مدل خدمات ابری متمرکز کاربر در بخش های دولتی: پیامدهای سیاست ابر سرویس ها

This study examines the acceptance of cloud computing services in government agencies by focusing on the key characteristics that affect behavioral intent. The study expanded upon the technology acceptance model by incorporating contextual factors such as availability, access, security, and reliability. The research model was empirically verified by investigating the perception of users working in public institutions. Modeling results showed that user intentions and behaviors were largely influenced by the perceived features of cloud services. Also these features were found to be the significant antecedents of cloud computing usefulness and ease of use. The findings should guide governments' promotion of cloud public services to increase user awareness by enhancing usability and appeal and ensuring security.

Cloud computing is one of the rising topics in information technology (IT) today. In fact, there is a growing awareness among consumers and enterprises of the ability to extensively position their IT resources through a utility model that is broadly called cloud computing. This term refers to the internet-based design and use of technology in which the cloud represents the internet (Behrend, Wiebe, London, & Johnson, 2011). The cloud includes a substantial technical infrastructure that users do not need to understand at any level of detail; rather, they only need to connect to it to access its resources (Voi, Light, & Rowland, 2011). Technically, cloud computing enables end-users to access computer software and hardware resources over the internet without the need to have any detailed or specific knowledge of the infrastructure used to deliver those resources (Tian, Lin, & Ni, 2010). As the diffusion of cloud computing has increased rapidly, the technology is increasingly being adopted in diverse public sectors (Jaeger, Lin, & Grimes, 2008). Companies are leveraging cloud computing to provide increased standardization of IT infrastructure and to increase efficiency in running technologies. While the private sector is fast moving toward the cloud, governments have begun to assess the potentials that this technology would bring. Cloud computing represents a fundamentally different way for government to architect computing resources, allowing governments to leverage powerful IT infrastructures in a fraction of the time it takes to provision, develop, and deploy similar assets in-house. Around the world, governments of Asia Pacific countries are particularly gearing toward cloud computing. For example, cloud computing solutions are also very prevalent in Korea. The Korean government and private industry together have developed a cloud project named “Next-generation Digital Service in a Cloud Computing Environment” that is aimed at developing and constructing so-called “Cloud storage,” a subset of cloud computing. One example of such a project is the N-Screen Service, which enables data sharing on multiple platforms for mobile phones, tablet computers, televisions, and personal computers. It can be said that the surge in smart device has triggered the expansion of the cloud market in Korea. Along with their phenomenal growth, various cloud computing services have recently suffered from increasing problems of security, privacy, and systematic risk (Lee et al., 2008 and Paquette et al., 2010). In cloud computing, users directly use and operate the software and operating system, and even the basic programming environment and network infrastructure are provided by cloud service providers. Thus, the impact on destruction of the software and hardware cloud resources in cloud computing is worse than those on current internet users. This issue becomes even more critical in the public sector, in which security and reliability are key factors in delivering stable public services (Hamner & Qazi, 2009). Thus, evaluations of user behavior and perceptions of safety are important research topics in cloud computing. For example, concerns about the possible risks of using radio frequency identification (RFID) have increasingly been indicated. Although concerns over cloud computing are not always obvious and direct, risks associated with RFID include the impact of electromagnetic radiation on health and indirect economic consequences such as job elimination through increasing automation (Lee, Lee, & Kong, 2007). The most frequently voiced concern relates to the misuse of data generated by RFID, resulting in an undesirable invasion of individual privacy (Svantesson & Clarke, 2010). Against rising concerns over security and usability, such issues have only been addressed in a few studies (Hossain and Prybutok, 2008, Jaeger et al., 2008 and Thiesse, 2007). In light of such concerns, this study explores the factors influencing user perception of cloud computing to theorize its acceptance model. It also applies the theory of reasoned action (TRA) and modifies the technology acceptance model (TAM) to propose a new model that can be used to examine the acceptance of cloud computing. Technology acceptance research has been relatively limited in its application to the public sector. Therefore, there is a concurrent need to develop and gain empirical support for models of technology acceptance within the public sector, and to examine technology acceptance and utilization issues among public employees to improve the success of implementation in this arena. The new model builds upon the existing TAM by integrating contextual and specific features as primary influencing factors. These factors are driven by underlying perceived beliefs, namely benefit, availability, access, and security as enhancing constructs to predict user motivations to accept cloud computing technologies. The three research issues that guide this study are: RQ1: What perceptions and attitudes exist among public sector officials in Korea that could contribute to, or impede, the acceptance of cloud services? RQ2: What motivational factors contribute to public sector users' intentions and behaviors regarding cloud computing in the public sector?1 RQ3: What policy implications can be derived from cloud computing? This study provides a new framework to identify the antecedents of users' intentions to adopt cloud computing in the public sector. TAM has been criticized for its lack of contextual understanding (Venkatesh & Brown, 2001), a factor that is even more essential in policy implications. It is unclear which specific factors facilitate and/or inhibit user acceptance in cloud computing in the public sector (Loo, Yeow, & Chong, 2009). This study addresses this question by providing structural correlations among the context-specific factors of cloud computing for use in policy provisions. The findings herein should guide governments in promoting cloud computing services as a way to improve government business and public service. One of the most promising cloud computing opportunities for the public sector is the capacity to share information and communication technology (ICT) resources among various agencies simultaneously. One challenge is the deployment of cloud computing in the public sector featuring secure and user-friendly services (Kshetri, 2010 and Zissis and Lekkas, 2011). So far, the difference between public and private has not been researched enough. Key differences existing between private commercial organizations and public sectors may result in different adoption behaviors of technology acceptance in public sectors. Private commercial organizations are ultimately motivated by maximizing profit whereas public sectors are not. In fact, cloud service is designed to provide more cost effective and efficient governmental services to citizens. Further, other differences between private commercial organizations and public sectors such as organization structure, organizational culture, and social norms, may also result in cloud service adoption in public sectors being different from that in private commercial organizations. The findings of the present study offer a set of guidelines that will help governments better understand the development of user-perceived features and how they contribute to usability.

