Federated Collaborative Networks

PEER

Project description

PEER is a federated object-oriented information management system, designed and implemented at the University of Amsterdam, developed to support the sharing and exchange of information among cooperating autonomous and heterogeneous nodes. The PEER federated architecture consists of a network of tightly/loosely interrelated nodes. Both the information and the control is distributed within the network. PEER nodes have no single global schema defined on the shared information, unlike many other distributed object-oriented database systems to support the entire network of database systems. The interdependencies between two nodes’ information is established through the schemas defined on their data; thus there is no need to store the data redundantly. There is no central (global) control within the network.

For each PEER node there is one schema that specifies the type structure of all objects stored locally. This schema is called the local schema (LOC). Derived from the local schema are one, or more, export schemas (EXPs) that each define a particular view on the local objects. Usually, an export schema contains only a part of the concepts (types and mappings) defined in the local schema. An export schema can be imported by other PEERs; that will be called import schema (IMP). Each PEER node has one integrated schema (INT), which is derived from the local schema and the various imported schemas. The integrated schema provides a single uniform type structure defining all the objects that are accessible by this PEER, both locally and remote. The Schema Definition/Derivation Language (SDDL) of PEER offers different `schema definition environments’ to define/derive these four different kinds of schemas. This language offers a set of ‘type and map derivation primitives’ to support the formulations of the detailed derivation specifications in the derived schemas. The schema derivation mechanism of PEER supports both the end-user and the application program developer. For the end-user, the mechanism is a tool to enforce at a high level his own view on the information that he/she manipulates. For instance, the mechanism enables the user at a node to preserve his own view on the information, as defined in his local schema, and restructure the imported information in order to match the existing structures, and then integrate the two schemas. For the application program developer also, it supports the isolation of his data structure design choices in his/her program from their distributed database representation.

The PEER model is a binary-based object-oriented database model. Any identifiable piece of information is uniformly represented as objects. It represents atomic, composite, and type objects. Atomic objects are strings of characters, numbers, booleans, text, etc. usually referred to as `values’ in other systems. Composite objects are non-atomic entities and concepts of the application environments and can be decomposed into further objects. Mapping objects, usually referred to as `attribute’, are a special kind of composite object. Mappings can be single-valued or multi-valued and represent both the descriptive characteristics of an object as well as its association with other objects. A type object is a structural specification of a group of atomic or composite objects. It denotes a collection of database objects called its members (instances). The PEER supports multiple inheritance; the subtype/supertype relationships defined among types form a Directed Acyclic Graph (DAG). The information retrieval/manipulation language designed for PEER supports both the local access and modification of nodes’ information, and the remote access and sharing of information among nodes. The retrieval of information is based on queries on binary relations among objects.

In PEER, a collection of inter-related objects can be defined and viewed as a single object-cluster. An example object-cluster is a patient-record that contains the patient’s name, and several cases of illnesses; where every case of illness is defined by the doctor who treated it, and the examination-record; where the doctor is defined by the name; and where an examination-record is described by the tests-results and the status report. PEER’s approach to object-clusters is to define it as an instance of a cluster-template which regulates its definition uniformly to the type/member definition, and it further supports the definition and manipulation of object-clusters that are distributed over several nodes in the network. To define an object-cluster as a single entity, its boundary has to be specified. An object-cluster’s boundary is defined by a `template’, that specifies its collection of relevant objects. A template for object-clusters is defined by a type (that is the `root’ of the template) and a `derivation expression’, as an ordered set of mappings defined among the types, starting at the root and continued on to the next type using a mapping on the path. The boundary of an object-cluster defined by a template is independent of the aggregations defined on its types. So, it is very well possible to define several different templates for patient-records. One template can only contain the test-results of the patient-record while another may only contain the doctors’ information. An object-cluster can be defined as an instance of this template by a tuple, with the object-id of its root object as the first element, and the name of its template as the second element. Both interactive users and the application programs need support for retrieving and updating large object-clusters as single entities. Application developers need to interface various development environments (e.g. programming languages) to the object management system. The impedance mismatch between the information base environment and the application development environment is most severely felt when object-clusters are retrieved. Here, we resolve this problem by retrieving object-clusters as a single entity in a linear format. Linear representations of object-clusters are needed for transporting them as messages between different PEER objectbases, between the PEER objectbases and the application programs, and for conversion of object-clusters into files.

