A comprehensive framework to increase efficiency of IT-services management systems (ITSMS) is proposed, which resolves 3 interconnected tasks in a target organization: 1) providing an effective configuration of ITSMmodules according to their specific features and needs; 2) integration a given ITSMS with existing enterprise architecture; 3) advanced incidents management in ITSMS. The applicability of this approach was tested successfully on the case-study at the National Technical University “Kharkov Polytechnic Institute” (www.kpi.kharkov.ua ).
(MDA) [
Taking into account some ITSMS-issues mentioned above, the main objective of the research presented in this paper is to propose the first vision for intelligent complex approach to increase efficiency of typical ITSMS, with a proof of concept basing on the ITSMS university case-study. The rest of this paper is organized in the following way: Section 2 analyses some existing ITSMS, introduces our vision about their typical functionality, and shows the list of prioritized tasks to be resolved to increase an efficiency of an ITSMS. In Section 3 we present the method elaborated for effective ITSM-modules configuring in a target business organization, and Section 4 reports the first version of ITSMS - ontologies to integrate the selected modules into enterprise architecture (EA). In Section 5 the designing perspective for the combination case-based reasoning (CBR) with ontology-based approach to advanced incident management it ITSMS is briefly outlined. In Section 6 we present the university casestudy for our method to estimate an effectiveness of different ITSMS configurations and discuss the results achieved. In Section 7 the paper concludes with a short summary and with an outlook on the next steps to be done in the proposed development framework. 2
In order to elaborate a way how to provide a complex approach to increase an efficiency of ITSM-system operating, it is necessary to understand its typical functionality and to analyze its specific features. 2.1
We have analyzed some already existing ITSMS [
To the group (a) belong such systems as, e.g., HP OpenView Service Desk [
ITSM-solutions from the group (b) also are used in practice, but they definitely
have limited functionality and provide less level of IT-services management. The
typical open source ITSMS are, for instance, GLPI [
And, objectively, the business organizations, which are not ready to buy advanced
software products from group (a), and which are not satisfied with functionality
provided by ITSM-systems from group (b), because they have some specific IT-needs
and challenges, exactly these companies try to develop their own ITSM-solutions to
be considered as members of the group (c). The more detailed comprehensive study
of some existing ITSM-systems is presented in [
Based on the given analysis of the real ITSMS (see above), we have elaborated the following vision for their typical functionality (see Fig. 1).
There are 5 main subsystems (or packages) of system functions, namely: 1. IT Business Alignment: this subsystem is supposed to implement a ІТ-strategy in given business organization with respect to its main goals and needs, and to provide a base for costs assessment to whole IT-infrastructure; 2. Service Operations: this facility is responsible for customer’s requests management (regarding to a current incident and to a related problem), and for providing of ITSM-support functions; 3. Service Delivery Assurance: this functional package implements a configuration and change management of all ITSM-software tools thus is extremely important for a stable IT-environment; 4. Service Design and Management: this ITSMS-functionality provides detailed information about new perspective IT-services to be designed with respect to their availability and quality for IT-customers; 5. Service Development and Deployment: this subsystem allows to create and to test new ITSM-services and appropriate IT-infrastructure solutions, including installation of new hard-ware components, development of additional software applications, and training programs for ITSM-staff’ and for end-users as well.
As we can see on the structure presented in Fig.1, each of these 5 subsystems is built from several functional modules (they are depicted as UML-classes). The most important of them are the following ones: Module М1 =”Incident management”: it includes organizational procedures and appropriate tools to resolve current incidents, which IT-service users are facing with (hard-and software errors, network connection problems, request for consultations, etc.);
Module М2 = ”Problem management”: this facility provides tools to detect and to eliminate any problem situation which is a reason for different incidents; Module М3 = ”Configuration management”: this module supports all operating sub-schemes in the IT-infrastructure of given business organization; Module М4 = ”Change management”: it supervises and coordinates all changes which arise in IT-infrastructure; .
Module М5 =”Service level management”: this unit is responsible for definition and implementation of an appropriate level of IT-services to be provided for customers.
In ITIL-best practice manuals (see e.g. in [
S1 M1, M 3, M 4, M 2, M 5
This approach quickly allows to resolve the most actual communication problems between IT-service department and customers basing on incident management (module М1), and it provides some tools for all IT-services support (the modules М3 and М4), and after that a platform for future IT-infrastructure development is introduced (modules М2 та М5 respectively). But in this case it has to be taken into account this scheme is a most expensive way for a given business organization, and it requires a lot of resources exactly at an initial phase of whole ITSM-configuring framework.
