Lack of long-term maintainability e.g. no platform independence
The EGOS Initiative
Harmonise the technologies/processes used by all infrastructure products
Develop solutions which are generic enough to serve the same need for different user communities
Promote re-use of the same implementation across infrastructure domains at various levels:
Common libraries (development)
Common components (integration)
Common services (run-time)
Objective minimise the development and long term sustaining efforts of the infrastructure code base
The initial ambition
Move from a fully ‘vertical’ to a largely ‘horizontal’ implementation model:
Layered implementation relying on extensive re-use of the same implementations across all domains
Identify, design and implement all potentially common elements of the future infrastructure
Develop new systems according to the EGOS ‘re-use’ paradigm
Progressively re-engineer the existing systems to replace custom functionality with generic solutions
Difficulties experienced
The larger systems (e.g. S2K) required significant re-engineering activities to rationalise their architecture
Lack of resources to accommodate the ambitious EGOS objectives in parallel with the need to evolve and expand the scope of the infrastructure
General/abstract solutions risk to be over-engineered for small applications e.g. the ones outside the ‘execution’ environment
The selected technology for the ‘native’ EGOS components (CORBA Component Model) is not well supported/widely used
The Current Implementation Model Platform Baseline Support Services Generic Applications & Standards Implementation Drivers Adapters External Interfaces User Interfaces Space Systems Monitoring & Control Operational Data Dissemin. Analysis & Reporting Operations Preparation Space Systems Simulators Operations Planning & Automation
Description of the Implementation Model
Common platform baselines (OS + 3rd Party Products) across domains
Run-time support services and generic implementations delivering functionality which is not domain specific
‘ Infrastructure functional domains’ supporting generic features which are re-usable across the traditional boundaries of user communities e.g. for controlling or simulating all systems of the ground and space segment
Common service provision layer ensuring ‘integrability’ and ‘inter-operability’
Common user interface platform
The Deployment Model Operations Preparation Databases Management Procedures Definition Operational Systems Tailoring Files Files Databases Multi-purpose Data Archive Post-Operations Analysis + Reports/Alerts Data Dissemination Office PCs Internet Remote Users Planning Data Interfaces Mission Data Automation M&C Kernel Service Provision Operators Interfaces Operational consoles Office PCs Support Services
The support services and generic applications Databases management (DABYS) Engineering Data Archive (DARC) Data Dissemination (EDDS) Data replication (SFT) Data Management Virtual Filestore (FARC) File routing and delivery (GFTS) Files Management Build and Installation (GABIS) Configuration Access (EGOS Core) Run-time directory (EGOS Core) Actions execution (EGOS Core) Events logging (EGOS Core) Services management (SMF) Real-time monitoring (OSMOSYS) System Management Identity and Session Management (EGOS Core) Secure Access Control (EGOS Core) Users Management
Harmonisation within functional domains: The simulators example Simulation Infrastructure Ground Station Monitoring and Control Validation OBSW Maintenance Flight Dynamics Ground Systems Validation Spacecraft Control Today Yesterday Tomorrow
M&C Domain Harmonisation: The ECS example MCS MCS MCS MCS STC STC STC STC ECS EPS/ESS Distribute products ( GFTS ) Distribute GS schedule ( FIDES ) Inject Station Parameters ( SMF S2K ) Receive Station Parameters and log messages ( GSC CORBA / SMF ) ECS History Repository (DARC) Store monitoring data ( DARC API ) ECS Product Repository (FARC) Store plan / schedule ( FARC API ) ECS Monitoring (S2K + MATIS) Inject Station Data ( SMF S2K )
M&C Domain Harmonisation: The GSMC example Specific Mission Control System SCOS-2000 Spacecraft M&C GSMC Ground stations M&C Common M&C Platform Common M&C Platform
GSMC based on a Common M&C Platform Ground Station Centralised Monitoring Script Execution MATIS (+) Control Center Archive User Desktop based GUI MON Model per Client Live Retrieval File archive Acquisition node
The idea of Reference Architectures
A Reference Architecture provides a framework for developing and integrating generic and specific implementations
It consists of the definition of components, interfaces and lifecycle of a complete system
Only some of the components are delivered as generic implementations, the others are implemented specifically for a given system
It is a useful platform in all cases where fully generic solutions are not viable for any reason
It encourages the commonality across systems/missions through a progressive generalisation process (rather than the classical specialisation process of generic infrastructure)
Reference Architectures: The EGOS User Desktop
Reference Architectures: EUD Contributors
Reference Architectures: the Post-operations domain Virtual Data Store Integration-API On-line Archives (OPSLAN) Off-line Archives (Pre-OPSLAN / Relay LAN) PARC FARC DARC …… . Firewall Firewall Web-Based User Interfaces EGOS views MUST Clients EDDS WS-API HTTP Internet SFT EDDS
A common support and development environment Change Management: Synergy Change Requirements Management: DOORS Test Management: Mercury Quality Centre Sharepoint EGOS Portals User Management Configuration Management: Synergy CM BIRF Project Management Services License Server Software Development Services Central Services
Achievements Advanced in the simulators domain, started in the pre&post-operations domain, awaiting programmatic decisions in the M&C domain (ECS + GSMC) Harmonisation/rationalisation within functional domains Implementation of support services and generic applications nearly completed. Deployment in operational systems is behind schedule. Common implementation across domains Good progress. Systems which are expected to be maintained in the long-term have or are being re-engineered Legacy systems re-engineering Good progress. Converging on Linux 64 bits platform + common 3 rd party products Technology/processes harmonisation
Next (concrete) steps
Complete consolidation of platform baselines (SLES 11 64 bits)
Roll-out generic applications developed so far
Extend the generic applications (e.g. FARC, DARC, SMF) to support the needs of new systems (e.g. ECS and EDDS)
Define the Reference Architecture for the ‘new’ functional domains (i.e. Mission Planning, Pre & Post Operations) and provide a framework implementation
Re-implement the user interfaces of the main systems on the basis of the EGOS User Desktop
Deploy the EGOS Core Components as a multi-system service provider
Develop a common Monitoring and Control platform for spacecraft and ground station control
Conclusions
The importance and the scope of infrastructure ground data systems has constantly increased in the last years
The EGOS initiative has promoted a ‘cross-communities’ culture which is now starting to deliver visible effects
Harmonisation within functional domains is ongoing and should be further supported/promoted
It is important to have strategic long-term visions but it is equally important to map them to concrete steps delivering visible outputs in the short term
Cultural changes take a long time to become visible…
For many years, ESA/ESOC's OPS-GI team have promote more
For many years, ESA/ESOC's OPS-GI team have promoted a harmonisation process for ground data systems infrastructure under the name of 'EGOS' (ESA Ground Operations Software). The basic idea behind EGOS is to develop and deploy solutions that are not only generic within a specific domain but that can also cover the same need across all relevant domains of ground data systems. What is the current status and plans of this initiative? Are the initial objectives still realistic? What is the target architecture for data systems infrastructure? How will this impact the evolution of the existing infrastructure products? less
0 comments
Post a comment