One of the main tasks of any distributed computer system is the analysis of the properties of the obtained data, and their further use for logical reasoning. This is a difficult task due to the frequently encountered dynamics and heterogeneity of information. The aim of the work is to present a knowledge-based web service model. The web service integrates with the technologies of the semantic web in order to study the data, their processing and inference. The goal is achieved as follows: the meta-metamodel of the web service ontology and the metamodel of the semantic web service are structured using UML class diagrams and formalized using ALC description logic. The scientific novelty of the results is the presentation of a service description through an ontology. Through the introduction of ontologies, a transition is being made in data mining. In contrast to the application of standards: SOAP, WSDL, HTTP, the use of ontologies is argued that they contain structured information about the functional and non-functional characteristics of the service, and give flexibility to the description. When working with ontologies, there is a powerful logical apparatus for searching, combining and comparing ontologies. By introducing ontologies, knowledge of the web service is represented. The development of an ontology / web service pair gives an advantage over the SAWSDL, OWL-S standards, because the web service developer has the ability to create a more accurate semantic description of the service. We presented the language features and constructs of ontology of web service descriptions. Step by step, we created semantic description to describe web services. After learning how to use OWL to formally express web service semantics, we moved on to the issue of actually adding semantics to service descriptions. For the first time, formal models of service ontology and semantic web service are proposed.
Copyright © 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Distributed computer systems (DS) based on a service-oriented architecture (SOA)
are independent of development technologies and platforms, while applications
running on one platform can call services running on other platforms in a standard way
[
Since WS can operate at a higher level of abstraction, analyzing and
processing data types in a dynamic way, the individual software components are given the
opportunity to interact more openly [
SOA is a heterogeneous environment, and it is developing towards creating a more
structured set of solutions where WS should be represented in some unified way and
equally detectable, able to communicate with other objects, and also be directly
integrated with the Internet infrastructure and other services, independently from the
functionality of the service [
Using semantic web technologies in the development of WS in the DS allows you to: go to the intellectualization of the description of XML-based interfaces and interactions, combining any type of application with another application, and providing freedom of change and development over time as long as the corresponding interface is supported; use a higher level of software abstraction and data description logic; take into account the weakness of the software, due to which the interaction between the applications of the service is not broken every time the design or implementation of a service changes; provide existing or legacy software service interface without changing the original applications; make decisions based on available data.
Key idea of semantic web services is to annotate web services with concepts which
are defined in formal logic-based ontologies such that, from an artificial intelligence
perspective, intelligent agents and service-based applications can actually reason on
such formal service semantics. In contrast to web service descriptions, this may
facilitate not only the semantic interoperation between services but their automated
logicbased composition planning and a more precise service search [
The aim of the work is to present a knowledge-based web service model. The WS integrates with the technologies of the semantic web in order to study the data, their processing and inference. The scientific novelty of the results is the presentation of a service description through an ontology. Through the introduction of ontologies, a transition is being made in data mining.
SOA is an approach for designing and developing DS [
Web services are modular, self-describing, selfcontained applications that are
accessible over the Internet [
Description of services in a language e neutral manner is vital for the widespread
use of WS. For general usability, a service must be described and advertised. WSDL
takes care of the description by providing a language to describe a service in enough
detail to invoke any of its operations. Service providers describe their WS and
advertise them in a universal registry called UDDI [
Prominent languages and formats for semantic service description are OWL-S
(Web Ontology Language for Web Services) [
SOA The SOA information model is registered with UDDI. Further, the client of the service can find it, call it, perform a certain task. The standards for exchanging information between services are SOAP, WSDL, UDDI, and a normal HTTP request. The SOAP and WSDL protocols provide a unified markup language for transmitted messages. It represents the functionality of a device connected to its managed resource. A user interface that provides a clear and standardized interface offers all the necessary functionalities for interacting with objects and related processes.
In order for SOA to enjoy greater success than it predecessors, it should consider the following attributes: scalable: The past solutions were not designed with the scale of the web in mind.
SOA should work in a variety of settings, such as within an organization, between
business partners and across the world [
This is accomplished through services such as a directory of service descriptions;
abstraction: A SOA abstracts the underlying technology. Developers can
concentrate on building services for business users rather than connecting systems and
applications [
The current description of the interaction between WS is as follows [
The semantic web should enable greater access not only to content but also to services
on the web. Users and software agents should be able to discover, invoke, compose,
and monitor web resources offering particular services and having particular
properties, and should be able to do so with a high degree of automation if desired. Powerful
tools should be enabled by service descriptions, across the web service lifecycle [
The semantic seb concept introduces formal definitions called ontologies, which allows you to build models of heterogeneous objects in a domain, share knowledge and support the automation of the formulation of logical conclusions from this knowledge.
