SELECTING NETWORK MONITORING SYSTEM SOFTWARE WITH ANALYTICAL HIERARCHY PROCESS METHOD

Nowadays there are lots of Network Monitoring System (NMS) software that are reliable and easy to use, such as CiscoWork, HP-OpenView, and IBM-Tivoli. It is just that the software is quite expensive because it has classified as a commercial product. Fortunately, the availability of NMS products is not limited to commercials, but also many other alternatives, namely products that classified as Free and Open Source Software (FOSS). Unfortunately, most of these FOSS products are not only difficult to implement, both in terms of installation and configuration, but also limitations in the number of nodes and types of monitored, including network devices, servers, and applications. The method of selecting FOSS-based NMS software based on the Oyku Alanbay and ISO 7498-4 research methods through the Analytical Hierarchy Process (AHP) approach using the Expert Choice tool. Among GroundWork, Hyperic, Nagios, OpenNMS, and Zenoss, based on the results of this study put Nagios as a fair reliable software for monitoring networks, servers, and applications.


I. INTRODUCTION
In today's era of global competition, computer networks are the backbone of information systems, which can measure whether a company is competitive or not. The vital of this computer network, so the failure will result in loss of productivity and financial losses. Therefore, companies that want to remain competitive want a computer network infrastructure that is reliable, and can work optimally with the lowest possible failure rate.
Computer networks are complex. The complexity is not only limited to the number of devices but also the variety of systems and technologies that are in it, as well as the wide coverage of network areas including Local Area Network (LAN), Wireless Local Area Network (WLAN), Wide Area Network (WAN), Metropolitan Area Network (MAN) and Campus Area Network (CAN) [1].
Even reliable infrastructure cannot be used as a guarantee that the network function can work optimally without the support of good management. Ourselves will never be able to observe with the naked eye and then report what happens just like that, especially when there are obstacles in the network. All these needs can be met with a program that calls itself an NMS.
NMS is a software whose main function is monitoring and managing network devices with the aim that reliability and availability are maintained and can be an anomaly or can be terminated early on the network. Current there are lots of software that is reliable and easy to use, such as, CiscoWork, HP Open View, and IBM Tivoli NetView. But the software is quite expensive because it is classified as a commercial product. Fortunately, the availability of NMS products is not limited to commercial (propriety), but also many other alternatives, namely open-source products [2].
Although there are many choices, it's not necessarily no problem. New problems will arise when we want to choose which products are reliable and by the needs of our network infrastructure. Reliability here is at least the same as licensed products (commercial). Because there is no means if the product is is free but not reliable. If this happens, in the end the cost must be greater than the price of the commercial product itself.
The next problem, most open source products are not only difficult to carry out, both in terms of installation and configuration, but also limitations in the number of nodes and types that are monitored, including monitoring network devices, servers, and applications. This choice is more complex when the software is to use on a large network (enterprise network).
Network Monitoring System (NMS is software whose main function is to watch network devices with the aim that the reliability and availability are maintained and can be an anomaly or can be terminated early on the network. A monitoring system is a process for collecting data from various resources. Usually, the data collected in realtime data [4]. Broadly speaking, the stages in a monitoring system are divide into three major processes as shown in the following pictures: a. Process in collecting monitoring data b. Process in monitoring data analysis c. The process of displaying monitoring data Action that occurs between the processes in a monitoring system is in the form of service, which is a process that is constantly running at a certain time interval. The processes that occur in a monitoring system starts by collecting data such as data from network traffic, hardware information, etc. Then the data is analyzed in the data analysis process and finally, the data will be display.
The first study to select NMS software as software to be review in this study based on research conducted by Jack Hughes of Tech Teapot. On November 2, 2007, Jack Hughes posted the results of his research in the form of statistics based on how many users have downloaded open source-based NMS software from sourceforge.net [5].  [6] Fault management means the ability to detect mistakes. This application will watch the device by doing pooling periodically and notifying when the device inaccessible or error.
The main purpose of configuration management is to watch network and system configuration information so that all hardware, software, and configuration versions can be tracked, and all potential problems can be eliminated (or anticipated). Here it is very important to have a standard configuration, so that with just a glance at the network staff will be able to know exactly where this device located, what type of device, what its role is, and so on.
Accounting management is the process of measuring network use so that network users can set up for accounting or billing purposes. Like performance management, the first step is to measure the use of network devices. The measurement results can be used to get network usage patterns for SLA (Service Level Agreement) guarantees.
