Towards Semantic Interoperability between C2 Systems Following the Principles of Distributed Simulation

1. Increased focus on multi-functional and multi-national operations brings new requirements to military command and control, in addition to other capability requirements. Parallel but separate from this development, interoperability has been of major concern within Modeling and Simulation (M&S) for years, especially in connection with standards for distributed simulations, e.g. High Level Architecture (HLA). Both areas share a need to create configurations of systems where elements of information exchanged are interpreted similarly among all participating parties, preserving the intended meaning (i.e. semantic interoperability). An effort to address this need is currently under development within NATO IST-094, Semantic Interoperability Framework (SIF), which includes tool and methodology support for harmonising data/information models on a semantic level, as well as mediators to translate between heterogeneous abstractions. The framework builds on a knowledge-based approach utilizing emerging semantic technologies, such as ontologies. In this paper we investigate how, and to what extent, concepts, solutions and experiences from the distributed simulation community can help fulfill the requirements of the SIF. Based on this, a common process is presented which is aimed at governing the development and execution of system configurations to meet expressed business requirements. 1.1. Background Introduction Coordinated efforts, collaborations and interdependencies have increased the need for information exchange between heterogeneous systems that are owned and designed by different organizations. Semantic heterogeneity is a particularly challenging form of heterogeneity which occurs when there is disagreement regarding the meaning, interpretation and intent of information or when information is described in different ways in two different systems. Semantic heterogeneity causes new and unexplored problems in decentralized information systems. Multiple interpretations of data by different users in different contexts must be handled. In current information systems, much of the semantics of a system resides in the application code, or in the assumptions which the application makes about the data rather than in a conceptual schema. Such a situation can be accepted in a centralized system in which the applications use a shared set of assumptions, but in a decentralized environment it gives rise to severe problems. It is therefore important to develop methods, tools, and techniques for achieving semantic interoperability, i.e., cooperation among semantically heterogeneous systems [1]. 1.2. Problem The ongoing globalization poses new challenges for military operations. In particular, it has become much more common to carry out activities together with other nations' civil and military organizations, i.e. to interoperate in multinational and multifunctional contexts. In order to cooperate efficiently, it is necessary for different organizations to exchange information between their command and control (C2), management and information systems (IS), i.e., to be interoperable. It is therefore essential to develop future IS that can adapt to different types of situations in which the information exchange needs are not known in advance. A prerequisite for an improved interoperability between IS of different organizations is to create standards, methods and tools which can align different terminology, and facilitate translation of data between heterogeneous systems. Successful operations involving organizations which are not trained together requires at least one very important function reliable communication of critical information like threats and risks. This, in turn, requires that any two co-operating parties are interoperable on a semantic level. Problems related to the use of different data formats, or pure connectivity issues, all too often distract from what is really the main problem to understand and interpret information in a consistent manner. The importance of this aspect cannot be emphasized enough, especially so when information exchange is taking place between multiple domains. The aim of semantic interoperability is to achieve a common understanding of one or more domains before the exchange of operational information between systems takes place. Since current development processes do not take into account the preservation of the intended meaning of terms in the exchange of information, any attempt to integrate heterogeneous systems for multi-national operations without taking semantic interoperability into consideration will be insufficient. Thus, it is of a great importance that command and control systems are developed with flexibility in mind, in order to be able to adapt to different situations in which the need to exchange information between systems exists. Within NATO, semantic interoperability (SI) has been consequently identified as a core capability for future command and control systems. Semantic interoperability may contribute to various capabilities of the armed forces, but is above all expected to increase the safety of international operations. Despite this, semantic interoperability has not yet been identified as a requirement for information exchange between different systems. SI is formally defined in Section 2. 