CHSS Library & Academic Computing Committee

CHSS Library & Academic Computing Committee

20 May 2014

Paper G

Disclosable

For Information

College data storage developments and plan

Introduction

The University-wide RDM project is delivering a number of research data initiatives. One of these initiatives is the provision of large scale storage facilities. The University will provide up to 0.5Tb of storage free at the point of use for research active staff and research students. This storage is available for use now.

The new storage platform also presents an opportunity for the College to move from its current platform which is reaching end of life. It presents an opportunity to increase overall storage, provisioned in one place and which offers greater numbers of access mechanisms, particularly for off-campus.

Proposal

Phase 1 – research data storage

IS have completed the provisioning of user spaces for all CHSS researchers[1] on the DataStore service, in their "personal" allocation spaces at:

\\chss.datastore.ed.ac.uk\chss\SCHOOL\users\USERNAME

This is available now and is being rolled out by IS and School-based storage managers – see appendix A.

Datastore is only available to be used by researchers at the present time. Allocations have been created for researchers for each School, and School-based storage managers will check and confirm the status of each at roll-out.

Since central systems have been used to provision these spaces, errors or omissions will need changes and amendments to upstream, golden-copy data.

Phase 2 – all College users

The College currently uses approximately 20Tb of storage on IS-managed storage infrastructure with a mixture of personal, School and College level storage for which the College pays annual maintenance charges.

College will use these funds to pay for storage on Datastore in addition for research data and IS will decommission the current platform. At current costs, this would be approximately 50Tb and so presents greater value for money than at present. This space would be in addition to the 0.5Tb per researcher that is free at the point of use.

Having 50Tb would help “smooth” the current issues with research data allocation policy – see appendix B.

Proposed structure[2]:

All users / \\chss.datastore.ed.ac.uk\chss\SCHOOL\users\USERNAME / Quotas applied for each user. 500Gb for researchers, 25Gb (tbc) for all others
Research Groups / \\chss.datastore.ed.ac.uk\chss\SCHOOL\groups\GROUPNAME / Quotas for research groups are made up from sacrifices by their constituent researcher plus funded top-up
Other groups / \\chss.datastore.ed.ac.uk\chss\SCHOOL\shared\SHARENAME / A direct replacement for existing shared directories, where they are not naturally accommodated within research groups above

Issues to note:

·  Quotas – management of separate individual quotas; 500Gb for researchers and, suggested, 25Gb for non-researchers. Management of researchers who go on to become non-research staff; e.g. PGR to admin will require careful management. Overall quota applied at College level to give maximum flexibility

·  Permissions – need to consider a more rigid application of user groups than currently required due to the different underlying storage technology. Assigning permissions to individual users for shared folders is problematic

·  Provisioning – use for home directory requires provisioning of user directory before first logon. Timing is crucial

·  Timescales – migrate and decommission Khyber and Atlas by the end of summer

·  Migration – fully-managed background migration over a set period and at a predetermined and agreed time out of normal working hours or at weekends. Old data set to read-only then deleted at a set time after migration

Recommendations

LACC is asked to note the paper, previously issued to and discussed by College CPAG, to disseminate as relevant to their Schools and to contact the College CIO for further information as appropriate.

Further communications to all those affected regarding the migrations will be undertaken locally.

