Information system infrastructure planning - threshold setting based on risk analysis I. Golub and B. Radojević * * Croatian Academic and Research Network (CARNet), Zagreb, Croatia Ivana.Golub@CARNet.hr, Branko.Radojevic@CARNet.hr Abstract - Information systems infrastructure planning performed based on the data collected from network monitoring and asset management systems can be developed further to include risk assessment and analysis as an input to the threshold setting methods in the two-dimensional load-age balance graph. The two-dimensional load-age balance graph is used to present generic value for network devices load gathered as measured values through the network managements system, and age - gathered from the asset management system. Threshold lines are defined in such graph to present areas of different interest from the perspective of low-high age, low-high load, or some other selection criteria. This paper explores how the threshold values can be set to reflect risk assessment methodology of the owner/maintainer/user. The position of each networking device in the graph is unique, disregarding the threshold values. However, the total score for the system, determined based on the position of a device in the graph area, depends on the defined threshold. The paper analyzes the impact of the threshold change based on the risk assessment criteria to the total score for a system. The total score can then be used to compare different systems, or to compare the same system in time. I. INTRODUCTION Information systems infrastructure planning has to be based on the information about the current state of the system received from the system itself (e.g. via monitoring system), and information about the plans for further development of services and users needs. It is usually not in the domain of the system maintainers to have information about neither the future services development nor the users growth. The only data that can be gathered and examined are those from the system itself. Risk in a general sense represent "a probability or threat of damage" or "loss or any other negative occurrence" "caused by external or internal vulnerabilities, that may be avoided through preemptive action" [1] or the possibility that something bad or unpleasant will happen [2]. In the sense of an information system, a risk can be seen in the degraded level of service, or as the total unavailability of the service offered to users (customers). Risk assessment has its two dimensions - the probability for an event to happen and the level of danger that the event can cause to the system. In addition, as explained in [3], [4], the risk has its objective and subjective side, and both determine the assessment of a risk in a system. While the "risk as analysis" (as defined in [4] starts from the information gained through measurement systems of the current state of an information system, the "risk as feelings" [4] has its final decision in the determination of the risk levels and categories. The model for the determination of the current stage for an information system is presented in [5] and is re- printed in the Figure 1. System selection (1) Device categorization (2) Usage/Load parameter selection (4) (N) Parameter value determination (5) Age determination (6) Exploration graph Determination (7) Graph values determination (3) Figure 1. Exploration Determination Model (Source [5]) Based on that model, using the data from a monitoring system, and as explained in the [5] it is easy to design two-dimensional load-age graph with the current status of each observed device represented with one dot in the graph. X axis will contain generic value for the load of the system-relevant components (e.g. CPU and memory load and the number of used ports for networking equipment, average CPU and memory load for servers) expressed in percent of the maximum load and represented as the sum of values for each measured parameters. Y-axis contains the age calculated as the current year minus year of the equipment acquisition. Thus created graph represents current state of the observed system. However, in order to be able to interpret the data and use it for the purpose of system development and planning, it is necessary to put this data into the relevant context for the system. II. RISK ASSESSMENT AND THRESHOLD DEFINITION Each information system can be seen differently from different user groups - the system owner's view does not necessarily have to be the same as the view of a user or a view of a service owner or the information system