A contract scheme S2 actually aims to formalize a communication process between IT-service department and customers, and it has the following modules-workflow:
S 2 M 5, M 3, M1, M 4, M 2
In this case all customer requirements to IT-services have to be collected and specified (in module М5), and appropriate IT-infrastructure sub-schemes can be built (using module М3), in order to define prospective IT-strategy in the target organization, next an operative ITSM-functionality is provided, including incident management (in module М1), change management (in module М4), and problem management (in module М2). Obviously, this scheme definitely has some risk factors regarding its efficiency, if the initial IT-service specifications were done not correctly (in module М5).
And, finally, an infrastructure-centered scheme S3 proposes the modules sequence indicated as following:
S3 M 3, M 4, M 2, M1, M 5 (3) that is, firstly, to provide tools for all IT-services support (modules М3 and М4 respectively). Secondly, this approach allows to manage all typical problem situations (in module М2), and already based on this one to detect and to resolve corresponded incidents by IT-service customers (in module М1). Thirdly, it creates an opportunity to define in computer-aided way the necessary composition and the IT-service level management (in module М5).
It is necessary to note that besides some empirical recommendations concerning the possible ITSM-modules configurations defined as (1)-(3), in the appropriate technical documentation there are no more or less proved suggestions about possible quantitative estimations for effectiveness of these alternative approaches. 2.3
Taking into account the results of performed analysis (see above), and based on some modern trends in the domain of ITSMS-development (see Section 1), the following list of prioritized tasks can be composed in order to increase ITSMS-efficiency, namely (1) (2) 1. to provide an effective configuring of ITSM-modules for a target organization, taking into account its specific features and needs; 2. to elaborate an integration framework for a given ITSM-system’s configuration and for an existing enterprise architecture (EA); 3. to support an advanced incidents management in the already installed ITSMsystem.
In our opinion, the task (I) can be resolved basing on some expert methods for
multi-criteria ranking, with respect to specific IT-infrastructure’s features and
customer needs in a concerned business organization [
To formalize the task (I) from their list considered in the Section 2.3, namely: to provide an effective configuring of ITSM-modules for a target business organization, the following factors have to taken into account: such a problem has a high complexity grade and it is semi-formalized; estimation criteria for it are of different nature and they are multi-valued; an information base to solve this task mainly can be collected basing on expert data only; available expert data could be quantitative and qualitative values both.
To solve this task we have chosen one of the multi-criteria ranking methods,
which is presented in [
K {K1, K2,..., Km} {K j , j 1, m} have to be defined.
Step 2. Each global criteria K j is characterized by a subset of appropriate local criteria
K j {k j1, k j2,..., k jQ} {k jq , q 1, Q} further, a set of membership functions according to all local criteria alternatives
{ k j1 (xi ), k j2 (xi ),..., k jQ (xi )} { k jq (xi ), q 1, Q, j 1, m} and the weight coefficients of their relative importance for these local criteria {w j1, w j2 ,..., w jQ } {w jq , q 1,Q} have to be determined, where the following condition has to be fulfilled Step 3. To determine membership functions of alternatives { xi , i 1, n} to criteria K j ,{ j 1, m} based on an additive convolution of their local criteria Q w jq 1 q1
Q k j (xi ) q1 w jq k jq (xi ) all alternatives xi ,{i 1, n} it is possible to determine a joined membership function for a generalized criterion K : (7) (8) (9) (10) (11) (12) m K (xi ) w j K j (xi )
j 1 (x*) max{ K (xi ),i 1, n}) where w j , j 1, m are coefficients of their relative importance K j , j 1, m . Step 5. Finally, an alternative with a maximum value of membership function for generalized criterion K can be chosen as a target solution:
Below in Section 6 we present the case-study, which was performed to prove this method, and we discuss the results achieved.
As already mentioned above (see Section 2), any ITSMS has to be integrated into an existing EA of a target organization. In our approach this task (II) has to be resolved for an ITSMS-configuration defined with the method presented in Section 3.
