Version: | 1.1 |
---|---|
Author: | Ivan R. Judson <judson@mcs.anl.gov> |
Status: | Draft |
This AGEP describes a proposed factoring of the initialization code for the AGTk. The goal is to unify and clarify the process of initialization for anything that is developed using the AGTk.
In order to unify and clarify the initialization process there are a few activities that need to be done by every piece of Access Grid related software. These activities are enumerated, described, and a factoring is presented that encapsulates them into a well defined set of interfaces.
The basic activities that are done during initialization are:
- Globus Initialization
- Bring Linux Initialization up to par with Windows
- Looking for system wide configs and either referencing or copying them
The types of things that can be built using the AGTk (and thus it's initialization functionality) includes:
Class | Type | Run As |
---|---|---|
Venue Clients | Application | User Id |
Shared Application Clients | Application | User Id |
Node Services | Service | Service Id / User Id |
Service Managers | Service | Service Id / User Id |
Services | Service | Service Id / User Id |
Venue Server | Service | Service Id / User Id |
Network Services | Service | Service Id / User Id |
Application Services | Service | Service Id / User Id |
None
This work should make the toolkit more secure, by enforcing all software built using the AGTk conforms to the security environment created by the AGTk.
This internal refactoring should not affect interoperability.
FILL IN IMPLEMENTATION
None
This document is Copyright 2003, The University of Chicago/Argonne National Laboratory.