1.3. Our contribution Interoperability problems have been a major concern within the Modeling and Simulation community for years, especially in connection with standards for distributed simulations. As in the C2 domain, there is a need to create configurations of systems where elements of information exchanged are interpreted similarly among all participating parties, preserving the intended meaning (i.e. semantic interoperability). As mentioned, an effort to address this need in the military world is currently ongoing within NATO, with the proposal of a Semantic Interoperability Framework (SIF). Therefore, it is of great interest to study and gain experiences from the M&S community which has struggled with similar interoperability issues for many years and apply that knowledge to the SIF which is still under development. The distributed simulation community has been successful in creating conditions for integration of simulation models, but has historically been most concerned with interoperability at the syntactic level. However, two very important results from this domain are of interest when developing SIF. The most important concerns the achievement of a standardized process for development and execution of distributed simulations. The other one concerns component-based development which relies on simulation components (models) being interoperable in different configurations without much need for manual work. In this paper, theories and best practices that have been accumulated by the distributed simulation community are adapted and applied to SIF in order to develop a robust framework for semantic interoperability of C2 systems. In particular, we investigate how, and to what extent, concepts, solutions and experiences from the distributed simulation community can help in fulfilling the requirements of SIF. By doing so, we aim at conceptualizing a common process for governing the development, execution and analysis of heterogeneous systems in a C2 context, meeting semantic interoperability requirements. 1.4. Paper layout The rest of the paper is organized as follows. Section 2 introduces the concept of semantic interoperability, as well as other levels of interoperability both below and above the semantic level. The concept of ontology is defined and its role in a knowledge based solution for semantic interoperability is described. Section 3 introduces the Semantic Interoperability Framework – SIF, which proposes an augmented ability of information systems to facilitate exchange of data and share information and knowledge. After a brief description of SIF, its main components and functionality are described. Section 4 gives a brief introduction to distributed simulation standards, the High Level Architecture (HLA) and the Federation Development and Execution Process (FEDEP). In Section 5 a general comparison of distributed simulation efforts and SIF’s requirements is provided. Based on that, and a survey of other, collaborationand interoperability-related management processes, a similar process for SIF is defined. The process is named Semantic Interoperability and Development Execution Process (SIDEP). In this study, SIDEP is explored to a high-level meta-model and the major activities are defined. Section 6 concludes the paper by summarizing our current research results and pointing out intended future work. 2. Knowledge-based solution In this section we recall the basics of knowledge-based solutions for semantic interoperability. 2.1. Semantic Interoperability Wikipedia defines Semantic Interoperability as the ability of two or more computer systems to exchange information and have the meaning of that information accurately and automatically interpreted by the receiving system. NATO's primary research group in this field, NATO RTO IST-075, has slightly modified Wikipedia’s definition and defines Semantic Interoperability as the ability of two or more computerized systems to exchange information for a specific task and have the meaning of that information accurately and automatically interpreted by the receiving system, in light of the task to be performed [2]. Hence, two actors that are semantically interoperable can not only exchange information, but can also interpret and understand the intended meaning of the information in a common way. This is a key issue in the interaction between groups that do not share common frames of reference acquired through a common culture or through education. Support for semantic interoperability is therefore a prerequisite for the ability to participate in international operations with allied forces. Interoperability is more than only the technical compatibility of systems. In a Network Centric Warfare scenario, the C2IS of all engaged elements must be connected (physical interoperability), exchange data in such a way that automatic processing is possible (syntactic interoperability), exchange information and guarantee identical interpretation (semantic interoperability), cooperate and realize situational awareness (pragmatic interoperability) that assures the coherent cooperation of all participating actors (social/cultural interoperability). 