Performance management is closely related to SLA (Service Level Agreement), which is a written agreement between the service provider and its customers about the promised network performance level. The SLA consists of several parameters agreed by both parties. Values in the SLA must realistic, meaningful, and measurable by both parties. All measurements entered into the SLA, such as packet drops, latency, bandwidth, and so on.
The main purpose of security management is to regulate access to network resources in accordance with agreed rules so that the network cannot be intentionally or unintentionally intruded. A security management subsystem, such as, can watch all users who enter the network, and reject users who do not have the correct access codes [7].
GroundWork Open Source (GWOS) is one of the NMS software that is quite widely used in monitoring network devices, servers, and applications. In accordance with its mottos The Open Enterprise Monitoring Solution This software classified as NMS for enterprise-class network monitoring. [8] GWOS Community Edition provides essential features needed to watch networks, servers, and applications, with the following features: a. Easy  Nagios is an application for monitoring networks, hosts, and services contained on a network. This application can tell users before there is a network problem in the client or enduser. Although this application is design to run on Linux operating systems, in general, this application can run on most * NIX variants equally well [10].
Nagios already has several features, which makes this monitoring application quite powerful. Some features of Nagios include: a. It can be used to watch network services (SMTP, POP3, HTTP, NNTP, PING). b. It can be used to watch the use of host resources (processor load, memory, and hard disk usage, running processes, log files, and so on). c. It has a simple plugin design, which allows users to develop easily their host and service inspection methods. d. It has the ability to define a network host hierarchy, which allows detection and distribution between host computers that are down, and those that are not down. e. It has the ability to define handling events that will run, before a problem occurs on the host computer. f. Having the ability to recognize problems through a web-based display. The support found in the Nagios project is also very good. Various support services ranging from documentation, FAQs, mailing lists, and forums, have been providing properly by the Nagios.
Like Hyperic HQ, OpenNMS based on a java platform (java-based platform). Other similarities, they both released two editions, open-source (free) and enterprise (commercial). Although service area coverage covers network infrastructure, servers, and applications, the difference is that OpenNMS is more focused on monitoring network devices, while Hyperic HQ focuses primarily on application monitoring (software).
The features of OpenNMS include: a. Service polling -Information service on data availability and latency and measurement. b. Data collection -Collection, storage, and reporting of data derived from nodes that are monitor through the protocol, including SNMP, JMX, HTTP, WMI (Windows Management Instrumentation). c. Thresholding -Evaluating polled latency data or collected performance data against configurable thresholds, creating events when these are exceeded or rearmed (Evaluate d. Event management -Setting events both internal and external events, including SNMP traps e. Alarms and automation -Equipped with alarms with centralized automatic control. f. Notifications -Sending messages based on events via email, SMS, and other media [11]. Zenoss (Zenoss Core) is a network application and monitoring system for the enterprise-class, which can simplify the work of network administrators and system administrators, who can effectively manage the status and performance of the network contained in an infrastructure. This application also can automatically search, agent-less data collection, reporting, and real-time information access through AJAX-based web portals.
The purpose of this study was to choose a reliable FOSS-based NMS software, by determining selection criteria and sub-criteria based on the combining of Oyku Alanbay theory [3] with ISO 7498-4. The objective was to select a reliable FOSS-based NMS software to be used as a computer network monitoring tool. The following two research questions were then proposed for guiding the implementation of the study. RQ1. : Does the software selection method based on Oyku Alanbay Theory and ISO 7498-4 will produce reliable NMS software? RQ2. : Which NMS software is more reliable based on the research method above? This paper divided into five sections. The first one elucidates the research programs of the study, then followed by the literature review, research method, results and discussions, and the conclusion parts in the last section.

Conceptual Framework
The thinking framework in determining selection strategy of FOSS-based NMS software that is reliable and can be implemented properly as a monitoring tool for networks, servers, and computer applications as illustrated in the mindset framework as follows: where the criteria are prepared using the AHP approach method and then questionnaire/ questionnaire data from respondents experts obtained and analyzed with Expert Choice 2000. While the determination of alternative software to studied was obtained based on the statistics of Open Source NMS downloads from the sourceforge.net site. Alanbay research method and change and integration with ISO 7498-4. H1: It is suspect FOSS -based NMS software that is reliable and complies with the criteria and sub-criteria for use in monitoring networks, servers and applications are Nagios.