To create a holistic system for deploying distributed knowledge-based systems (DKBS) semantic web services in a specific field of activity, it is necessary to develop a formal ontological model of objects (web services) that are part of SOA.
By combining the key elements of SOA with knowledge-based systems, you can
get a formal DKBS model. Figure 1 shows the structural model of DKBS. We single
out the concept of the main artifact of the system through the Element class, its
subclasses are such DKBS elements as a service, system, event, human actor, semantic
service [
A formal semantic service model should include: 1. Analysis of the structure of knowledge of the subject area of the DKBS, the main objects and relations: (a) the use of least-logical and ontological methods for the formation of the terminology of the system domain and knowledge structure; (b) structure of OWL ontology. 2. Repeated use of existing thesauruses, taxonomies, and ontologies of the DKBS domain: (a) semantic search for relevant objects and analysis of means of representing knowledge and standards in the DKBS; (b) a brief overview of relevant ontologies and other knowledge structures; (c) integration of existing taxonomies and domain ontologies. 3. The architecture of methods for the automated improvement of the formal ontological model of the DKBS: (a) the architecture of methods for the automated extraction of knowledge (terms and relationships) from natural language texts that relate to the DKBS; (b) methods of automated linguistic processing of natural language texts; (c) advanced OWL ontology. 4. Semantic search in the system based on the domain ontology: (a) semantic search for objects; (b) methods of semantic search for objects of the DKBS; (c) methods of semantic search for SOA services; (d) recommendations regarding the use of WS.
Search data for web services, their interactions, reviews, recommendations from service customers can be analyzed and converted into active knowledge, which allows us to better understand the physical world and create more value-added products and services.
DKBS is a synthesis of SOA and semantic technologies, which defines a serviceoriented presentation of software and hardware components and a description of their formal semantics. Table 1 shows the architecture of the meta-modeling of the DKBS.
A web service is a software system identified by a unique web address (URL) with standardized interfaces.
Semantic web service (SWS) - complete elements of program logic with uniquely described semantics, accessible via the Internet and suitable for automated search, composition and execution, taking into account their semantics.
A semantic meta-modeling of a WS is proposed (a meta-metamodel of a web service ontology, a meta-model of a WS) Thus, a new system object is formed - a SWS. In contrast to the application of standards: SOAP, WSDL, HTTP, the use of ontologies is argued that they contain structured information about the functional and nonfunctional characteristics of the service, and give flexibility to the description. When working with ontologies, there is a powerful logical apparatus for searching, combining and comparing ontologies. By introducing ontologies, knowledge is represented in the system. Developing an ontology / web service pair provides an advantage over SAWSDL, OWL-S, because a web service developer can create a more accurate description.
The meta-metamodel of the service ontology is shown in Figure 2, where Concepts are shown, i.e. units of information or data. Each concept has its own Logical Definition and a Textual Definition. In accordance with the concept, there may exist Instances, which, through the presentation of specific entities, connect abstract concepts with objects from the real world. In the same way, each Property of a concept has different Term - these are interpretation functions. Therefore, the meta-model of the service is its ontology, which describes all the characteristics of the service and its parameters (Fig. 3). A service primarily consists of Methods and Operations. The next component of the service is its Interface, which depends on already defined methods and operations. Binding is a component that provides the detailed information needed to access EndPoint. EndPoint is a specific service implementation where EndPoint associates a specific binding with a real address so that the service can be called. It is also important to understand the obvious difference between OWL and OWLS: OWL is an ontology language; it provides constructs and features we can use to create an ontology document. OWL-S, on the other hand, is just one such ontology created by using OWL.
Ontology is based on description logic (OWL-DL), which provides a platform for formal and machine-knowledge. To establish its syntax and semantics. Syntax, certain expressions (concepts, axioms, roles, etc.) are considered correctly constructed in this logic. How to interpret these expressions, i.e. gives them formal meaning. Let CN {A1,..., Am} and RN {R1,..., Rn} be finite nonempty sets of atomic concepts and atomic roles (also called concept names and role names). There are several dialects of descriptive logic that differ in expressive capabilities. The main operators of DL:
T – concept «THING» ; – concept «NOTHING» ; – logical connective (conjunction); – the quantifier of existence; – quantifier value limitation; - logical connective (negation).
The syntax of DL, namely the ALC dialect, is a lot of concepts that are defined by an inductive definition: ─ symbols T and - concepts (called truth and falsehood); ─ - every atomic concept A is a concept; ─ - if C is a concept, then C is a concept (called a complement to the concept); ─ - if C and D are concepts, then C D and C D are concepts (intersection and union); ─ - if C is a concept, and R is an atomic role, then R.C and R.C are concepts; ─ - no other expressions are concepts.