This issue is already discussed intensively in a lot of publications, and their authors
consider both its conceptual and technological aspects. E.g., an ITSMS-EA
integration based on well-known SOA – framework is presented in [
Onto _ ITSM EA Onto ITSMS ,Onto EA,Onto TO (13)
Obviously, some already existing ITIL / ITSM ontological specifications can be
used for this purpose, e.g.: Onto-ITIL ontology elaborated in [
Onto ITSMS Onto ITIL,Onto SPEM ,Onto WF
There are also several ontologies developed to specify EA, and according to one of
recent and comprechensive researches in this domain presented in [
Onto EA Onto BT,Onto AC,Onto RS where: Onto_BT is a sub-ontology of Business Terms (BT), Onto-SC is a subontology of Architecture Components (AC), and Onto-RS as a sub-ontology of RelationShips (RS) among items of AC.
And finally, to define an Onto-TO ontology for target organization given in expression (13), its specific features and needs related to ITSMS-usage within existing EA have to be taken into account. As a small excerpt of such domain-specific Onto-TO, which is elaborated in our University-ITSMS case-study (see Section 6), the following UML-class diagram in Fig. 2 is shown. The proposed ontology-based approach for ITSMS-EA integration can also be used to elaborate the solution for the task (III) from their list completed in Section 2.3. 5
In order to solve the task (III), namely: to provide an advanced incidents management in ITSMS, accordingly to our inter-disciplinary vision about the ITSMS-development in general, we propose to amalgamate the following design-principles (i)-(iv) listed below (i) an incident management as a weak-formalized and complex task within the ITSMS-support for its customers can effective be resolved using one of the intelligent decision-support methods, e.g., using CBR-method; (ii) to enhance a CBR-functionality, especially with respect to specific needs in a target organization, an appropriate domain-ontology should be elaborated and used combining with CBR; (iii) because of the permanent changes in an IT-infrastructure of a given organization, and of the changes arising in its environment as well, such a domain-ontology has to be constructed as an adaptive ontology; (iv) to provide a possibility for knowledge gathering and their reusing in ITSMS, some e-Learning models and technologies can be applied.
There are already the approaches elaborated to combine a CBR-method with
ontologies [
These results provide a solution for the tasks (i)-(ii), but in our opinion to cover the
task (iii) in efficient way, with respect to permanent changes in IT-infrastructure of a
target organization, an appropriate ontology has to be constructed as an adaptive
facility [
Onto TO (adapt ) C , R, P,W (C ) ,W ( R ) (16) where, additionally to the basic components of any ontology, namely: C – set of concepts, R – a set of relationships among these concepts, and P – a set of axioms (semantic rules), the following ones have to be defined: W(C) is a set of weight coefficients for concepts of C, and a W(R) is a set of weight coefficients for relationships of R respectively. Usage of these weight coefficients allows us, e.g., to take into account an appropriate importance grade in several types of ITSMS-customers (see Fig. 2) to provide IM - services for them.
In order to get all information resources needed for a completed solution of tasks
(i)(iv), we propose to apply some e-Learning models and technologies within an
ITSMS, especially, for skills training and experience gathering by ITSMS-staff,
designated in the Onto-IM ontology as Incident Manager, ServiceDeskEmployee,
Specialist [
Summarizing aforementioned issues concerning the tasks listed in (i)-(iv), the conceptual mechanism to provide an advanced incidents management (AIM) in ITSMS can be represented at the high-architectural level as the UML-package diagram shown in Fig.4.
Below the approach to resolve the task (1) from their list given in Section 2 is illustrated using the real case-study within our research and practice activities to apply ITSMS to manage IT-infrastructure of National Technical University “Kharkov Polytechnic Institute” (www.kharkov.ua) referred in following as NTU “KhPI”.
It is to mention that exactly university- and /or a campus-domains are considered from
many authors as a suitable example of ITSMS-usage (see, e.g., in [
The NTU “KhPI” is one of the largest technical universities of Ukraine located in the city of Kharkiv, which is the important industrial and cultural center at the East of the country. The university has about 22000 students, ca. 3500 of faculty members, and accordingly there is the advanced IT-infrastructure to support all educational and research tasks. The main characteristics of IT-operating at the NTU “KhPI” are summarized in Table 3.
In cooperation with the IT-staff at the University IT control office we have analyzed retrospective data about some typical problem situations occurred, and about the corresponded incidents, which daily have been resolved within the direct communication with IT-service customers. In this way the main types of ITSMincidents and their initial problem situations were identified, and they are described in Table 4. Incident type Cause ( problem situation) No Internet-connection at - router was turned off; Dept or on local PC - network cable breaked or failure on router hardware; - incorrect network setup; - problems with software on local PC Incident type Cause ( problem situation) High-loading of PC processor - computer viruses with a small number of active - high degree of PC hard driver’s de-fragmentation. user’s programs Installing problems for new - computer viruses software - absence of additional (middleware) software needed for installation. 4 Failure to send email -incorrect setup of local network server (proxy) -problems with central e-mail server. 5 Troubles in the use of third- -lack of specific configuration,
party software - improper use of system services.
