SIMPLIFIED CBA CONCEPT AND EXPRESS CHOICE METHOD FOR INTEGRATED NETWORK MANAGEMENT SYSTEM

SIMPLIFIED CBA CONCEPT AND EXPRESS CHOICE METHOD FOR INTEGRATED NETWORK MANAGEMENT SYSTEM Mohammad Al Rawajbeh 1, Vladimir Sayenko 2 and Mohammad I. Muhairat 3 1 Department o Computer Networks, Al-Zaytoonah University o Jordan, Amman, Jordan 2 Kharkov National University o Radio Electronics (KhNURE) Kharkiv, Ukraine 3 Department o Sotware Engineering, Al-Zaytoonah University o Jordan, Amman, Jordan ABSTRACT The process o choosing and integrating a network management system (NMS) to an existing computer network became a big question due to the complexity o used technologies and the variety o NMS options. Most o computer networks are being developed according to their internal rules in cloud environments. The use o NMS requires not only inrastructural changes, consequently increasing the cost o integration and maintenance, but also increases the risk o potential ailures. In this paper, conception and method o express choice to implement and integrate a network management system are presented. Review o basic methods o cost analysis or IT systems is presented. The simpliied conception o cost beneits analysis (CBA) is utilized as a basis o the oered method. A inal estimation is based on three groups o parameters: parameters o expected integration risk evaluation, expected eect and level o completed management tasks. The explanation o the method is provided via example. KEYWORDS Network Management; NMS; CBA; Integrated System; Cost Analysis; Eiciency Criterion 1. INTRODUCTION It is early or late, but or any complexity computer network the updating process begins. This process changes not only the hardware o computer network but also its inrastructure. The sotware structure is changed, and unctionality properties are changed too. This problem is described very nice in [1, 4]. The responsibility on these actions is an administrator. He makes the update operations according to the pre-determined network management concept. The basic principals o those concepts are considered in [1, 2]. The simplest schema o this concept according to [1] is FCAPS (Fault, Coniguration, Accounting, Perormance, and Security). The more complex schema is Microsot Framework MOF 4.0 [3]. For the considered task we can use the simplest schema FCAPS and choose the Coniguration Management unctions. According to MS MOF 4.0 and Management unctions (SMF) we can choose "Change and Coniguration SMF". In the MS MOF 4.0 or C&C SMF we ind a good description o the updating process or the change procedure and the descriptions o the correspondence responsibilities roles. The choose procedure or the alternative sotware elements is an administrator's decision. For the existence computer network inrastructure any changes leads to changes o the computer network state. These changes carried out the positive (A) and negative (B) eects. As a result, or each Sot i (Sotware updates) there are unique evaluations (A, B). The igure 1 shows that the administrator has to chose the best optimal solution. DOI: 10.5121/ijcnc.2016.8304 47

Figure 1. Choice solution or integrated sotware And he has two main questions How? and What?. The question How? means the evaluation method or (A, B). The methodology questions o that task is considered in [5]. The question What? means the rules and criteria to chose the concrete system rom the existence set. This choice is dependent on the type o sotware updating service or apps Sot i. Let s consider the task when Sot i = NMS (Network Management System). It means that the existence NMS should be changed or a new NMS should be implemented. The usage o management tools becomes actual, and it improves network administration, automated storage operations and management inormation analysis. The example o this tool is Network Management System (NMS) [1, 2]. The complexity o the managed system increases the complexity o the management system. In this situation, inrastructural changes will not only increase the cost o integration and maintenance, but will also increase the risk o potential ailures. Integration o the NMS may sometimes reduce the eiciency o the whole system as well. Based on the above, the problem o estimating the eiciency o integrating the new network management system into an existing inrastructure o inormation space is actual. In view o certain situations, the decisions about the integration o a special system oten must be taken rapidly. It depends on the real inancial condition o the company. These solutions could be considered as an independent IT e-space [13] or within separate unctions o network management rameworks. Such rameworks are inancial management MOF SMF [3] or ITIL [16]. They are parts o the general complex concept o the unctions and areas o network management rameworks. As a result, the considered task is presented as Figure 2. 48

