Ontology Mapping for the Interoperability Problem in Network Management
Wong, Ray, Parameswaran, Strassner
network snmp management ontology integration
@article{wong:jsac-2005,
title={Ontology Mapping for the Interoperability Problem
in Network Management},
author={Wong, A.K.Y. and Ray, P. and
Parameswaran, N. and Strassner, J.},
journal={{IEEE} Journal on Selected Areas in Communications},
volume={23},
number={10},
pages={2058--2068},
issn={0733-8716},
year={2005},
publisher={{IEEE}}
}
[ PDF ]
Network management abstractions exist, but don't have formal semantics
- Ability to relate to other standards/domains is important
Interoperability means one or all of:
- Data/knowledge exchange
- Coordinated behavior
- Cooperative problem solving
Interoperability on four levels of increasing complexity and difficulty:
- Physical
- Data
- Specification
- Semantic
Interoperability may occur at different places:
- Applications
- Platforms
- Management Information Base (MIB)
Two programs for a similar configuration task:
Currently, "network administrator must translate a high-level specification of the desired functions (e.g., traffic conditioning for gold, silver, and bronze services) into different sets of commands for each vendor"
Automated mapping is required for effective real world use
Concepts to be modeled in network management:
- Tangible network objects (router interfaces)
- Intangibles (BGP parameters)
- Management operations (router configuration commands, security monitoring)
Router configuration can be expressed as state transition of network objects
Ontology mapping process
- Ontology acquisition
- Ontology formalization
- Concept similarity estimation
- Ontology mapping
- Results presentation
Divides similarity measurement into three tasks:
- "Translation of the ontology from the initial otology formalization language into the semantic ontology description language"
- "Generation of the semantic content of the concepts"
- "Calculation of the semantic similarity estimates between the concepts"
Defines a similarity metric for FOL statements
Evaluation doesn't really compare to other similarity metrics
- Constructs their own ontologies for evaluation...
Of note:
- SNMP SMI
- TMF SID
- OSI GDMO
- DMTFCIM
- ARM API
- ITU-T TMN (Telecommunications Management Network)