Basing on the analysis results obtained, we can apply the elaborated method to estimate alternative ITSMS-module configurations (see Section 3). 6.2
configurations and criteria definition estimation method: alternative According to the Step 1 of the method presented in Section 2.2, the list of alternative ITSM-module configurations have to be defined, and in our case they are presented: X1 = Service Desk subsystem (SDS) and Incident Management Module X 2 = SDS, Incident Management Module and Configuration Management Module X 3 = SDS, Incident Management Module and Change Management module X 4 = SDS, Incident Management Module and Problem Management Module On the next Step 2, according to the formulas (4) - (10), we determine the criteria for the quantitative evaluation of the proposed alternatives and their performance indicators, which are shown in Table.5. These criteria and their indicators (metrics) are taken from [35], and they are recommended to evaluate effectiveness of ITinfrastructure in any business organization. Global and local criteria
Semantics performance measurement criteria and target values K1 k11 k12 k13 k14 K 2 Global and local criteria
Effectiveness of incident management Average time incident resolution →min Percentage of incidents resolved proactively →max Semantics performance measurement criteria and target values Percentage of incidents resolved at the first level of support →max Percentage of incidents that have been resolver from the first time →max Effectiveness of problem management k11 (opt =20м) 5 6 5 7 5 7 5 6 5 6 5 8 6 6 6 7 X1 X 2 X 3 X 4 … Insecure value >30 min. 0% Insecure value <65% <75%
Effective value To continue the usage of our method presented in Section 2.2 (Step 3 and Step 4 respectively) using the pair-wise comparison the weight coefficients of relative importance (WCRI designated as w(ki, j ) ) for the local criteria regarding their global ones were determined: The WCRI values of the local criteria for the global criterion K1 : w(k11) = 0,239458, w(k12 ) = 0,239458, w(k13) = 0,432749, w(k14 ) = 0,088335 The WCRI values of the local criteria for the global criterion K2 : w(k21) = 0,68334, w(k22 ) = 0,19981, w(k23 ) = 0,11685 0,332516, w(k32 ) = 0,527836, w(k33) = 0,139648 The WCRI values of the local criteria for the global criterion K3 : w(k31) = The summarized WCRI values for the global criterion Ki : K1 = 0,527836, K2 = 0,332516, K3 = 0,139648
And finally, according to Step 5 of this method (see Section 3.2), and using the multi-criteria ranking formulas (11) - (12), we obtain the following ultimate results of the effectiveness assessment for the considered alternatives (see Table 5), namely X1 0.537, X 2 0.671, X 3 0.578, X 4 0.727 (17)
To confirm the reliability of the results given in (17), the comparative analysis with
some "best practices" in ITSMS implementation was carried out, using the data of
IDC-company [
As we can see, to provide Change Management and Configuration Management is necessary to have within an IT-infrastructure database (DB) of IT-configurations, and DB of problem situations as well, these facilities are rather costly for the University, and therefore the implementation of these modules is not a priority task. The most effective ITSM-modules configuration for NTU "KhPI" includes an Incident Management module and a Service Desk subsystem. 7
In this paper we have presented the intelligent approach to increase efficiency of ITSMS, which has to resolve 3 interconnected tasks for its effective usage in a target organization: 1) providing an effective configuration of ITSM-modules according to its specific features and needs; 2) elaboration an integration framework for a given ITSMS with existing EA; 3) advanced incidents management in ITSMS. To solve these tasks in a comprehensive way the interdisciplinary framework is elaborated, which includes: the expert method for multi-criteria ranking of alternative ITSMmodules configurations, the ontological specifications for ITSMS-EA integration, and the approach to enhanced incident management based on the combination of adaptive ontologies and CBR-methodology. To implement the first part of this approach the appropriate software tool was elaborated, and its applicability was tested successfully within the case-study at the NTU “Kharkov Polytechnic Institute”.
In future we are going to implement and to test the appropriate software solutions for other tasks in the proposed framework, using such technologies as OWL, BPMN, XML /XLST, and Web-services.