Figure 2.Choice solution or integrated network management system The questions How? and What? are remained open, but the choice o NMS could be done according to FCAPS demands. It means that the evaluation o positive and negative eects could be considered according to FCAPS concepts: A(FCAPS), B(FCAPS). The answers on the questions How? and What? are bound with the evaluation procedures o A(FCAPS), B(FCAPS). They belong to the area o eiciency and cost analysis or the implemented systems Sot i. Additionally, there is the time restriction to make optimal decision (Time < T*). In such cases, the availability o methods that allow justiying such a decision is actual. Such methods and techniques are Total Cost o Ownership (TCO) [7], Total Beneits o Ownership (TBO) [8] and Return On Investment (ROI) [9]. These techniques were successully used in assessing the eectiveness o new inormation systems implementations, or example, Cloud system [11] and Data Center. O course, these are good techniques, but they have one big disadvantage. Actually, they are the ultimate economic methodologies that require qualiied experts, high research level and time expenses. Most o the techniques are based on three main approaches: estimation o costs associated with the implemented systems, estimation o implementation risk, and estimation o beneits. Here are the examples o some popular techniques that can help estimate the impact o the new IT solution integration or complex inormation systems. For example, TCO [7] allows us to estimate the total costs, TBO [8] helps evaluate the beneits o such integration, IT IR (IT Integration Risk) [12] allows analysing and evaluating the risk, ROI (ROI) [9] allows evaluating the eiciency o an investment. In [19] proposed the several dimensions o complexity that have to be considered during the building o cost management model (CM), and ound that, the reliability o the model is depending on the degree o complexity. Out o all described concepts, two o them are the most widely used; namely TCO and TBO. Both concepts are based on the determination o the purchase price o an asset plus the costs o operation, but they possess opposite ideas under the hood. The TCO shows how the costs will increase ater integration o a new system, whereas the TBO points to the beneits o the integration with the same system. The Cost-Beneit Analysis belongs to the ratio approach to decision-making processes [18]. 49

Another popular concept is Cost beneit Analysis (CBA) [6, 7]. It uses two components; one o them considers the cost o integrated inormation technology and the problems that rise ater the main system coniguration change (e.g., additional cost and eort associated with the restructuring o the management, the sta adaptation period and so on.). The second one considers the advantages that we get ater the system integration. It appears that the CBA is a more advanced approach or these tasks. The article consists o the review part, ormulation o the problem, description o the concept, description o the parameters, proposed method o analysis and examples. 2. EVALUATION CONCEPT 2.1. PROBLEM FORMULATION AND DESCRIPTION OF THE STUDIED OBJECT 2.1.1. THE AIM The aim o the article is to identiy the ways o express (simpliied) selection o Network Management Systems implementation and their integration. 2.1.2. THE RESEARCH OBJECT Assume that some company has deployed a computer network. There is a need or additional integration o specialized network management system. The question o the implementation or integration o complex (expensive) sotware is considered. At the same time, there are alternative ways to purchase one o several types o such sotware systems. Suppose there is an ITdepartment in the company. It is responsible or the network administration. The head o department should take the decision about integration in limited time. 2.1.3. PROBLEM FORMULATION It is necessary to have a method (technique) that allows making the decision o the NMS systems selection rom available alternatives. This method should be relatively easy to use. It should not demand additional specialist involvement rom other departments or companies. The number o method parameters should be small, the criteria should take into account the level o solving the basic problems; the results o the analysis should be understandable or managers and the network administrator. The basic approach o the problem solving is assessing the beneits o new system integration, estimating the level o existing problems, and resolving and assessing the possible problems arising rom the new sotware (system) exploitation. 2.2. COMMON CONCEPT OF EVALUATION Let us have three groups o criteria. The irst group describes the level o probable risk (E r (NMS)). It means that we highlight the most signiicant problems on administrator's mind. These problems could appear in the network ater new system integration. The second group covers the eectiveness o the network maintenance (E e (NMS)). We suppose that ater integration o the NMS, the general eectiveness is going to increase. In general the eectiveness is characterized as, D0 = E1 / E0, ( E1, E0) [0,1], where E1, E0 are the current and basic variables. In this case, we use time and money as the main characteristics. This is a complex 50