2.2. Utilizing ontology for Semantic Interoperability Knowledge-based solutions for semantic interoperability often exploit the ontology notion. Within the knowledge engineering community, ontology is defined as an explicit, formal specification of a shared conceptualization [3]. Here, conceptualization refers to an abstract, concept-based model of some phenomenon in the world. Explicit means that the type of concepts used, as well as the constraints on their use, are explicitly defined. Formal refers to the fact that the ontology should be machine-readable. Shared reflects that an ontology captures consensual knowledge, that is, the knowledge accepted by a group. Every term used in natural languages has several meanings. In an ontology we constrain the semantic interpretation of these terms, and provide formal definitions. This is called ontological commitment and means mapping between ontology terms and their intended meanings. The major task here is to determine precisely what meaning the term has. Ontologies have been proposed for the following uses [4]: • sharing common understanding of the structure of information among people or software agents, • enabling reuse of domain knowledge, • making domain assumptions explicit, • separating domain knowledge from operation knowledge, and • analyzing domain knowledge. More recently, ontologies have become recognized as an emerging mechanism for dealing with semantic interoperability of Information Systems. This is entirely aligned with the lately recognized fact that semantic understanding and interoperability is a key challenge for organizations and their systems to successfully and competitively provide their services. By specifying the conceptualization in terms of an “agreement” on meaning between the parties involved, the ontology becomes a reification of an agreement on knowledge. 2.3. A solution for Semantic Interoperability The traditional means of exchanging information between systems do not guarantee that the intended meaning of information (the semantics) is preserved. To ensure that meaning is preserved, we need shared terminologies (ontologies); every message between communicating actors may then include references to one or several ontologies according to which the message should be interpreted. Common representation of semantics through ontologies represents one important step towards information interoperability. However, in addition to the use of ontologies and related tools, a consensus on a common process is needed, i.e. on how they are to be used in the lifecycle of a system interoperability task. One way to achieve semantic interoperability between two systems is to align the ontologies of those systems. Ontology alignment is the result of an ontology matching process which is the task of determining correspondences between the concepts of different ontologies. Ontology matching and alignment are required when two heterogeneous systems want to harmonise their ontologies in order to achieve semantic interoperability. This process of harmonising two different ontologies is known as ontology reconciliation [5]. FOI (Swedish Defense Research Agency) has as of 2007 worked to clarify the concept of semantic interoperability, to build skills in this area, and to propose solutions. In cooperation with NATO's primary research group in this field, NATO IST-075, a general logical framework in the shape of an architecture for semantic interoperability has been developed, called Semantic Interoperability Framework (SIF). The framework will be explained in next section. 3. SIF Semantic Interoperability Framework In this section we describe the Semantic Interoperability Framework (SIF) proposed in the report of NATO task group IST-075 [2]. 3.1. Background Semantic Interoperability (SI) is difficult to measure, and the challenge of achieving SI between independent and heterogeneous systems is far too complex for any “one-size-fits-all” universal solution. Nevertheless, in order to achieve a common view and describe this challenge, the NATO group IST-075 has conceived a framework, called SIF Semantic Interoperability Framework, intending to provide a generic approach to SI. This section briefly introduces SIF, gives an overview of SIF, and finally outlines the requirements for it to function efficiently. IST-075 was an RTG (Research and Technology Group) coordinated by the IST (Information Systems Technology), and included members from several countries that cooperated, for the period 2007-2009 under the umbrella of NATO RTO (Research and Technology Organization). The group worked on the problem area "semantic interoperability" with a focus on ontologies and created SIF as one of its results. The Swedish Defense Research Agency FOI received a mandate from the Swedish Armed Forces to join, follow and contribute to this work during that period. A continuation group IST-094, which FOI also joined and supports, proceeds with this activity for the period 2010-2012. 3.2. Overview of SIF In order to ensure semantic interoperability of several systems, an architecture is needed which includes a party-wise set of common ontologies between communicating parties, which the involved systems can understand and use. Such is always implied by actors who exchange messages (otherwise communication is impossible), but in this architecture it is made explicit. This allows each message between communicating parties to be provided with references to one or more of the ontologies according to which the message should be interpreted. SIF is a high level view of such architecture that supports semantic interoperability among heterogeneous information systems. In terms of features, SIF is a middleware that performs interoperability in a communication medium and not as part of the communicating systems. SIF applies means of knowledge-based systems, using ontologies, for mediation purposes. SIF can be described from various perspectives from a functional point of view one could say that SIF has a preparatory phase and an implementation phase. During the preparatory phase, all necessary information about the participants in the communication is collected or created in the form of ontologies. During the implementation phase a number of different ontology methods and mapping tools are applied on those ontologies. The end result is a transformation of the message structure from one information system A to another information system B with preserved semantics. 