Document Filename: TechnicalDescription.pdf
Type of document: Technical Description
Version: 0.2
Document classification: PUBLIC
Abstract: This document contains the technical description for the infrastructure and hardware specification of the Resource Centers, hosted in the University of Cyprus and in the University of Nicosia.
2 GENERAL DESCRIPTION
... 4 3 CYGRID TESTBEDS
... 5
3.1 CY01KIMON TESTBED DIAGRAM
... 5 3.2 CY03INTERCOLEGE TESTBED DIAGRAM
... 6
4 SUPPORTED SERVICES
... 7
1.1 AMGA SERVICE
... 7 1.2 WORKLOAD MANAGEMENT SYSTEM (WMS)
... 7 1.3 INFORMATION SERVICE (BDII)
... 7 1.4 USER INTERFACE (UI)
... 8 1.5 COMPUTING ELEMENT (CE)
... 8 1.6 STORAGE ELEMENT (SE)
... 8 1.7 MONITORING SYSTEM (MON)
... 9 2 HARDWARE ... 10 1.1 CY01KIMON ... 10 1.2 CY02CYGRIDCERT HARDWARE ... 12
p
1
Document Log
Versio n
Date Summary of changes Author
0.1 28/01/08 Initial Contribution Andoena Balla 0.2 30/01/09 Revision Nikolas Stylianides
•
1
General description
CyGrid was initiated by the High Performance Computing systems Laboratory (HPCL) at the University of Cyprus during the year 2001-2002. The CyGrid initiative, aims at establishing a contact point for local scientific community users who are interested in Grid computing by promoting the technology and encouraging both academic and industrial researchers to become familiar with it. Once the potential benefits of the technology are realised, academic/industrial researchers can experience the benefits from deploying their applications on the CyGrid testbed and beyond. Cygrid support and provides knowledge to organizations that are interested to establish new resource centers. In 2007, the University of Cyprus and the University of Nicosia together with the Cyprus Research & Academic Network CyNET and the Ministry of Health signed a Memorandum of Understanding with the intention of providing a framework for research and academic cooperation, aiming at the enhancement of the existing infrastructure in Cyprus.
To assist in the enhancement of the local infrastructure and enable researchers to safely use the Grid, CyGrid has established two Resource Centers, hosted in the University of Cyprus and in the University of Nicosia (former Intercollege), and connected to the European Grid Infrastructure of EGEE.
b
1
Cygrid Testbeds
ÈÈÈ
1
Supported Services
1.1AMGA service
AMGA (ARDA Metadata Grid Application) is an interface for metadata access on the Grid, designed to be scalable and support flexible schema management, efficient retrieval of large result sets and to allow a broad range of implementations. AMGA can be regarded as a simplified database that can be used from jobs running on the Grid, which often need to retrieve or store metadata, in an "attribute-value" way. The AMGA server can use many different database servers for a backend, so it can answer SQL-like queries concerning the metadata stored. This makes it a very good choice for storing and retrieving metadata about files stored on Grid Storage Elements.
1.2
Workload Management System (wms)
The WMS (Workload Management System) is the gLite 3.1 component that allows users to submit jobs, and performs all tasks required to execute them, without exposing the user to the complexity of the Grid. It is the responsibility of the user to describe his jobs and their requirements, and to retrieve the output when the jobs are finished.
1.3
Information service (bdii)
The first level of MDS (Monitoring and Discovery Service) information publication is the GRIS, which provides specific information for a particular service. The GRIS
normally runs on the same node as the CE, SE or other service for which it publishes, although it may be on a different node.
The BDII collects information about all resources present at site where the service is installed (i.e. data from all GRISes at the site).
1.4
User Interface (ui)
Apart from registering with EGEE, a user must also have an account on a EGEE User Interface in order to access the Grid. To obtain such an account, a local system administrator must be contacted, either at the user’s own site or at a central site like CERN.
1.5
Computing Element (ce)
A CE (Computing Element), in Grid terminology, is some set of computing resources localized at a site (i.e. a cluster, a computing farm). A CE includes a GG1 (Grid Gate) , which acts as a generic interface to the cluster; a LRMS (Local Resource Management System), sometimes called batch system, and the cluster itself, a collection of WNs (Worker Nodes), the nodes where the jobs are run.
1.6
Storage Element (se)
A SE (Storage Element) provides uniform access to data storage resources. The SE may control simple disk servers, large disk arrays or tape-based MSS (Mass Storage Systems). SE can support different data access protocols and interfaces. Simply speaking, GSIFTP (a GSI-secure FTP) is the protocol for whole-file transfers, while local and remote file access is performed using RFIO or gsidcap. Most storage resources are managed by a SRM (Storage Resource Manager) a
to tape, file pinning, space reservation, etc.