The functionality of the PEER federated system is supported by the specific architecture of each PEER node and its PEER-kernel, and by the existence of a community dictionary in the network. The PEER-kernel is itself a predefined schema that acts as a unifying superstructure to primarily support supports the representation of multiple schemas, the information on the node’s involvement in the cooperation network and provides an environment to model the schema integration/derivation relationships. The community dictionary is the `source of information’ within the network. Its function is to provide up-to-date information on all nodes in the network. It contains the network addresses of its active nodes and their current state. For every node, it also stores its export schemas and the specific access rights and schema modification rights that the node supports. The dictionary can also be used as the general store for other static information that concerns the entire community such as objects’ name-tables.

An implementation of PEER federated system is developed in the C language that runs on a network of SUN and IBM RS/6000 workstations. The main applications where the PEER system has been or is being applied are within industrial automation, CIM and biomed. For higher level access to the data and meta-data, there are two interface tools developed, these are the Schema Manipulation Tool (SMT) and the Database Browsing Tool (DBT). The Schema Manipulation Tool supports the definition, modification and checking of schema definitions of an node. The Database Browsing Tool supports a high level browsing interface through the data and meta-data of an node.

Contact person

Hamideh Afsarmanesh

The PEER development group:

– Hamideh Afsarmanesh
– Ammar Benabdelkader
– Cesar Garita
– Michiel Wiedijk
– Frank Tuijnman
– Martin Bergman
– Peter Trenning
– Arnoud Visser
– Pedro Pinto
– Olindo Iglesias

PEER related publications:

Bibliography of PEER related papers

1997

  • Camarinha-Matos, H. Afsarmanesh, C. Garita, and C. Lima. Towards an Architecture for Virtual Enterprises. In Proceedings of the II World Congress on Intelligent Manufacturing Processes & Systems IMPS’97, Budapest, Hungary, June 1997.
  • H. Afsarmanesh, L.M. Camarinha-Matos, and F.J. Martinelli. Federated Knowledge Integration and Machile Learning in Water Distribution Networks. In Proceedings of the IFIP/IEEE/OE International Conference on Integrated and Sustainable Industrial Production ISIP’97, Lisbon, Portugal, May 1997.

1996

  • M. Wiedijk, H. Afsarmanesh, and L.O. Hertzberger. Co-working and Management of Federated Information-Clusters. In Proceedings of the 7th IEEE International Conference on “Database and Expert Systems Applications DEXA’96”, Zurich, Switzerland, Lecture Notes in Computer Science (LNCS 1134), pages 446-455. Springer Verlag, September 1996.
  • M. Wiedijk, H. Afsarmanesh, I. Mazón, F. Rojas, and A. Pérez. Information Management for Intelligent Supervision of Hybrid Process Control. In Proceedings of the IEEE/ECLA/IFIP 2nd International Conference for Balanced Automation Systems BASYS’96, Lisbon, Portugal, pages 115-125, June 1996.
  • I. Mazón, H. Afsarmanesh, M. Wiedijk, F. Rojas, A. Pérez, and L.O. Hertzberger. Control Information Management in a Balanced Automated System. Journal of Studies in Informatics and Control, 5(2):119-130. IC Publications, Romania, June 1996.
  • H. Afsarmanesh, M. Wiedijk, L.O. Hertzberger, F.J. Negreiros Gomes, A. Provedel, R.C. Martins, and E.O.T. Salles. Cooperation of CIM Expert Systems supported by PEER. Journal of Studies in Informatics and Control, 5(2):157-169. IC Publications, Romania, June 1996.