problem or direct values estimation, so we should use the indirect characteristics such as the maintenance o the network. We assume that i we improve the maintenance o the network, the main characteristics will be improved as well. The maintenance and eiciency will be considered or the independent management areas. There are several concepts o representation o management areas [1, 6]. In accordance with the concept based on the OSI RM (ISO) the management is divided into 5 areas or 5 basic management unctions - FCAPS (Coniguration, Perormance, Fault, Accounting, and Security) [1, 2]. It should be noticed that the proposed division o Microsot SMF MOF 4.0 suggests 18 unctions [3]. In addition, there is also a TMN (ITU-T) [1, 2] classiication. The third group evaluates the completion level o the general solutions on the system integration (E (NMS)). It is possible that the integrated system will not solve all o the management problems completely, and ater the purchase o the system, we will have to buy another one (This is an incomplete solution). So that, instead o abstract positive (A) and negative (B) variables, we can evaluate the eectiveness levels, completion and probable risk. The task is as evaluation o values or (Ee, Er, E). The Figure 3 shows that the task became clear or administrator. Instead o the question How? he can use the evaluation procedures or each considered NMS. Instead o the question What? he can use the analysis procedure and the set o criteria. All solutions will be conormed to FCAPS and MS MOF 4.0 (C&C SMF). All previous demands will be in orce. Figure 3. Choice concept or integrated network management system The concept o a NMS comparison may be represented in the orm o ive stages: 1. Evaluation o the level o probable risk or the integrated system. 2. Evaluation o the beneits ater system integration in the network maintenance. 51

3. Estimation o completion level ater system integration. 4. Evaluation o the generalized criterion o selection expediency o the system to be integrated. 5. Estimation scale ormation or comparative variables. The generalized criterion should be an integrated variable. Estimation scale ormation or comparative variables means that or all variables we should have a single scale. It should be with relative values on the interval [0, 2] with the control ixed points {0, 1, and 2} = {<bad>, <OK>, <good>}. This approach allows getting the normalization and standardization o all variable values and makes them uniorm. 3. EVALUATION CRITERIA 3.1. CRITERIA OF THE LEVEL OF PROBABLE RISK Let us introduce our types o criteria or this group: load ( Load ), complexity o the administration ( Adm ), network security ( Sec ) and cost o ownership or the system ( Cst ). "Load" means the approximate amount o traic that is generated by the new system throughout on network. Actually, it is an evaluation o the additional network load, created by the service management traic. It is determined as the total generalized estimate, based on the average load at the workstation Load( WS ), the average load at the segment Load( Seg ) and the average load at the server Load ( Ser ). Load WS = Load Sg Load Sr Nw 1 ( ) N Load ( WS ) w i= 1 i Ng 1 ( ) = N Load( Sg ) g i= 1 i Ns 1 ( ) = N Load ( Sr ) s i= 1 i (1) Where Nw, Ng, Ns are the number o workstations, segments, and servers, respectively. { Load ( WS), Load ( Sg), Load( Sr)} {( < low >, < medium >, < high > )} ={0,1,2}, (2) Load( Net) = l1load ( WS) + l2load( Sg) + l3load( Sr), (3) with the criterion Load ( Net) min, l1 + l2 + l3 = 1, hence assuming same values o coeicients l 1 = l 2 = l 3 = 1/ 3 «Complexity o the administration» ( Adm ),is an expert estimation. Adm { < low >, < medium >, < high > } ={0,1,2}, (4) with the criterion Adm min. The point is as ollows. The integration increases the administration process and could increase the complexity o the base network and its inrastructure. «Network security» ( Sec ) means that using another s NMS reduces the security level. This is an expert estimation 52