1.7
Monitoring System (mon)
The ability to monitor resource related parameters is currently considered a necessary functionality in any network. In such a heterogeneous and complex system as the Grid, this necessity becomes fundamental. A monitoring system implies the existence of a central repository of operational information. The monitoring system should be able to collect data from the resources in the system, in order to analyze the usage, behavior and performance of the Grid, detect and notify fault conditions, contract violations and user-defined events.
2
Hardware
1.1 CY01KIMON
• 1 Generic Server thales.grid.ucy.ac.cy
o IBM xSeries 335 (2x Intel Xeon HT 2.8GHz CPUs, 2GB RAM, 2x200GB
HDD)
• Generic Server euclid.grid.ucy.ac.cy
o 24GB RAM , 8CPUs each 2.83GHz, 2x500GB HDD
o EGEE Disk Pool Manager (DPM)
o 3.8TB in RAID 5 configuration, giving 3.2TB of storage space
o 1 IBM DS400 SAN Storage System (14 Hard Disks SCSI 300GB)
o 1 IBM eServer 326m Machine, DS400 controller
o NFS mount to EGEE computing element
• 2 Regional Grid Resource Brokers
o rb101.grid.ucy.ac.cy (IBM ThinkCentre, 1x Intel Pentium 4 HT
3.6GHz / 1MB L2, 2GB RAM)
o wmslb101.grid.ucy.ac.cy (SuperMicro server, 2x Intel Xeon 2.4GHz
CPUs, 4GB RAM) BEST EFFORT SERVICE
• 1 Regional Grid Information Index (BDII) bdii101.grid.ucy.ac.cy
o IBM ThinkCentre (1x Intel Pentium 4 HT 3.6GHz / 1MB L2, 1GB RAM)
• 1 Grid Gate (Computing Element) ce101.grid.ucy.ac.cy
o Super Micro (2x Intel Xeon 2.4GHz CPUs, 4GB RAM)
• 35 Grid Worker Nodes (82 CPUs) wn107 to wn141.grid.ucy.ac.cy
o 7 IBM eServer 326m Machines (2.0GHz dual AMD Opteron 64-bit CPUs,
1GB RAM)
o 1 IBM SuperMicro Machines (2.4GHz dual Xeon CPUs, 1GB RAM)
o 22 IBM eServer 326m Machines (2.6GHz dual AMD Opteron 64-bit
CPUs, 2GB RAM)
o 5 IBM system x3455 Machines (2.6GHz dual AMD Opteron dual-core
• 1 Grid User Interface ui101.grid.ucy.ac.cy
o Fujitsu-Siemens (1x Intel Pentium 4 2.0GHz CPU, 1GB RAM)
• 1 Network Monitoring node mon101.grid.ucy.ac.cy
o Fujitsu-Siemens (1x Intel Pentium 4 2.0GHz CPU, 1GB RAM)
• 1 Micronet Gigabit Ethernet switch 24 ports
• 1 CISCO Gigabit Ethernet switch 48 ports (module)
The cluster is connected with GÉANT via an ATM switch at the Network centre of the
University of Cyprus over 155Mbps fiber optics links.
1.2 CY02CYGRIDCERT HARDWARE
• 2 x IBM xSeries 335 (2x Intel Xeon 2.8GHz CPUs, 1GB RAM, 40GB HDD)
VIRTUAL MACHINE HOSTS o ce201.grid.ucy.ac.cy o wn201.grid.ucy.ac.cy o wn202.grid.ucy.ac.cy o mon201.grid.ucy.ac.cy o wmslb201.grid.ucy.ac.cy o bdii201.grid.ucy.ac.cy
• 1 x Dell PowerEdge (2x Intel P3 800MHz CPUs, 768MB RAM, 2x35GB
• 1 x Dell Precision 330 (1x Intel P4 1.5GHz, CPU, 512MD RAM, 40GB HDD)
o ui201.grid.ucy.ac.c
• 1 x Generic PC (1x Intel P4 1.7GHz, CPU, 256MD RAM, 40GB HDD)
o amga201.grid.ucy.ac.cy
• 1 x Generic PC (1x Intel P3 800MHz, CPU, 384MD RAM, 40GB HDD)
o lfc201.grid.ucy.ac.cy
The cluster is connected with GÉANT via an ATM switch at the Network centre of the