3.3. Assumptions and Conditions The application of SIF assumes that the lower levels of interoperability have already been achieved between the concerned systems. This means that the systems are connected (physical interoperability is established) and that they can exchange data in such a way that automatic data processing is possible (syntactic interoperability is also established). It also assumes that semantic descriptions of systems can be obtained in some way. These descriptions can more or less automatically be (partly) derived from systems, but in order to achieve the necessary quality of the descriptions the process normally requires human intervention. It is important to note that the starting point for SIF is that existing systems have a need to share information in order to be able to interact in some kind of coalition. This must also be done without claiming major changes to the systems, and without any requirements of knowing the other systems' intention beforehand. Nations will unlikely change their C2 systems in order to be able to interact with other nations. Nor is it likely that they want to adapt their C2 systems every time a new nation will integrate. The optimum for each C2 system is to "talk and listen" in their own language. In addition, the general situation is that of a sender creating a message without knowing in advance who the receiver will be. 3.4. Brief description of SIF, its main components and functions The basic idea of SIF is to foster the use of a semantic description of all of the information to be exchanged and then take advantage of a number of existing and emerging semantic technologies, mainly ontologies, to improve interoperability. Figure 1 shows an overall view of SIF which can be described as follows. SIF mediates an exchange of information between systems A and B, which do not necessarily know each other. Furthermore, the assumption is that the systems information structures are different and therefore the exchange of information cannot happen painlessly. This means that to make the communicated information correctly interpreted in accordance with the semantics of system B a transformation is required for all information that system A communicates. A number of ontology operations take place in order to define and produce the rules necessary for these transformations. Input to these ontology operations and transformations are not only semantic descriptions of systems A and B, but also references to potential shared concepts and definitions which will exist in the "Common Ground" (CG). Figure 1: An overall view of SIF. The most important components of SIF according to Figure 1 are as follows. The main purpose of Common Ground (CG) is to provide knowledge resources that will serve as common references for the semantic descriptions supplied by independent systems, in order to produce accurate ontology mappings. The idea here is that a portion of "all knowledge" available in the world, either exist or can be made available in machine-readable form. If this available machine-readable knowledge proves to be useful, reliable and validated for military use, it can be placed in CG to support SIFs ontological activities. An ontology manager within SIF provides services for ontology operations that identify similar concepts across ontologies and otherwise harmonise and align ontologies. Translation rules are the output of the mappings between concepts in the Common Ground, schema definitions, etc. Transformation is used to convert a message from a form which was suitable for system A into a form which is appropriate for system B. It is important to note that the structure of the message is converted without loss of semantics. For more details on SIF the interested reader is directed to NATO IST 075 Final Report [2]. The major functionality of SIF is to facilitate the exchange of messages (information) by the help of above described components. The information exchange is orchestrated into a number of stages, which we have directly considered when proposing our solution for a semantic-interoperability process (Section 5). 4. Modeling and Simulation Modeling and simulation (M&S) technologies play an ever-increasing role in supporting military applications such as training, research and development, analysis, test and evaluation. The M&S community has tackled interoperability-related problems for many decades. Since the late 1980’s, there have been serious efforts to address the related problems of interoperability and reuse by encouraging the development of simulations according to well-defined standards. The Simulation Interoperability Existing system Existing system Common Ground Transformation Semantic Description (A) Semantic Description (B) World Knowledge Translation Rules Ontology Operations