Sec { < low >, < medium >, < high > } ={2,1,0}, (5) with the criterion Sec min. «Cost o ownership» is complexity estimation. It consists o three estimations: cost o purchase ( Cst( In ) ), cost o deployment ( Cst( Use ) ), and cost on supporting ( Cst( Sup ) ). Each o criterions Cst( In ), Cst( Use ), and Cst( Sup ) will be estimated by an expert. Cst( p) { < low >, < medium >, < high > } ={0,1,2}, where p { < In >, < Use >, < Sup > }. There are three types o estimation: 1) each criterion Cst( p) is considered separately, 2) only one criterion is used, and 3) the average value o all criterions is used. The type o estimation is given by parameter c i. Cst( Net) = c Cst( In) + c Cst( Use) + c Cst( Sup), (6) 1 2 3 where c1 + c2 + c3 = 1, hence assuming c 1 = c 2 = c 3 = 1/ 3. To exclude type o estimation, use c i = 0. The criterion Cst( Net) min. Any system could be estimated by usage o the Cost o ownership. E ( NMS) = r Load ( Net) + r Adm( Net) + r Sec( Net) + r Cst( Net), (7) r 1 2 3 4 with the criterion Er ( NMS) min, r1 + r2 + r3 + r4 = 1, hence assuming r 1 = r 2 = r 3 = r 4 = 1/ 4. 3.2. GENERALIZED CRITERIA FOR THE EFFECTIVENESS According to ive Management areas (FCASP), it is recommended to use ive criterions: perormance ( Pe ), coniguration ( Ce ), Fault tolerance ( Fe ), security ( Se ) and accounting ( Ae ). In this group, Perormance means the state o data perormance (real throughput) in case o the integrated system usage. It is expected that the value o perormance criterion will increase ater the integration. This is an expert estimation. Pe { < not changed >, < changed >, < very high > } = {0,1,2}, (8) with the criterion Pe max. Coniguration means the act o presence o Management coniguration. Coniguration is an expert estimation. The coniguration management is optional. In case o presence, it could have dierent levels o depth and complexity, or example, logical and unctional management coniguration. In this case, it is multiple. Ce { < no >, < partially >, < multiple > } ={0,1,2}, (9) 53

with the criterion Ce max. Fault tolerance means the improvement in the detection, prediction and prevention o possible ailures elimination. It could be estimated by Minimum Time beore Repairing (MTBR) or by reducing the number o aults. It is recommended to use an expert estimation. Fe { < no improving >, < partially >, < very good > } = {0,1,2}, (10) with the criterion Fe max. Security means that the integration o a new system will increase the security level. This is an expert estimation. Se { < no >, < partially >, < very good > } ={0,1,2}, (11) with the criterion Se max. Accounting means the possibility to maintain user access control and release control o personal QoS. This is an expert estimation Ae { < no >, < partially >, < very good > } ={0,1,2}, with the criterion Ae max. (12) To sum up, we get estimation o generalized eectiveness criteria. E ( e NMS ) = e ( ) ( ) ( ) ( ) ( ) 1Pe Net + e2ce Net + e3fe Net + e4se Net + e5 Ae Net, (13) with the criterion Ee ( NMS) max. e1 + e2 + e3 + e4 + e5 = 1, hence assuming same values o coeicients i.e. e1 = e2 = e3 = e4 = e5 = 1/ 5. 3.3. COMPLETION LEVEL CRITERIA In general, this group can be represented by a list that does not coincide with the previous group; it will be more suitable when the lists are the same. Perormance is a number o criterions that characterize the problems, which are being resolved, and their level o completeness. Use the estimation P { < no >, < partially >, < yes > } ={0,1,2}, (14) with the criterion P max. Coniguration means the act o solving the task o coniguration management in the network C { < no >, < partially >, < all > } ={0,1,2}, (15) with the criterion C max. 54

Fault tolerance means the act o solving the task o ault management in the network. Use the estimation F { < no >, < partially >, < all > } ={0,1,2}, (16) with the criterion F max. Security means the act o solving the task o security management in the network. Use the estimation S { < no >, < partially >, < all > } ={0,1,2} (17) with the criterion S max. Accounting means the act o solving the task o accounting management in the network. Use the estimation A { < no >, < partially >, < all > } ={0,1,2} (18) with the criterion A max. To sum up, we get a generalized criterion that shows the level o problem solving completion. E ( NMS) = a1p ( Net) + a2c ( Net) + a3f ( Net) + + a4s ( Net) + a5a ( Net) (19) a + a + a + a + a, hence assuming same values o coeicients i.e. 1 2 3 4 5 =1 a = a = a = a = a. 1 2 3 4 5 =1/5 3.4. GENERALIZED CRITERION This inal generalized criterion is estimated as the ratio o sums o eiciencies ( E (NMS) + E e (NMS) ) to the criterion o expected risk E r (NMS). Since the value o criterion E (NMS) could be equal to 0, use a biased estimate (1 + E (NMS)). The inal equation will be r Q( NMS) = ( E ( NMS) + E ( NMS)) /(1 + E ( NMS)) (20) e r with the criterion Q( NMS) max. This inal equation takes into account all criterions and allows to compare all decisions. As a result, we have a convolution o the evaluations as it shown on Figure 4. r 55

