INFN-GRID Globus evaluation (WP 1)

1
INFN-GRID Globus evaluation(WP 1)

• Massimo Sgaravatto
• INFN Padova
• for the INFN Globus group
• globus_at_infn.it
• http//www.infn.it/globus

2
Globus

• Some basic services (security, information
  service, resource management, ) must be
  implemented in order to implement and use a Grid
  for real applications
• Globus identified as possible Grid framework
  providing these services
• but it has been developed mainly for
  traditional computing, different from computing
  in HEP
• High performance vs. high throughput
• Supercomputers vs PC farms
• Distributed data intensive computing not
  addressed
• Need to assess what can be used for HEP
  environment
• ? WP1 Installation and Evaluation of the Globus
  Toolkit of the INFN-GRID Project
• Goal evaluation of the Globus toolkit
• Which services can be useful ?
• What is necessary to integrate/modify ?
• What is missing ?

3
Globus Architecture
Applications
High-level Services and Tools
GlobusView
Testbed Status
DUROC
globusrun
MPI
Nimrod/G
MPI-IO
CC
Core Services
GRAM
Nexus
Metacomputing Directory Service
Globus Security Interface
Heartbeat Monitor
Gloperf
GASS
4
Proposed work plan

• Security
• To access GRID resources mechanisms for user
  authentication and authorization needed
• ? Evaluation of GSI service
• Information Service
• To discover the GRID resources (CPU, storage,
  network, ) mechanisms to publish them must be
  defined
• ? Analysis of GIS service to publish
  information using a uniform and standard
  interface
• Resource Management
• Necessary a uniform interface to submit jobs on
  GRID resources
• Uniform standard interface to different resource
  management systems
• Uniform standard language for task management
• ? Assessment of Globus services for resource
  allocation and process management

5
Proposed work plan

• Data Access and Migration
• High performance and reliable tools needed to
  manage data (access to remote data, data
  transfers, wide area replica, )
• ? Assessment of Globus tools for data management
  (GASS, Globusftp)
• Fault Monitoring
• Faults in a GRID environment must be promptly
  detected and recovery mechanisms must be
  implemented
• ? Evaluation of HBM service for fault detection
• Execution Environment Management
• Code migration (moving the application where the
  job will actually be executed) as a possible
  implementation strategy
• ? Evaluation of GEM service to support code
  migration
• Globus installation tools
• Reduce complexity and manpower for Globus
  installation and maintenance

6
Globus installation tools

• Flavias presentation
• INFN-GRID installation tool to shorten the
  installation time of the Globus toolkit, avoid
  common mistakes, support for specific
  customisations
• Possibility (option) to install optional
  software, to proceed with INFN specific
  customizations (INFN CA, configuration of a
  hierarchical GIS architecture), to install and
  use specific INFN tools
• Proven to be successful within INFN (used to
  setup a INFN GRID testbed) and also outside
  (CERN, FNAL, )

7
Security

• Evaluation of Globus GSI
• User authentication (implementation based on
  X.509 certificates)
• User authorization managed by grid-mapfile
  (mapping between Grid users and local users)
• Some shortcomings, but the GSI security model
  seems to satisfy our requirements
• Some shortcomings already addressed
• INFN-CA used to sign certificates
• CRL (issued by INFN CA) distribution
• Centralized management of grid-mapfiles

8
Security

• Centralized management of the grid-mapfiles
• Goal Ease the sharing of the same access
  policies (represented by the grid-mapfiles) for
  groups of hosts with common purposes
• Proposed system
• Central repository (LDAP server) to store user
  certificates and to define groups of users
• Certificates published by CA manager
• Group manager responsible for editing group
  memberships (using a LDAP client)
• Resource owners (Globus administrators)
  periodically (i.e. cron job) connect to this
  repository, download the subject of the
  certificates that meet a specified criterion
  (i.e. all users of group X), and produce
  grid-mapfile entries

9
Security

• AFS tests
• Analysis of what can be done now with the
  existing tools (quite unfit for any real need)
• Possible ways to address the existing
  shortcomings identified
• New Globus tool (gsiklog) available

