VISTA DATA FLOW SYSTEM
(VDFS)
---------------
for VISTA & WFCAM data









Management and Planning

author
P.M.Williams (WFAU Edinburgh)
WFAU Manager
number
VDF-WFA-VSA-003
issue
Issue 1.0
date
13 September 2006
co-authors
A.Lawrence, N.Hambly


Contents


SCOPE

This document considers and addresses the issues posed by management of the development of the Science Archives for the WFCAM on UKIRT (WSA) and the VIRCAM on VISTA (VSA). It identifies the resources available for the work. It describes the mechanisms for reporting and monitoring progress. It also considers the standards adopted by those working on the project and assesses the risks faced.


INTRODUCTION


Context

This document is part of the reference documentation for the VDFS Project.

The development of the VISTA Date Flow System (VDFS), including that of the WFCAM and VISTA Science Archives considered here, is funded by PPARC through a specific VDFS grant. This work is part of the programme of Wide-Field Astronomy and e-science at the Institute for Astronomy funded by PPARC through a Rolling Grant. The programme was last reviewed in 2004-5, and funds were sought, and granted, for the operation of the Science Archives. Despite separate funding routes, the development and operational staff work very closely together.

The WSA work, in addition to its intrinsic importance, was seen as a step towards (and test-bed for) the VSA. For this reason, it was agreed that the two be planned and designed together so that we could become more confident that the planned Archive is scalable to the data volumes expected from the VISTA IR camera, and ensure that no design decisions were taken which militated against scaling for use for VISTA science. It was also necessary that the archive be VO-compatible as soon as possible.


Stakeholders

Although the WSA is managed as part of VDFS, other groups, to whose needs VDFS is sensitive, have a strong interest in the products. This group includes the Joint Astronomy Centre, which is responsible for the operation of UKIRT and WFCAM, and the UKIDSS Consortium, which planned and is currently undertaking a set of public infrared surveys using WFCAM for most of its time on the telescope. These groups were instrumental in generation of the Science Requirements leading to the first Science Requirements Analysis Document (SRAD, AD01) and, following the release of WFCAM Science Verification data and publication of Data Release 1, have provided valuable feedback on the data delivery tools and user-interface of the of the Archive.

The group of stakeholders has broadened during the progress of the Project, and now includes astronomers from the ESO community.

The project is overseen by the VISTA Data Users' Committee (VDUC), which was constituted to provide independent and scientifically informed advice to the VDFS Project on behalf of VDFS's UK user community.

The VDUC is chaired by a nominee of PPARC and the membership includes the UKIDSS Survey Scientist, the VISTA and AstroGrid Project Scientists, and four other representative active researchers nominated by the UKIRT Board, the VISTA Project Board, the UKIDSS Consortium and the VISTA Consortium. It first met in 2004 May.


PROJECT DEFINITION AND MANAGEMENT


Definition

The Science Requirements Analysis Document (SRAD, AD01), based on the UK User Requirements document (URD, AD02) developed by the VISTA Project Scientist, is the primary determinant of the work to be undertaken.


Schedule

The original project plan set out in the 2004 VEGA-VDFS proposal described five versions (VDFSv1-v5) of the Science Archive, to be released annually at the ends of 2003-2007. Versions VDFSv1-3 were for WFCAM data, offering progressively more functionality, while VDFSv4 is the WFCAM archive scaled for VISTA VIRCAM data volumes, required to be ready for VIRCAM commissioning in December 2006, and VDFSv5, is the Archive fully shaken down with VIRCAM science data. Owing to the delay in delivery and commissioning of WFCAM, the time-scale between WFCAM coming into operation and the scheduled VISTA VIRCAM delivery has been greatly compressed, necessitating revision of the plan. The following revised plan was presented (AD03) to the VDUC at its meeting on 2005 September 6.

The principal change schedules preparatory work for the VISTA-ready version, VDFSv4 = VSA1 (scalability of WSA design, OS, DBMS and hardware for VISTA data volumes) in parallel with further development of data access tools for the WFCAM Science Archive (WSA3). Some of the design work for the VISTA Science Archive (VSA), namely the database schemas for the public surveys, can only be completed when the surveys themselves have been defined and accepted by ESO. This process is currently underway.

The top-level project plan is divided into five phases (we use the term `phases' to allow the term `versions' to be used for processed data):

  1. WSA1: WFCAM Science Archive ready for receiving the first WFCAM data.

    As described in the Project Overview (AD03), in order to provide experience with scaling SQL Server to databases larger than the 6dF Survey database developed in 2002, this phase of the programme included the building of a copy of the 4-TByte SuperCosmos Object catalogue on SQL Server (the `SSA'). This provided time for tuning the DBMS and Catalogue Server hardware and also a platform for testing the data access tools as they were developed, which was valuable as there was some delay in delivery and commissioning of the WFCAM.