Figure 4. Convolution o the evaluations. 4. EVALUATION METHOD The inal method o estimation consists o six steps. 1. General analysis o all considered existing systems. 2. Evaluation o the level o probable risk or the integrated system. This step should cover the ollowing points: the evaluation o the additional network load created by service traic (1), (2), (3), the probable increase o complexity o administering process (4), the probable decrease o security level (5), cost evaluation depends on integrated system such as: cost o purchase ( Cst( In ) ), cost o deployment ( Cst( Use ) ), cost on supporting ( Cst( Sup ) ) (6) and general estimation o probable risk Er ( NMS ) (7). Figure 5 illustrates the lowchart or the method o Evaluation o the probable risk level. 56

Figure 5. Flowchart or the method o Evaluation o the probable risk level 3. Evaluation o the beneits ater integrating the system in maintaining the network. This step should cover the ollowing points: the evaluation o the probable improving o perormance (8), coniguration management (9), network in general (10), security level (11), accounting or users (12) and evaluating the total value o expected eiciency Ee( NMS ) (13). Figure 6 illustrates the low charts or the method o Evaluation o the beneits (a) and Evaluation o the completion level (b). 4. Evaluation o the completion level ater system integration. This step should cover the ollowing points: the evaluation o the supporting level or perormance (14), or coniguration management (15), or ault tolerance (16), or security management (17), or user accounting (18), and evaluation o the total value o expected completion level E ( NMS ) (19). 5. Evaluation o the total criterion value Q( NMS ). 6. Processing o the inal analysis and decision making on selection o the system to integrate (20). Figure 6 (c) illustrates the low charts or general method. 57

(a) (b) (c) Figure 6. Flowchart or the sub-methods: (a) or beneits ater integrating, (b) or completion level, (c) the general method 5. ANALYSIS OF THE SUGGESTED SOLUTIONS The backbone o the proposed method is based on the act that this method and technique use expert assessment and express evaluation o the complex technical and cost criteria. They are oriented on the evaluation o total risk and beneits when there is a need o integration o a new inormation system to the existing inrastructure o a computer network. The proposed method and the concept should be compared to CBA analysis [1, 2], or other methods o analysis, that are used or analysis o dierent inormation systems [10, 14, 15]. First o all, the proposed concept is a simpliication o the existing one. There are three proposed groups o criteria. These are the main eatures o the proposed method. The assessment methodology is presented or each group o the criteria. In opposite to traditional methods [10, 7, 14, 15], it is proposed to use expert estimates. In this case, head o the IT department could make the evaluation process. There is no need to do detailed cost analysis. The proposed method is simple, lexible and allows perorming the required analysis and justiication in a short time. The number o criteria parameters in the method is small. The criteria take into account the basic 58

network management problems. The results o the analysis are clear or managers and the network administrator. The proposed method is used or ast pre-assessment. TCO methods can be used as a supplement to make more extended evaluation. One more additional implementation o this approach is creating the dierent IT e-spaces [13]. This eature helps to manage inrastructure solutions. The main eatures o the suggested solution are the ollowing. Firstly, it oers a symbolic language that is understandable by all o oice managers. This approach helps to orm the development strategy o the computer network inrastructure, as well as to coordinate the strategy according to the company's inancial plans and policy o inancial management. The common terms and categories or proposed conceptions are shown in Figure 7. Figure 7. Schema o the methodology Secondly, it oers a simple method o dierent solutions comparative assessment. This allows the network administrator to make a proper decision quickly or the required system selection. This situation is very typical in real lie. Thirdly, the proposed solution is ocused on the integration o NMS, although it can be applied to the integration o any complex inormation system. In this case, however, it is necessary to review a set o criteria. Why should we choose the estimation scale [0, 2] instead o [0.1]? This is the choice o the authors. We can airm that values {0, 1, 2} are more suitable to interpret than ractional values. They are integer, but in general, it would be possible to use a scale [0, 1]. All o the arguments would stay in orce. For the inal criterion, we should select a range [0, 4]. I we select a range [0, 2], we get the values close to 0.5. In this case, it is diicult to determine and interpret the dierence between various values. How do the changes o value or Er inluence on the values o Q(NMS)? The diagram on the Figure 8 shows these with conditions Er =(2, 1, 0) and Sum=Ee(NMS)+E(NMS)= (1, 2, 3, 4). 59