In what follows, we will use a shorter notation to formulate the syntax. So, the syntax for the ALC logic concepts in this entry is as follows: | T | C | C D | C D | R.C | R.C (1) where A is an atomic concept, R is an atomic role, D are arbitrary concepts.
A complete description of the meta-metamodel of the service ontology in the formulations of the discription logic with the class diagram (Fig. 2) is given in Table 2.
The semantics of logic is defined using the concept of interpretation. Interpretation is a pair I (,I ) , consisting of a nonempty set , called the domain of this interpretation and the interpretive function I , which is: each atomic concept A CN is an arbitrary subset of AI ; each atomic role R RN is an arbitrary subset of R I . EndPoint Specification Binding Methods Operation Interface Service Taxonomy Result Text Name
For a service ontology, domain interpretation will be the set of all ontology definitions. Atomic concepts are comparable to the set of all existing rules and terms for describing the ontology. We interpret atomic roles as two-place relations connecting all knowledge with a concept. Then everything listed in Formula 2 acquires the following semantics: a concept with this interpretation means:
Concept Relation - a lot of concepts of interconnection (i.e., relationship); Concept Term - many concepts of terms (term); hasPartOf.Concept - many concepts that are part of this concept; Property hasProperty.Concept - many properties that are property of the concept.
A full description of the semantic web service model in the formulations of the description logic is given in the table 2. hasObject Thus, the formation of ALC logic concepts for the SWS model is as follows: Service Methods | Definiton Text | hasParamet ers.Service | Value hasParamet ers.Service (3)
Service Methods - many service methods;
hasParameters.Service - many concepts that are service parameters; Value hasParameters.Service - many values that are service parameters.
Thus, a semantic web service is a pair of ontology and web service, which are described above by the presented models. Each formal model of descriptive logic is easily ported to OWL. 6
Formal models of ontology and semantic web service are aimed at supporting the process of automated deployment of intelligent applications in heterogeneous environments and allows: hide the technological heterogeneity that is characteristic of many heterogeneous web services; hide the semantic heterogeneity inherent in the used heterogeneous domain ontologies in order to semantically annotate the data of semantic web services. The following is the process of finding the web services the user needs. In this case, the user can specify both the functions or properties of the service itself, as well as the characteristics associated with the end result of the functioning of the service.
At the same time, the user can apply the terms from any ontologies available to him that describe the DKBS space and the corresponding subject area (with reference to the ontologies themselves). But in the description of the semantic web service, terms from other ontologies related to close domains can be used.
Therefore, the problem of comparing ontologies arises. In the general case, this problem is extremely complex and laborious, but if there are a number of restrictions on the comparable ontologies, it can be solved in an acceptable time.
For the intellectualization of data and the processing of knowledge stored in ontologies, logical inference blocks (reasoners) are used, which play a key role in knowledge-oriented intelligent systems.
There are many implementations of logical inference processors (reasonіng engіne)
for OWL ontologies that differ in capabilities, applications, and quality of task
execution. The generalized analysis allows us to divide them into three groups depending
on the implementation method:
1. Table Dl-processors. Traditionally, they were developed first to solve such
problems. They have low performance, but are able to make conclusions on complex
ontologies with many non-trivial constructions. This class includes resonators
Pellet, RacerPro, FacT ++, as well as Hermіt and SHER [
This group belongs to KAON2. [
There are two approaches to the implementation of inference: based on the rules (using the algorithms forwardchaіnіng and / or backward-chaіnіng) and based on the semantic scoreboard (semantіc tableau). Based on the rules, Semantіcs, SDK and Owlіm are implemented, and on the basis of the semantic scoreboard - Pellet.
Today, the Pellet system implements the most expressive descriptive logic using a high-performance logic inference (tableau-based algorіthm), which is used to process ontologies described in the OWL DL language. 7
We presented the language features and constructs of ontology of WS descriptions. Step by step, we created semantic description to describe SWS. For instance, given that WSDL is also used to describe WS, understanding the relationship between WSDL and OWL is important for Semantic Web developers. We first built two models (structural and formal) to describe the ontology and web service. Examples of semantic interpretation of each model are given. Presented existing tools for further work with the obtained ontologies.
The advantages of using ontologies for WS: 1. Automatic discovery of SWS: Finding the desired service can be hard, especially when the service requester does not know of the existence of the service provided. However, to make SWS a real success, a way to discover the requested service should be provided; also, it has to be discovered automatically, with great accuracy and efficiency. 2. Automatic composition of the necessary services. Clearly, a software agent should be able to find all the necessary services and invoke them in the correct order to accomplish the business goal.