Fraser Muir

CIO, College of Humanities and Social Sciences

13 May, 2014

Appendix A – RDM contacts in CHSS

College / Role
Fraser Muir / CIO and College representative on RDM steering group
Richard Lomax / Storage manager
Jacq McMahon / College Research Officer
Alvin Jackson / Dean of Research and Deputy Head of College
School / Director of Research / Administrative Contacts / Open Access Champions / Storage Manager
Business / Jonathan Crook / Research Support Manager: Charis Wilson / Jonathan Crook / Paul Caban
Research Support Manager: Lynn Walford
Research Support Manager: Caroline Leburn(Mon, Wed, Fri)
Divinity / Helen Bond / Research Officer: Karoline McLean / Arkotong Longkumer / Eli Donald
Economics / Jozsef Sakovics / Research/REF Administrator: Gina Reddie / Andrew Clausen / Ivan Salter
Maia Guell (on Sabbatical)
Edinburgh College of Art / Remo Pedreschi / RKEO Office Manager: Louise Fleming / Tahl Kaminer / Geoff Lee
Deputy : Mr Ed Hollis / RKE Administrator: Valentina Guerrieri / Genevieve Warwick
Research Administrator: Janet Black / Sean Williams
Health in Social Science / Heather Wilkinson / Research Administrator: Jane Richards / Nick Jenkins / Fraser Muir/Chris Kant
History Classics and Archaeology / Louise Jackson / Research Administrator: Morag Cherry (DDPS) / Martin Chick / Karen Howie
Literatures Languages and Cultures / Andrew Newman / Research Administrator: Eve Equi / James Loxley / Fiona Carmichael
Deputy (KE and Impact): Peter Dayan / Research Office Assistant: Anne Mourad / Andrew Newman
Law / Graeme Laurie / RKE Office Manager: Alison Stirling / Daithi Mac Sithigh / Nick Dyson
Moray House School of Education / Sheila Riddell / RKE Officer: Simon Temperley / Fraser Muir/Toby Morris
Research Administrator: Lesley Thomson
Philosophy Psychology and Language Sciences / Robert Logie / Research Manager: Susan Hamilton / Sergio Della Sala / Morag Brown
Research Administrator: Melissa McLaughlin
Social and Political Science / Vernon Gayle / Research Administrator: Marjorie Drysdale / Radhika Govinda / Ian McNeil
Deputy: Sotiria Grek / Research Office Assistant: Mr Craig Landt
Deputy (KE and Impact): Nicola McEwen
Institute / Director / Administration / Storage Manager
Office of Lifelong Learning / Con Gilllen / Nicola Davidson / Mark Galloway
Institute of Academic Development / Louisa Lawes (Assistant Director/Research Development) / Nicola Cuthbert / Karsten Moerman

Appendix B – Research Datastore allocation policy

Proposed RDM File-Store Allocation Policy

Version 2.7

This document presents the allocation policy for the Research Data Management (RDM) file-store, which is part of a suite of new services being rolled out across the University.

1.  The fundamental purpose of the RDM file-store is to provide each researcher with a guaranteed minimum amount of storage for research purposes. An allocation of space to the RDM file-store is provided much in the same way as an email account when joining the university.

2.  The file-store service provides a free at point of use allocation of 0.5TB for each researcher (a researcher is an academic member of staff or PGR student.)

3.  By default, the entire allocation to each researcher will be provided as a personal researcher space.

4.  A researcher can assign up to 50% (0.25TB) of their free individual allocation to shared project spaces, where those spaces are for research groups or projects in which they are actively collaborating.

5.  Additional capacity above the free allocation can be purchased as required. As appropriate, researchers should ensure that legitimate costs are recovered from grants to fund this additional capacity, or seek internal funding if grant funding is not available.

6.  It is expected that the primary use of the free storage allocation will be for research purposes – however it is accepted that some personal teaching or administrative data may be held on the RDM file-store and associated services.

7.  The allocation is for the RDM file-store and associated services (e.g. dropbox-like). Allocations to other services, such as data publication (DataShare) or Data Vault, are currently outside of this policy.

8.  The allocation and use policy will be reviewed in 12 months time (1st November 2014) to ensure it continues to be fit for purpose. A minimum of 0.5TB will always be free of charge, sharing will always be possible up to 50% of this 0.5TB, and the costs of charged storage will be minimised and as stable as possible.

Administrative Details

In the case that researchers choose to aggregate space into group space, then to allow the RDM file-store to be appropriately managed, the following directions must be followed:

1.  Where a research group spans schools they must be associated with a nominated school.

2.  There are three roles which must be fulfilled by Schools/units when using the RDM file-store

i.  Storage Manager – who will instruct Information Services on the construction and administration of group spaces.

ii.  Data Owner – who will be responsible for the data.

iii.  Storage Administrator - this role will technically administer group file spaces.

Above this, the Head of School or similar retains overall responsibility and control of data held within the area operational management should be devolved to the roles described above. Storage managers and data owners must be fulfilled by staff from the school/unit or group. If there are no appropriate technical staff within the school/unit or group to fulfil the storage administrator role, Information Services will take on this role. Roles can be fulfilled by multiple people. When an individual leaves, the school/unit must ensure a replacement takes on their responsibilities.

[1] Essentially all academic staff and research students

[2] School can be read as School, Group (e.g. IAD, OLL) or College Office.