Q(NMS) 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 ٢ ١ ٠ Er Sum=1 Sum=2 Sum=3 Sum=4 Figure 8. Diagram or Q(NMS) by changing Er For the inal evaluation we can chose the basic values as a classiication o general estimation. These values are given in the Table 1. Table 1. Basic values Value Er+E Er Q (NMS) very bad. 0 2 0 bad 1 2 0,33 middle 2 1 1 good 3 1 1,5 best 4 0 4 The diagram on the Figure 9 shows the general estimation. Q(NMS) 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 very bad. bad middle good best Diagr 1. Figure 9. Diagram or general estimation Q(NMS) Why does it put a limit on the number and type o criterions? The answer is that, all solutions are oered or speciic systems, NMS. 60

6. EXAMPLES 6.1. USING THE CRITERIA OF LEVEL OF PROBABLE RISK Suppose there are three systems: C1 a simple multiunctional system, C2 a multiunctional system, and C3 a ull multiunctional system. The irst system does not make additional load on the network channels; Load ( Net) = 0 X, does not change the existing network administration; process Adm( Net ) = 0, does not change the existing level o security; Sec( Net ) = 0 and has a low cost; Cst( Net ) = 0. The total estimation is equal to Er( NMS ) = 0. The second system is a dierent one. The load o service traic is an average o Load( Net ) = 1, the complexity o administration process is an average o Adm( Net ) = 1, the level o security is an average o Sec( Net ) = 1, and the cost is an average o Cst( Net ) = 1. Thereore, the total estimation is equal to Er ( NMS ) = 1. Third system is a complex multiunctional system, perhaps like management platorms MS System Center or IBM Tivoli. This system has an additional impact on the existing network administration process and complicates it's Load ( Net ) = 2, Adm( Net ) = 2. This can change the security level; Sec( Net ) = 2, and has a high cost; Cst( Net ) = 2. Thus, the inal estimate o the expected risk is equal to Er ( NMS ) = 2. 6.2. USING THE CRITERIA FOR THE EFFECTIVENESS For the irst system, we ind that the beneits will not be signiicant and they will not change the perormance, security aults or accounting criterion. Suppose that they are ocused only on the collection and monitoring processes are related to the state o network coniguration. Then, we have respectively: For the irst system, P = 0, C = 1, F = 0, S = 0, A = 0 as total E (NMS)=0.2., e e e e e e n i= 1 2 i For the second system, P = 0, C = 2, F = 1, S = 0, A = 1, E (NMS)=0.8. e e e e e e For the third system, P = 1, C = 2, F = 1, S = 2, A = 2, E (NMS)=1.6. e e e e e e 6.3. USING THE CRITERIA OF THE COMPLETION LEVEL The irst system does not have high level o unctionality, so the estimations will be: 61