1995

  • L.M. Camarinha-Matos, H. Afsarmanesh, and J.R. Reinaldo Silva. Multiagent Systems in FMS/FAS. In J. Paiuk and J. Weisz, editors, Preprints of the 4th IFAC Symposium on Low Cost Automation, Buenos Aires, Argentina, pages 180-187. IFAC/AADECA, September 1995.
  • G. Loria, I. Mazón, F. Rojas, H. Afsarmanesh, M. Wiedijk, and L.O. Hertzberger. Supporting the Information Management of Control Activities in a CIM Shop Floor Environment. In Proceedings of the IEEE/ECLA/IFIP International Conference on Architecture and Design Methods for Balanced Automation Systems BASYS’95, Vitoria, Brazil, pages 253-264, July 1995.
  • H. Afsarmanesh, M. Wiedijk, L.O. Hertzberger, F.J. Negreiros Gomes, A. Provedel, R.C. Martins, and E.O.T. Salles. A Federated Cooperation Architecture for Expert Systems Involved in Layout Optimization. In Proceedings of the IEEE/ECLA/IFIP International Conference on Architecture and Design Methods for Balanced Automation Systems BASYS’95, Vitoria, Brazil, pages 287-300, July 1995.
  • J. Reinaldo Silva, H. Afsarmanesh, D.D. Cowan, and C.J P. Lucena. An Object-Oriented Approach to the Design of Flexible Manufacturing Systems. In Proceedings of the IEEE/ECLA/IFIP International Conference on Architecture and Design Methods for Balanced Automation Systems BASYS’95, Vitoria, Brazil, pages 91-106, July 1995.

1994

  • H. Afsarmanesh, M. Wiedijk, N.P. Moreira, and A.C. Ferreira. Design of a Distributed Database for a Concurrent Engineering Environment. Journal of the Brazilian Society of Mechanical Sciences, 16(3):298-310, December 1994.
  • H. Afsarmanesh, M. Wiedijk, F.J. Negreiros, R.C. Lopes, and R.C. Martins. Integration of an Optimization Expert System within a CIM Distributed Database System. Journal of the Brazilian Society of Mechanical Sciences, 16(3):318-327, December 1994.
  • M. Wiedijk and H. Afsarmanesh. Information Modeling in a Robot Assembly Environment. In Proceedings of 27th ISATA International dedicated conference on Lean/Agile Manufacturing in the Automotive Industries, Aachen, Germany, pages 181-188, November 1994.
  • H. Afsarmanesh, M. Wiedijk, A.C. Ferreira, and N.P. Moreira. Distributed Database Support for a Concurrent Engineering Environment. Journal of Studies in Informatics and Control, Vol. 3, Nos. 2-3, pages 135-143. IC Publications, Romania, September 1994.
  • H. Afsarmanesh, M. Wiedijk, and L.O. Hertzberger. Flexible and Dynamic Integration of Multiple Information Bases. In Proceedings of the 5th IEEE International Conference on “Database and Expert Systems Applications DEXA’94”, Athens, Greece, Lecture Notes in Computer Science (LNCS) 856, pages 744-753. Springer Verlag, September 1994.
  • H. Afsarmanesh, M. Wiedijk, F. Tuijnman, M. Bergman, and P. Trenning. The PEER Information Management Language User Manual. Technical Report CS-94-14, Dept. of Computer Systems, University of Amsterdam, September 1994.
  • M. Wiedijk and H. Afsarmanesh. The PEER User Interface Tools Manual. Technical Report CS-94-15, Dept. of Computer Systems, University of Amsterdam, September 1994.
  • H. Afsarmanesh, F. Tuijnman, M. Wiedijk, and L.O. Hertzberger.The Implementation Architecture of PEER Federated Object Management System. Technical report, Dept. of Computer Systems, University of Amsterdam, September 1994.
  • L.M. Camarinha-Matos and H. Afsarmanesh. Federated Information Systems in Manufacturing. In Proceedings of the IEEE European Robotics and Intelligent Systems Conference EURISCON’94 Vol. 3 (Stream C), Malaga, Spain, pages 1598-1613, August 1994.
  • H. Afsarmanesh, M. Wiedijk, N.P. Moreira, and A.C. Ferreira. Design of a Distributed Database for a Concurrent Engineering Environment. In Proceedings of the ECLA.CIM workshop, Florianopolis, Brazil, pages 35-43, June 1994.
  • H. Afsarmanesh, M. Wiedijk, F.J. Negreiros, R.C. Lopes, and R.C. Martins. Integration of Optimization Expert Systems with a CIM Distributed Database System. In Proceedings of the ECLA.CIM workshop, Florianopolis, Brazil, page 2, June 1994.

