GMA Implementations

The Grid Monitoring Architecture (GMA) is an abstract description of the components needed to build a scalable monitoring system. This page lists implementations which contain these components. It is important to note that the implementations listed here are not necessarily interoperable, but instead they all have producers, consumers, and a directory service that can interact with some subset of the defined GMA interactions.

Implementation	Description	Authors	Links
pyGMA	Python module providing SOAP/Web-Services GMA interfaces .	DIDC Group , LBNL	SourceForge site
GrADS	The Grid Application Development Software Project (GrADS) is funded by NSF and based at Rice University.  The GrADS project is focused on building an integrated Grid execution environment that will enable easy and effective use of Grid resources, with the primary goal of providing predictable application performance.   This project brings together researchers in library development, compiler design, Grid middleware development, scheduling, performance prediction, and performance monitoring.  Areas of research include just-in-time compilation, Grid economies, and the use of performance contracts to formally state expected resource allocation and application progress.  If the contract is violated, the execution environment detects the violation and explores taking corrective action to deliver the predicted performance.	Grads Project
AutoPilot	Autopilot, developed by the Pablo Research Group, provides a library of C++ classes implementing software sensors and actuators that facilitate data distribution and remote control among processes running on the Grid.   Sensors and actuators can be inserted in application programs, or used in standalone programs that monitor system and network resources,  to allow easy sharing of the collected information.  The sensors and actuators are registered with a directory service, called the AutopilotManager, that can be searched by interested clients.  Once a client has located a sensor or actuator of interest, a direct connection is established to exchange data via Nexus RSRs.  The classes support data buffering and both push and pull modes of data exchange.	UIUC	Autopilot
MDS	Globus MDS: Information and Monitoring Service .	Globus	MDS
NWS	Network Weather Service .	Rich Wolski, UCSB	NWS
R-GMA	GMA implementation in Java using a Relational Data Model.	EU DataGrid WP3	R-GMA
CODE	GMA implementation.	NASA Ames (Warren Smith)	CODE

For additions to this page, email Dan Gunter