P = 0, C = 1, F = 01, S = 0, A = 0, E (NMS)=0.2. The second system can solve only some tasks or coniguration and accounting management. Thereore, we have: P = 0, C = 1.3, F = 0, S = 0, A = 1 as total E (NMS)=0.46.. A third system can completely solve the problem o coniguration management and other problems that have been solved in part. Thereore, we have P = 1, C = 2, F = 1, S = 1, A = 1 as total E (NMS)=1.2.. All results and the inal estimation Q( NMS ) are shown in Table 2. It can be seen that this technique does not allow directly speciying which system to choose, but it can help justiy any choice. Table 2. Comparative analysis Name o Values or criterion Goal Criterion C1 C2 C3 E (NMS) 0 1 2 Min=0 r E (NMS) 0.2 0.8 1.6 Max=2 e E ( NMS ) 0.2 0.46 1.2 Max=2 Q( NMS ) 0.4 0.63 0.7 Max=4 I we choose the irst system, we spend the least amount o money and make ourselves the least number o problems, but it solves a very small number o tasks. Oten, it cannot give us the possibility o unctional extensions. This is a good solution when the IT-Department is small. I we choose the second system, we restrict ourselves to a partial solution o basic problems. This solution is good or common types o the administrative department: the risk is low and the problem is alleviated. I we choose the third system, then deliberately and consciously we are switching to higher risks and costs, but the eect o the system integration will be much higher. This solution is good or the IT-administration department o a large corporation. In real lie, there are no such particular situations such as Systems 1, 2, and 3. The most values o criteria are very dierent, so in this case, the proposed method can be very useul. The generalized criterion is eective or choosing an integrated system at the comparative analysis. It can be seen that the result is not very dierent in all three cases (Recall that or this method we have the border values Q = 4). This means that in reality the overall eiciency dose not greatly increase because o the problems with integrating or implementing required systems and because o the incompleteness o inal solutions. 62

7. DISCUSSION AND FUTURE WORK The main advantage o the proposed method is the possibility o using it to make decisions at the real-time operational mode. It uses only the resources o experts. One question is valid is there a risk o errors by experts? Yes, o course, the error is allowed, but the criteria are chosen in such way that the error be minimized. It is expected to develop the proposed method to optimize the inrastructure o computer networks and extend these solutions or cloud systems. 8. CONCLUSION The questions o express choice when integrating a new inormation system on the class o Network Management Systems was covered in this article. All solutions are valid or a complex computer network. A review o basic methods o IT systems cost analysis is discussed. The main results include the presentation o the concept and method o express assessment o new inormation system selection, which is going to be integrated to the existing one. The concept describes a general ramework to solve the problem, the method and how to solve it. The results are scientiic and novel. General description o the scientiic results can be presented in the ollowing way. The simpliied method or evaluation o the appropriateness o the selection algorithm, based on the expert express risk assessment o the system integration, was urther developed. The level o the completion and success o the problem solution is accelerated as well. This method can help reduce the inancial costs and decrease inancial risk in integrating new systems. The practical signiicance o the results is that they can be successully used to solve problems o computer networks inrastructure management and in the assessment o dierent kinds o inormation system selection, not just the NMS ones. This reduces the investment risk and causes a reduction o inancial costs in general. ACKNOWLEDGMENT Special Acknowledgement to Al Zaytoonah University o Jordan, or supporting this research paper.all solutions were obtained according to the research project in the Network Management laboratory, Control Inormation System Department, Kharkov National University, Kharkov, Ukraine. 63