1993

  • F. Tuijnman and H. Afsarmanesh.Management of shared data in federated cooperative PEER environment.International Journal of Intelligent and Cooperative Information Systems (IJICIS), 2(4):451-473, December 1993.
  • H. Afsarmanesh, M. Wiedijk, A.C. Ferreira, and N.P. Moreira. An Approach to the Design of a Distributed CIM Database for a Brazilian Aerospace Industry. In Proceedings of the ECLA.CIM Workshop, Lisboa, Portugal, pages 81-88, November 1993.
  • H. Afsarmanesh, F. Tuijnman, M. Wiedijk, and L.O. Hertzberger. Distributed Schema Management in a Cooperation Network of Autonomous Agents. In Proceedings of the 4th IEEE International Conference on “Database and Expert Systems Applications DEXA’93”, Lecture Notes in Computer Science (LNCS) 720, pages 565-576. Springer Verlag, September 1993.
  • F. Tuijnman and H. Afsarmanesh. Sharing Complex Objects in a Distributed PEER Environment. In 13th Int. Conf. on Distributed Computing Systems, pages 186-193. IEEE, May 1993.
  • F. Tuijnman and H. Afsarmanesh. Distributed Objects in a Federation of Autonomous Cooperating Agents. In Int. Conf. on Intelligent and Cooperative Information Systems, pages 256-265. IEEE/AAAI, May 1993.

before 1993

  • H. Afsarmanesh and F. Tuijnman. Complex Object Management in a Distributed CIM Environment. Technical Report CS-92-14, Department of Computer Systems, University of Amsterdam, October 1992.
  • F. Tuijnman, H. Afsarmanesh, and G. Wiedijk. Specification of the Distributed Information Management. Archon Esprit Project P-2256 Technical Note Archon/UvA/TN11/7-92, July 1992.
  • H. Afsarmanesh, F. Tuijnman, P. Trenning, and G. Wiedijk. The 3DIS/ISL User Manual. Technical Report CS-92-08, Dept. of Computer Systems, University of Amsterdam, June 1992.
  • H. Afsarmanesh and F. Tuijnman. ISA: An Architecture to support Sharing and Exchange of Information among Archon Agents. Archon Esprit Project P-2256 Technical Report TR23, 1991.
  • H. Afsarmanesh, D. Knapp, D. McLeod, and A. Parker. An Extensible Object-Oriented Approach to Databases for VLSI/CAD. In Proceedings 11th International Conference on Very Large Data Bases, pages 13-24, 1985. reprinted in “Readings in Object-Oriented Database Systems” (S. Zdonik and D. Maier, editors), Morgan Kaufmann Publishers, 1990.
  • H. Afsarmanesh and D. McLeod. The 3DIS: An Extensible Object-Oriented Information Management Environment. ACM Transaction on Information Systems, 7:339-377, October 1989.

Articles describing PEER architecture:

  • H. Afsarmanesh, F. Tuijnman, M. Wiedijk, and L.O. Hertzberger. Distributed Schema Management in a Cooperation Network of Autonomous Agents. In Proceedings of the 4th IEEE International Conference on “Database and Expert Systems Applications DEXA’93”, Lecture Notes in Computer Science (LNCS) 720, pages 565-576. Springer Verlag, September 1993. (Postscript)
  • F. Tuijnman and H. Afsarmanesh. Management of shared data in federated cooperative PEER environment. International Journal of Intelligent and Cooperative Information Systems (IJICIS), 2(4):451-473, December 1993. (Postscript)
  • H. Afsarmanesh, M. Wiedijk, and L.O. Hertzberger. Flexible and Dynamic Integration of Multiple Information Bases. In Proceedings of the 5th IEEE International Conference on “Database and Expert Systems Applications DEXA’94”, Athens, Greece, Lecture Notes in Computer Science (LNCS) 856, pages 744-753. Springer Verlag, September 1994.(Postscript)
  • Wiedijk, M.; Afsarmanesh, H.; Hertzberger, L.O. (1996) – Co-working and Management of Federated Information-Clusters. 7th Int. Conf. on Database and Expert Systems (DEXA’96), Lecture Notes in Computer Science 1134, pp 446-455. Springer Verlag, Sept. 96.

The Federated Object Management System PEER [Restricted] (summary_1.ps) (summary_2.ps)

The Implementation Architecture of PEER Federated Object Management System [Restricted] (Postscript)