Types of Research
Judging from the type of information that managed, this type of research is a Quantitative Study where the researcher tests the hypothesis with statistical techniques. The statistical data obtained from the questionnaire using the analytical Hierarchy Process (AHP) approach and then tested using the Expert Choice 2000 tool or software.
Meanwhile, if viewed from the purpose of this study, this type of research is Descriptive Research in which this research intends to offer a description or description of a situation as clearly as possible without any treatment of the object to examine.

Sample Selection Method
In selecting samples, the authors take data from a limited population (limit population) by using purposive sampling, which is taking samples based on certain considerations.
Respondents take a selection of this sample are expert respondents namely network administrators from universities and Internet Service Provider (ISP) companies and academics in the field of networking.
Consideration of sample choice based on the categorization or background of the respondents above, among others, so that evaluation of FOSS-based NMS software is more varied and aim. Besides that, expert respondents in this case are network administrators from colleges/universities, ISPs, and academics in the field of networking because these institutions/companies usually carry out open source products in carrying out their business operations.
The following is a complete list of expert respondents:

Instrumentation
The main instrumentation used in this study was a questionnaire. The questionnaire arranged in the form of questions by referring to the hierarchy made from the criteria and subcriteria based on the Saaty scale of 1 9 using the Pairwise Comparison method.
In priority determining of strategic choice step in determining the FOSS-based NMS software, 20 (twenty) sub-criteria proposed which grouped into 3 (three) main criteria.
The preparation and grouping of these main criteria based on the hierarchy compiled by Alanbay. The sub-criteria modified and adjusted to the subject of this study, where Alanbay's research subject is ERP software that is commercial while the research subject that the author studies is FOSS-based NMS software.
The next change is the addition of subcriteria based on ISO 7498-4 by adding FCAPS (Fault, Configuration, Accountancy, Performance, and Security) elements.
Following the AHP hierarchy in FOSSbased NMS software selection strategy:

III. RESULT AND DISCUSSION
Respondents in the FOSS-based NMS software selection research were all expert respondents totaling 6 (six) respondents. The expert respondent's understanding, in this case, is that all respondents really understand the object under study and have implemented the software in their work.
The six expert respondents each represented 3 (three) groups/categories, namely the network administrator group at the university, ISP (Internet Service Provider), and practitioners/academics. Each group represented by 2 (two) respondents.
The responses of expert respondents for the questionnaire be seen in the results of the merger of respondents as follows:  The geometric calculation of the merger of the respondent's data obtained the value of alternative weights as presented in the following graph: Figure 9. Global priority weight value of alternative priorities based on goals determining FOSS-based NMS software Figure 9 can be seen that the top or highest priority alternative for determining the FOSS-Based NMS Software is Nagios with a weight value of 0.342 or equivalent to 34.2% of the total alternatives set. The results of these alternative weight values are following the hypotheses made by formulation of the problem in the previous chapter. Then the next alternative priority ranking is OpenNMS (weight value 17.8%), Hyperic HQ (weight value 17.7%), GroundWork (weight value 15.7%), and the lowest priority ranking is Zenoss (weight value 14.7%).

Inconsistency Ratio (CR)
Inconsistency ratio or respondent data inconsistency ratio is a parameter use to check whether a pairwise comparison has done consequent or not. The data inconsistency ratio considered good if the CR value is ≤ 0.1.
The following shows the inconsistency ratio values for each comparison matrix.  It concluded that the pairwise comparison provided by expert respondents has a value of inconsistency ratio that is smaller than 0.1 as the largest limit of inconsistency ratio values. Thus, the results of combined geometric calculations of respondents' data are quite consistent.

Conclusions
Some conclusions drawn from the implementation of this thesis are as follows: a. The method and evaluation in determining the choice of FOSS-based NMS software as an effective network, server, and application monitoring tool is the Analytical Hierarchy Process (AHP) approach which determines selection criteria based on the Oyku Alanbay research method with change and integration of ISO 7498-4. b. Nagios is a FOSS-based NMS software that is reliable enough to use as a network monitoring tool, server, and computer application because Nagios meets all the elements of the criteria and sub-criteria specified in this study by getting the highest weight.

Suggestions
Based on the results of the research implications presented in the previous chapter, the authors recommend: a. This research can be used as a reference in determining reliable FOSS-based NMS software for monitoring network devices, servers, and computer applications. b. For software developers, especially NMS software vendors are advised to pay more attention to matters on the user (User Related), both in terms of ease of use and in terms of customization, so users can adjust to existing needs, both from the network, server and application side. c. This research can further be developed, among others by exploring or breaking down in terms of its features, because on a choice of NMS software, often the first thing to look at is the completeness of the features.