10
Information Service

• Alessandros presentation
• Evaluation of Globus GIS (Grid Information
  Service)
• Definition and implementation of a hierarchical
  architecture of GIS 1.1.3
• Performance and scalability tests
• Web interface for browsing
• Various shortcomings must be addressed (to use
  the GIS in a production environment)
• Mixed push/pull model more suitable than a pull
  model
• Performance
• Lack of security

11
INFN GIS Topology
Dcinfn,dcit, ogrid
Top Level INFN GIIS
Expcms, ogrid
Dcbo, Dcinfn, dcit,ogrid
INFN CMS GIIS
Dcpd,Dcinfn, dcit,ogrid
GIIS
GIIS
GRIS
Padova
Bologna
12
Resource Management

• Most of these activities as collaboration with
  Grid Workload Management work package
• Evaluation of Globus resource management
  architecture
• Evaluation of Globus GRAM
• Tests with fork, Condor, LSF and PBS as
  underlying resource management systems
• The model is fine, but lack of robustness
  (needed for real production environments)
• Memory leaks in the Globus job manager (fixed)
• Scalability (one job manager for each job)
• Reliability (the job manager is not persistent)

13
Globus resource management architecture
(simplified design)
Resource Discovery
RSL
Broker
Grid Information Service (GIS)
Submit jobs
Broker chooses in which resources to submit the
jobs (not implemented in the Globus framework)
RSL
Information on characteristics and status of
local resources
RSL
Globus GRAM as uniform interface to different
local resource management systems
Globus GRAM
Globus GRAM
Globus GRAM
Local Resource Management Systems
CONDOR
LSF
PBS
Site1
Farms
Site2
Site3
14
Resource Management

• Evaluation of GRAM API
• Evaluation of GRAM Reporter (cooperation
  between GRAM and GIS) in particular for farms
• Many useless attributes (at least for our needs),
  attributes not calculated (always defined as
  0), some attributes not properly calculated by
  Globus shell scripts
• Some important information describing the farms
  and the submitted jobs (necessary for example for
  a resource broker) missing
• ? Draft proposal for a possible modification of
  the default schema
• Evaluation of RSL as uniform language to specify
  resources
• More flexibility required
• Submission of Condor jobs to Globus resources
• Condor-G (useful as a reliable crash-proof job
  submission service)
• GlideIn
• Evaluation of MPICH-G2 vs. MPICH
• Some shortcomings found (lack of support for
  shared memory, worse latency performance wrt.
  MPICH)

15
Data management

• Tests with GASS
• Service to ease the access to remote files
  without having a distributed file system and/or
  transferring files from/to remote storage systems
  
• Tests with command line tools and APIs
• Problems (huge decrease in transfer rate) when
  transferring big files
• Tests with Globusftp alpha release 2
• Collaboration with WP network INFN-GRID
• Tests of new features
• Support for GSI mechanisms
• Capability of resuming interrupted file transfers
• Throughput tests using parallel data transfers
• Antonios presentation

16
Other services

• Fault Monitoring (HBM)
• Evaluation of HBM for fault detection (for
  system and user processes)
• but the HBM package is not seeing active
  development
• Execution Environment Management (GEM)
• Evaluation of GEM as service for code migration
• but the GEM service now provides only limited
  capabilities (executable staging)

17
WP 1 Deliverables Milestones

• Deliverables
• Tools, documentation and operational procedures
  for Globus deployment (6 Months) ?
• Final report on suitability of the Globus toolkit
  as basic Grid infrastructure (6 Months) ?
• Milestones
• Basic deployment Grid infrastructure for the INFN
  GRID (6 months) ?
• Globus installed on 40 machines on 10
  different sites

18
Conclusions

• The activities of WP 1 are over
• The Globus toolkit can provide basic services
  useful to create and deploy usable Grids, but
  many shortcomings and issues must be addressed
• more details in the report
• Other info http//www.infn.it/globus