  2. WSA2: WFCAM Science Archive with experience of WFCAM science data (2004 October - 2005 September).

    It was recognized that the `Day 1' archive (WSA1) would not provide all the functionality required by the SRAD and that there would also be some issues arising from the instrument itself and pipeline that would require re-factoring of some of the data ingest, curation and delivery tools. WSA2 was envisaged to be available one year after survey operations began but, because of the delay in WFCAM commissioning and the need to adhere to the VISTA time-table, we proposed closing this phase at the end of September 2005. The work on WSA2 was envisaged to benefit from feedback from members of the UKIDSS consortium working on Science Verification of the system, as occurred at the Science Verification meeting on August 18.

  3. WSA3: WFCAM Science Archive with advanced data curation and access tools (2005 October - 2006 March)

    From October 2005, part of the WFAU VDFS team continued to work on the WSA, WSA3, operating the archive and ingesting re-processed data, re-factoring database schemas and code in the light of experience and feedback from users, and adding additional facilities for catalogue maintenance and data delivery. This phase was extended, leading to the public Early Data Release (EDR) on February 10 and Data Release 1 (DR1) on July 25.

  4. VSA1: VISTA Science Archive ready for receiving the first VIRCAM data (2005 October - 2006 December).

    The first phase of work from October 2005, in parallel with work on WSA3, was benchmarking of the DBMS/OS (MS SQL Server running under Windows) and hardware adopted for the WSA to see whether they were still appropriate for the VSA. The outcome from these studies is given in AD04. If a hardware architecture different from that hosting the WSA had been indicated, this would have given us time to procure and install the first nodes of the servers required for the VSA and iron out any bugs.

    Following DR1, we proceed to the following design work, leading to documentation for the present Final Design Review (FDR):

    Following the FDR, and in the light of its outcome, we will proceed with the implementation of the design. Further hardware for data servers will be acquired and installed as required. Data transfer between CASU and WFAU using UKlight will be implemented.

    Towards the end of the construction phase, we expect to receive specimen files from CASU based on engineering data to test the Curation Tasks. It is hoped that the data will include some on-sky frames, but this depends on the progress of the installation and engineering commissioning of VISTA and VIRCAM.

  5. VSA2: VISTA Science Archive shaken down with experience of VIRCAM data. (originally 2007 January - September)

    We expect a period of shake-down of the data-flow system and VSA once real data flow from VISTA. In the light of operational experience, and user feedback, we may have to re-factor some of the tools for data ingest, catalogue maintenance and data delivery. We expected that there would be time for feedback and re-factoring before the current VDFS funding ceases on 30 September 2007.

    Given that the scheduled date for handover of VISTA + VIRCAM to ESO, and the beginning of Science Commissioning is scheduled for August 2007, there is practically no time for this work and, as described below, funds are being sought from PPARC to extend the VDFS grant to allow it to be done.

    In any event, during this phase, we will add further data access tools as specified in the UK VISTA User Requirements (URD 3.0) and upgrade our VO-compatibility.

The programme is summarized in a Gantt chart (Fig. 1), which marks the significant milestones.

Figure 1: Top level Project Gantt chart. The open bars mark external tasks.
\includegraphics[width=6.9in]{VSAtop}


Management

The VDFS management structure is shown in Figure 2, where the Management Team are Emerson (VDFS PI), Lawrence (VDFS co-I), McMahon (VDFS co-I), Stewart (VISTA Software Manager), Irwin (CASU Manager) and Williams (WFAU Manager). Stewart's particular rôle is in ensuring coordination of the VDFS ESO deliverables with the VISTA camera work.

Figure 2: Management structure of the VDFS Project
\includegraphics[width=4.0in]{VSAmgt}

There are four time-scales (funding, quarterly, monthly, weekly) on which the progress of the Archive work is assessed and managed.

The longest is that defining the overall long-term aims. The VDFS Project is currently funded from 2004 October - 2007 September by PPARC through grants to WFAU and CASU.

On a calendar Quarterly cycle, detailed work plans and deliverables are defined and the work allocated amongst the available staff. This is regularly monitored throughout the quarter (see below) and at the end of the Quarter the progress against the plan and deliverables is assessed, and a new Quarterly plan is made which takes account of the successes, and any failures, in the previous quarter. The quarterly past progress and future plans are closely scrutinised by the PI+Management Team, and used as the basis for reporting to the Grid Steering Committee (at intervals of approximately six months).

On a monthly cycle, the WFAU Manager reports progress against plans to the Management Team who discuss them. This allows corrective actions to be taken within the quarterly time-scale. At the same time, the increases in Earned Value and expenditure on the VDFS Grant are reported to the VDMT.

Figure 3: Example of the quarterly chart on which progress is reported to the VDMT at its monthly meetings. The numbers under the months in the last three columns are pencentages complete, reported to high granularity (10% = started, 50% = nearly complete and 100% = complete).
\includegraphics[width=7.9in]{VDFS06Q3}

Figure 4: Plot of percentage Earned Value (EV) and VDFS grant expenditure for the first seven quarters of the Project.
\includegraphics[width=4.0in]{EVAchart}

On a weekly cycle, the staff working on the Archive project meet to review progress, allocate tasks and actions. These meetings are minuted and copies of the minutes are provided to the VDFS PI+Management team. They are also available to the JAC.

