File contents

FNAL Dcache Project
FY08 WBS V1 
28 June 2007 
by Timur Perelmutov

based on
V2.3.1
10 January 2007
Rob Kennedy

High Level WBS
1. Data Movement & Caching/Upper Storage/Dcache/Operations 2.1FTE
    1.1.Day-to-Day production operations 1 FTE
    1.2 Operations / Other Services 0.2 FTE
    1.3 Operations / Worldwide Support 0.9 FTE
2. Data Movement & Caching/Upper Storage/Dcache/Resilient Manager Development 0.5 FTE
3. Data Movement & Caching/Upper Storage/Dcache/Development 2.3 FTE
    3.1 Development /Namespace 0.2 FTE
    3.2. Development / Deployment Re-org: major project, transition 0.2 FTE
    3.3 Development / GridFTP 0.3 FTE
    3.4 Development / Infrastructure 0.5 FTE
    3.5 Development/ Monitoring 0.1 FTE
    3.6 Development/ Utilities 0.5 FTE
    3.7 Development / Other Features (incl. per node limits) 0.3-0.5 FTE 
    3.8 Development / Technology review 0.2 FTE
4. Collaborations	
5. Data Movement & Caching/Upper Storage/Dcache/Project Management 0.2 FTE
6. Data Movement & Caching/SCIDAC2/CEDS 0.5 FTE

Total 5.4 FTE
Staff Involved
----------------------------------------------------------------------------

Staff:   50% Timur P. - in the dcache primary shift rotation
           100% Alex K.
           100% Vladimir P.
           ..% Ted H. - CMS-specific support, GPlazma development and integration (charged to SRM)
           50% Andrew B.
           ----
           2.5

Openings:  0.00
New Hires: 2.00

Detailed WBS
1. Data Movement & Caching/Upper Storage/Dcache/Operations
    1.1.  Day-to-Day production operations 1 FTE
        1.1.2 Upgrades and Installations
        1.1.3 Operations Issues (introduce palliatives, if necessary)
        1.1.4 Facility/Customer Requirements, Config, Planning
        1.1.5 Administrative procedures, scripts, and related doc
        1.1.6 Developer Shift System Management (scheduling, duties)
        1.1.7 Communications: Operations notices and reporting
        1.1.8 Integration Services: customer accessible test stands
  1.2 Operations / Other Services .2 FTE
        1.2.1 PNFS - with other groups
        1.2.2 Chimera - after deployment
        1.2.3 Postgresql used by PNFS, DCache, SRM - with other groups
        1.2.4 Apache and dCache internal httpd
  1.3 WorldWide support .9 FTE
	1.3.1 US-CMS T2 sites support .2 FTE
        1.3.2 Non US-CMS OSG dCache sites .3 FTE
        1.3.3 The Rest of WLCG dCache sites .4 FTE

2. Data Movement & Caching/Upper Storage/Dcache/Resilient Manager Development 
      2.1 Near-zero administration features
      2.2 General V2 design, if needed
      2.3 Specialize configs: file in RAID + any other disk
      2.4 Pool scheduling coordination with PoolManager, eg.
                        in order to replicate on distinct nodes
      2.5 Integrate with central space management in dCache
3. Data Movement & Caching/Upper Storage/Dcache/Development 
     3.1 Development /Namespace
           3.1.1 Evaluation and Testing of Chimera 
           3.1.2 Evaluation and Testing of multiple PNFS under one dCache, if needed
     3.2. Development / Deployment Re-org: major project, transition
        3.2.1 Decouple from enstore CVS (encp, log server, etc)
        3.2.2 Adapt to dcache.org RPMs, under user ~dcache
        3.2.3 Package watchdogs, monitors, plotting, KDC mux
    3.3 Development / GridFTP: in hiatus due to lack of staff
        3.3.1 FTP command support: a few basic commands, most
                remaining to be done are gridftp-related
        3.3.2 On-demand start-up of (gridftp) doors by SRM to
                enhance scalability, reduce redundant copies
        3.3.3 Light-weight, distinct SocketAdapter for both Active and Passive mode (currently used as active only)
        3.3.6 Permit multi-door/multi-NIC configs on a node
        3.3.7 Gridftp X-mode development (done by NDGF)
        3.3.8 Test/validation suite for common "supported" clients
        3.3.9 FTP: Default "ls" should not check file permissions
        3.3.10 FTP: Command-specific monitoring via logs and/or DB
        3.3.11 FTP: Mitigation of high-performance transfer issues in OS
    3.4 Development / Infrastructure
        3.4.1 Packaging and Distribution (RPMs, FSL workgroup,...)
        3.4.2 Automated Build and Test Platform
        3.4.3 Dev Process Definition with Issue Tracking and Workflow
        3.4.4 Documentation and Tutorials
        3.4.5 FNAL Dcache Plone site (and scattered web sites)
        3.4.6 Storage developer test stand (WH8SE, FAPL): procure, operate
    3.5 Development/ Monitoring
         3.5.1 Plots - improve, integrate, DB use, RPM deploy
         3.5.2 Improve Gridftp, current transfers, current state monitoring
    3.6 Development/ Utilities
         3.6.1 Inventory and consolidation of existing utilities (scans, alarms, admin scriptlets, etc.)
         3.6.2 For Some Utilities:  Identification of dCache defects responsible for their existence and addressing them
         3.6.3 For Some Other: Conversion into internal dCache services
         3.6.4 Others should be developed into centrally supported and destributed products
    3.7 Development / Other Features
         3.7.1 HSM Interface, real-encp in python
         3.7.2 Logging - improve, archive
         3.7.3 Dcap and clients - as driven by FNAL customers
         3.7.4 Per-Node Limits - independent of per-pool limits
         3.7.5 Experience-driven development (replace palliatives) - LONG LIST
    3.8 Development / Technology review
         3.8.1 Investigate other caches, storage Systems
         3.8.2 File Systems - investigate f/s, f/s-specific features
         3.8.2.1 XFS and direct I/O - re-enable?
         3.8.2.2 ext4 versus XFS
         3.8.2.3 XFS and ext4 large entents feature
         3.8.3 Operating Systems
         3.8.4 Investigate Products that can replace parts of Dcache
         3.8.4.1 Namespace
         3.8.4.2 Schedulers
         3.8.4.3 File Repositories
         3.8.4.4 Transfer Protocol implementations
         3.8.4.5 Administration
         3.8.4.6 Messaging
4 Data Movement & Caching/Upper Storage/Dcache/Collaboration
    4.1 dCache
    4.2 OSG
    4.3 WLCG
    4.4 CEDS
    4.5 Other Grid Projects
    4.6 General Outreach: Conferences, Workshops, etc.
5. Data Movement & Caching/Upper Storage/Dcache/Project Management
     5.1 Group Steering: biweekly meetings and reports
     5.2 Section and Other Management meetings
     5.3 Groupware to reduce overhead, improve accessibility of info
     5.4 Hiring-related activities
     5.5 Timesheets, Effort Reports, etc.
     5.6 CompDiv Projects: reports, presentations 1-2 times per year
     5.7 Project WBS, Budget: once per year with follow-up
     5.8 Long-term facility/program-oriented planning: once per year
     5.9 Project tactical planning: roughly 4 times per year
     5.10 Funding, developer help: grants, SBIR, new collaborations
6. Data Movement & Caching/SCIDAC2/CEDS
     6.1 Integrity Verification features of GridFTP2
     6.2 Integration with another gridftp code base?
     6.3 Investigations
     6.4 Further CEDS activities