چكيده لاتين :
Introduction The importance of data center site selection according to different criteria and the protection of data and equipmenthas causedthe proximity to the business offices getting less important and there will be an opportunity for other cities to host the data centers.Thus, each organization when establishing a data center, must evaluate significantfactors in locating a national data center. In general, important factors in choosing a data center location include weather conditions, natural disasters such as floods and earthquakes, un-natural disasters such as war and terrorism, easy access to critical infrastructures such as communications, electrical power,and transportation and qualified people to work in the data center. Thenatural disasters in the developing countries have more impactsrather than developed countries. Since there is no specific study on the impact of natural disasters to locatingdata centers in Iran, in this article,we focus on the natural disasters criteria to locate a data center in Iran.Thisstudy attempts to evaluatetheimpact of this criteria on national data center and offerssome locations for them in Iran. 2. Material anf Methods In this work, disastersare divided into five main categoriesincludingearthquakes, drought, storms, floods, and volcano, and then, they are evaluated separatelywith respect to Iran's climatic conditions. This study is applicable and prepared based on content analysis and statistical analysis. In this study, the effective criteriaon data centersite location in terms of natural disastersis investigatedto find the most desirable province to establish a data center in Iran.Then, provinces are ranked, respectively,as the most favorable to host the data center, according to specific criteria. Animpact factor or weight is added to each criteria using analytic hierarchy processbased on expert’ssurvey, which shows the importance of it from the IT standpoint.Finally, by obtaining the weighted averageand the use of majority voting, the best provinces to host anational data centerare presented respectively. 3. results and Discussion The operation of a data center can easily be compromised by natural disasters. Natural disasters such as earthquake, volcano, storm, flood and drought are always a threat to a data center.It is commonly assumed thatthe probability of occurrence of these events is extremely low and their impact is not much if they happen.But generally, the location of a data center should be somewhere that is potentially far from these risks. Hence, for the first time in Iran, this paper investigates the effect of natural hazards criteria on locating options for hosting a national data center in the whole geography of Iran. Based on our investigation, a computer software is developed that uses these natural hazards criteria, as well as many other factors in mind, to automatically rank all provinces in Iran and their geographical coordinateswithin provinces that are suitable for data center location. Natural hazards criteria used include earthquake, drought, storm, flood, and volcano. An inference engine in the computer software takes in weights of natural hazards criteria that obtained from experts’ judgments and uses a hierarchical analytics processing, and then by a major voting mechanism, outputs its ranked recommended geographical areas. 4. Conclusion In this paper, by recognizingdata center site selection criteria, the natural disasters criteria are measured and their impact on a national data center is obtained with the help of experts. Finally, usingan analytic hierarchy process, a software is developed that receives weightsand ranks cities in terms of suitability for the establishment of a data center. The proposed ranking is just based on natural disasters. This is applicable for some data centers that do not have the security aspects, but depending on the type of data center operation, other criteria should also be considered.For example, if the aspect of national security of a data center is concerned, war and terrorism criteria can improve the rank of center townscompared toborder towns. To rank cities in this paper, a fixed distance is usedamong scores (a unit distance between 1 and 30, respectively). As a future work in this research a proper distance amongscores of cities for the final ranking can be expressed, and scores of cities are calculated according to their distance of criteria that was calculated individually for all cities to achieve higher accuracy in addition toa better illustration of the impact of their weights. Criteria measurementin this article is made by a province that is not a proper granularity to set up a data center. As a future work,we are going to measure these criteria for cities instead ofprovincesto reach to more accuracy