REFERENCES [1] Mani Subramanian, (2012). Network Management: Principles and Practices (2nd Edition. Prentice Hall; 726 p [2] Aleksander Clemm, (2007). Network Management Foundamentals. CISCO Press, 510 p. [3] Microsot Operations Framework (MOF) (2008). MOF Executive Overview version 4.0\ Microsot Accelerators, 32 p. Web: http://microsot.com/ technet/solution. [4] Manar Jammala, Taranpreet Singha, Abdallah Shamia, Rasool Asalb,, Yiming Lic, (2014), Sotware deined networking: State o the art and research challenges. Computer Networks Volume 72, 29 October 2014, Pages 74 98. [5] Shai i Muhammad Abdulhamid. Resource scheduling or inrastructure as a service (IaaS) in cloud computing: Challenges and opportunities, (2016), Journal o Network and Computer Applications 68 (2016) 173 200. [6] Scott Farrow & Richard O. Zerbe, Jr. (2013) Principles and Standards or Beneit-Cost Analysis. Edward Elgar Pub. [7] Per-Olov Johansson, Bengt Kriström, (2016) Cost-Beneit Analysis or Project Appraisal. Cambridge University Press. [8] Lambert M. Surhone, Miriam T. Timpledon, and Susan F. Marseken, (2010) 'Procurement: Total Cost o Ownership, Scarcity, Cost-Beneit Analysis",Beta script Publishers 132 p. [9] Supercharging Return on Investment with Rapid Application Development Tools, sponsored by FileMaker, Inc., Apr. 2010, 5 р. Web://www.bitpipe.com/tlist/ [10] Caesar Wu, Raj Kumar Buyya, (2015) Cloud Data Centers and Cost Modeling, Morgan Kaumann,1st ed 848p. [11] Matthew D. Adler, Eric A. Posner (2006) New Foundations o Cost-Beneit Analysis, Harvard University Press 256p. [12] Paul M. Duvall, Steve Matyas, Andrew Glover (2007) Continuous Integration: Improving Sotware Quality and Reducing Risk. Addison-Wesley Proessional, 336 p. [13] Sayenko V.; Al Rawajbeh M.; Levikin V., (2005)- The E-space and Models o the Services, ineer Conerence or Engineering Education and Research iceer-2005. Taipei, Taiwan, CD Proceedings o articles. Session paper N T27, 2005, p. 1-8. [14] E.A. Capuano.(Sep 2005) A TCO Analysis o Cluster Servers Based on TPC-H Benchmarks cluster, 2005 IEEE International Conerence on Cluster Computing, pp.1. [15] Luigi Lavazza. (Aug 2007) Beyond Total Cost o Ownership: Applying Balanced Scorecards to Open- Source Sotware, International Conerence on Sotware Engineering Advances (ICSEA 2007), Aug. 2007, pp. 34-40. [16] ITIL service lie cycle publication suite (2011), 2nd ed 211 p. Web: (http://www.itil.org.uk/ ). [17] Francisco Macia-Perez, Iren Lorenzo-Fonseca, Jose Vicente, Berna-Martinez (april 2014). A ormal ramework or modelling complex network management systems. Journal o network and computer application vol 40,april 2014.pp 255 269. [18] Petra Marešováa&, Kateřina Půžováb (2014) "Application o Cost Beneit Analysis method in Cloud Computing in the Czech Republic ".2nd World Conerence On Business, Economics And Management- WCBEM 2013.pp 674-678. [19] Daniel Ellström, Jakob Rehme, Maria Björklund, Håkan Aronsson,( July 2012)" Logistics Cost Management Models and Their Usability or Purchasing", Journal o Modern Accounting and Auditing 2012, Vol. 8, No. 7, pp1066-1073 64

AUTHORS Dr. Mohammad Al Rawajbeh is an associate proessor at the Computer Network Department, Faculty o Science and Inormation Technology, AL Zaytoonah University o Jordan. Received the High Diploma Degree in Computer Technology rom Yarmouk University, Jordan in 1991. Received B.SC and Master Degree in computers and networks rom Kharkov National Polytechnic University, Ukraine rom 1992-1996. Obtained his Ph.D. degree in Computer Network Management rom Kharkov National University o Radio-Electronic, Ukraine. His research interests are in the ields o network management, perormance evaluation, and internet technology. He has more than 18 published articles in International Journals and Conerences. Pro. Vladimir I. Sayenko. Proessor o Inormation Control Systems department in Kharkov National University o Radio Electronics, European Engineering Educator IGIP (ING-PAED), Head o Network Management Laboratory. In 1979 graduated with honours rom the Control Systems Department, Kharkov Institute o Radio Electronics (specialty Automated Control Systems), then worked as engineer, junior scientiic associate, senior scientiic associate.in 1987 deended the candidate dissertation (PhD) on the topic: Identiication o Objects o Technical Systems with the Use o Patch Functions". 1995 Received an associate proessor certiicate. Since 2001 has been working as a proessor, and at 2001 get the title European Engineer-Educator (ING-PAED IGIP). Supervised many postgraduate dissertations. His research interests in Network Management, Computer Network Optimization, Cloud Systems, and Internet o Things. He has more than 80 published papers. Dr. Mohammed I. Muhairat received the M.Sc. degree in Computer Engineering rom Kharkov State Technical University o Radio Electronics, Ukraine in 1997, and the Ph.D. degree in Inormation Technology rom Kharkov National University o Radio Electronics, Ukraine in 2002. Currently, he is Associate Proessor o Sotware Engineering Department and Deputy Dean o Scientiic Research and Higher Studies Deanship at Al Zaytoonah University o Jordan. His research interests in Sotware Engineering ield, such as, Requirements Speciication, Sotware Architecture, Sotware Development Process, Reverse Engineering, and Formal Methods. He has more than 20 published articles in International Journals and Conerences. 65