to become part of a larger community and would like to participate in some reciprocal replication.  

data volume: <5TB
data qualities: ecological, native peoples data, etc. (goal to collect data that isn't being picked up by other archives)
identifiers:  MD5sum; looking at obtaining DOIs (we can provide guidance there)

An example of how we may look at data objects:

An envelope (OAI-ORE, resource map) contains one or more metadata file(s) and data file(s)
 * A science metadata file in ISO19115 format describing the data
 * A science metadata file in FGDC format describing the data
 * ... maybe other metadata files
 * A science data file (the meat) in CSV format
 * A science data file in XLS format
 * ... maybe other data files in different formats

IARC currently operates a GeoNetwork system
also has an OAI-PMH point, ISO 19115 metadata 

DataONE can harvest metadata records from IARC with the OAI-PMH info and make IARC's data discoverable via DataONE.  However, we'd lose the enhanced capabilities provided by the Investigator Tool Kit (things like R) which allow expanded analysis of data.

Tiers
 * tier 1: unauthenticated reads (anyone can read)
 * tier 2: authenticated reads (requires a login)
 * tier 3: allows writing data to the MN
 * tier 4: allows replication of data from other MNs to the MN

IARC has login capability, managing own accounts.  Not doing LDAP, using the account manager that comes with GeoNetworks.  

CILogon and InCommon - InCommon (US-based identifier, based on SAML 2.0)
CILogon (built on top of InCommon, extends it a bit (allowing Google logons))

DataONE's Python stack is a full tier 4 implementation.

Certificate authentication can be a little squirrely to get hold of
    client certificates
    server certificates
    Coordinating Nodes (CNs) have certificates, too
    in replication, will communicate with other MNs with their certificates
Once you "get" it, certificates are manageable.

Environments in DataONE:
 * sandbox - play area, get things up and running, can wipe out if needed
 * testing (staging) - environment looks more like the production environment, test the MN's operations, communication with CNs, etc.; when it seems to be working well, can move to...
 * production

Bruce to send MN slide deck to Jim, has links to lots of good info

How does IARC handle data set updates?  If you delete it in GeoNetwork, it will save the delete.  User can delete a dataset and upload a new one.  

Desired behavior is that if a data object is changed, it gets a new identifier.  

There is a story in redmine to accomodate GeoNetwork; several entities in Europe are using GeoNetwork, as well as IARC.  

??  Could we expose IARC's metadata via DataONE and later work on a more robust MN (maybe GMN) ??  
Ultimate goal is to be a fully-functional MN; we can proceed in phases (expose metadata first, then work on how to expose data for direct access for full functionality) 

DataONE Users Group, meeting info: http://www.dataone.org/dataone-users-group in conjunction with the Federation of Earth Science Information Partners (ESIP)

ESIP started with funding by NASA, now NOAA, EPA, USGS, some NSF

Bruce's actions:  MN slide deck to Jim, contacts, DUG info, same stuff as sent to TERN