The Project maintains a Twiki collaborative web site on which all documents, formal and informal, are shared. These include notes from meetings and external visits, equipment set-ups and experiments with different hardware and software.


RESOURCES


Staff

Staff Effort for the Science Archive component of the VDFS Project is provided by members of the Wide Field Astronomy Unit. Those assigned to the Archive development, together with fractions of their effort as funded through the 2004-07 WFAU VDFS Grant, are:

The following effort is funded by the 2005-10 WFAU Rolling Grant:

The Project also receives specialist assistance and advice, particularly in the field of large databases from Dr R. Mann (WFAU, NeSC and AstroGrid). We also receive help and advice from NeSC and the University of Edinburgh, IBM and Microsoft.

As noted above, the VDFS Grant funding the development runs until 30 September 2007. After that date, one member of the development team (Dr Sutorius) is funded to move to Science Archive operations and the remainder of the team is expected to be dispersed to other projects. This is scheduled to occur about one month after the scheduled date for the beginning of VISTA+VIRCAM Science Commissioning, leaving practically no time for the Phase 5 work shaking down the Archive with experience of science data described above.

Funds are therefore being sought from PPARC to extend the VDFS Grant by six months to 31 March 2008, the review date of the WFAU Rolling grant. Proposals for the Rolling Grant will be prepared in March-April 2007 and will include requests for funding for:


Equipment

Preliminary work on the Project made use of equipment provided by the University of Edinburgh or to the University for the Wide Field Astronomy programme through a major JREI grant and the WFAU PPARC Rolling Grant. This included a 6-TB RAID system used for the SSA. The University also provided file server systems for the WSA Catalogue Servers. The VDFS grant has funded purchase of another Catalogue Server and funding for the pixel data servers is provided by PPARC though the WFAU Rolling Grant.

The equipment identified as being required for the Project is described in the Hardware/OS/DBMS design document (AD04). The equipment will be purchased in accord with standard University of Edinburgh procedures, including tendering where required, with funding provided for Archive Operations in the WFAU Rolling Grant.

__oOo__


APPENDICES


Standards


Risk Assessment

We have considered the risks, both external and internal, which might jeopardise completion of the Project on time. We estimate the likelihood (L), effect (E) and impact (I = L x E) of each event, scored on a scale of 1-3, and give for each our mitigation strategy. This table will be kept in a separate Risk Register and the risks re-assessed periodically.


Event L E I Mitigation
Serious illness or loss of staff 1 3 3 de-scope: delay lower priority tasks
Network too slow for data transfer 1 3 3 use same method as for JAC and ESO to CASU
CASU/external tools delivered late 1 2 2 use preliminary versions supplied
Failure to make interfaces work 1 3 3 seek specialist advice
Development funds cease too early for shake-down with real data 3 3 9 seek extended funding



ACRONYMS & ABBREVIATIONS

ADnn : Applicable Document No nn
CASU : Cambridge Astronomical Survey Unit
DBMS : Database Management System
ESO : European Southern Observatory
JAC : Joint Astronomy Centre (Hawaii)
JREI : Joint Research Equipment Initiative
NeSC : National e-Science Centre
SRAD : Science Requirements Analysis Document (AD01)
UKIDSS : UKIRT Infrared Deep Sky Survey
UKIRT : United Kingdom Infrared Telescope
VDUC : VISTA Data Users' Committee
VIRCAM : VISTA Infrared Camera
VISTA: Visible and Infrared Survey Telescope for Astronomy
VO : Virtual Observatory
VSA : VISTA Science Archive
WFAU : Wide Field Astronomy Unit (Edinburgh)
WFCAM : (UKIRT) Wide Field Infrared Camera
WSA : WFCAM Science Archive


APPLICABLE DOCUMENTS


AD01 Science Requirements Analysis Document VDF-WFA-VSA-002

Issue: 1.0 20/09/2006

AD02 UK VISTA User Requirements VDF-SPE-IOA-00009-0001

Issue: 3.0, 5/07/2005

AD03 VISTA DATA-FLOW SYSTEM SCIENCE ARCHIVE (VSA): REVISED PLAN VDUC(05)09

06/09/2005

AD04 VSA Hardware/OS/DBMS design VDF-WFA-WSA-007

Issue 1.0 02/09/06



CHANGE RECORD


Issue Date Section(s) Affected Description of Change/Change Request Reference/Remarks
Issue 1.0 13/09/06 All New document


NOTIFICATION LIST

The following people should be notified by email whenever a new version of this document has been issued:


WFAU:P Williams, N Hambly
CASU: M Irwin, J Lewis
QMUL:J Emerson
ATC:M. Stewart
JAC:A. Adamson
UKIDSS: A. Lawrence, S. Warren

__oOo__

About this document ...

This document was generated using the LaTeX2HTML translator Version 2002-2-1 (1.71)

Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, 1999, Ross Moore, Mathematics Department, Macquarie University, Sydney.

The command line arguments were:
latex2html -html_version 3.2,math,table -toc_depth 5 -notransparent -white -split 0 VDF-WFA-VSA-003-I1

The translation was initiated by Nigel Hambly on 2006-09-30


Nigel Hambly 2006-09-30