5.1. Structural model Table 1 shows the structural modeling estimate. The overall model fit was satisfactory, with all of the relevant goodness of fit indices being > 0.90. Chi-square statistics showed non-significance for both the measurement and the structural models, indicating that the two models adequately fit the data. The goodness of fit index was 0.95, the adjusted goodness of fit index was 0.91, and the Tucker–Lewis index was 0.91. Similarly, there was no evidence of misfit, with the root mean square error approximation showing a fairly satisfactory level of 0.067, which favorably compares to the suggestions by Joreskog and Sorbom (1996), who suggested that values ≥ 0.06 reflect a close fit. The standardized root mean square residual, at 0.027, was also very good, well below the threshold for a good overall fit. Another positive test statistic was the normed Chi-square value of 1.98, a value that was appropriately well below the benchmark of 3, indicating good overall model performance. Since the model shows satisfactory fit indices, the path coefficients of the structural model can be assessed.5.2. Structural paths and hypothesis tests The hypothesized causal paths were estimated to test the structural relationships, and nine hypotheses were supported while two were rejected. The results are reported and depicted in Table 2. Overall, the results support the proposed model, confirming the antecedents of PU and PEoU in cloud computing. Most paths in the model were statistically significant, highlighting the significant roles of the cloud computing features in determining user PU and PEoU, which then affect intent to use cloud computing (H2 and H3).Although intention is significantly influenced by exogenous variables, the effect of intention on usage behavior was moderate or weak in this model (H1, β = 0.21, t = 2.981, p < 0.05); probably because users want to confirm their behavioral intention using other factors. In other words, while users might have strong intentional will, they may still have lingering doubts that limit their actual behavior. In fact, there might be other factors that lead users to actual usage. Interestingly, actual behavior is significantly influenced by SN (β = 0.52, t = 3.304, p < 0.001). Both endogenous variables, intention to use and behavior, showed significant correlations with their exogenous variables. Thus, the question of why positive intention did not lead to actual behavior was raised. There is a clear gap between intention and actual behavior. Approximately 54% of the variance in the intention to use cloud computing was explained by the variables in the model (R2 = 0.541). The R2 of actual behavior was explained by about 29% using the exogenous model constructs. Because both R2 values were fairly high, it can be reasonably inferred that there is a missing link between the two. This finding implies that while cloud computing users might have a strong intention influenced by PU and PEoU, this intention does not automatically lead to actual behavior. While users might cognitively perceive the excellent features of cloud computing, they may not actually intend to use the system unless the important issues are addressed. These users may want to personally ensure that cloud computing is reliable and dependable. While the model shows a pattern that highlights the importance of both utility and usability, the model underplays the link between intention and behavior compared with those in previous studies employing TAM. A gap between intention and behavioral intention in cloud computing may be inferred. Social influences like SN can play a facilitating role between intention and behavior. This role sheds light on possible advancement in terms of theory and practice. Thus, further tests are needed to uncover possible underlying effects. 5.3. Comparing public vs. private As the findings in the public sector show a significant result, it is worthwhile to compare and contrast with the private sector. With the public sector adoption model in place, this study further collected survey data from the private sector. To compare with the public sector model, data from the private sector collected were the equal number of respondents of the public sector. A marketing firm specialized in data collection, survey design, and survey administration collected data from various private sectors (banking, IT, manufacturing, retailing, and research). Interestingly, the result from private sector adoption shows a similar result from the public sector, yet there were different patterns and importance (Table 3). Just like the public sector case, all twelve hypotheses were supported with sufficient path coefficients and t-value. There were noticeably different patterns between the sectors. For example, access factor shows to be moderately significant to the public sector (* p < 0.05; ** p < 0.01), whereas it shows high significance (both ***p < 0.001). Similarly, while availability shows a moderate significance to the public sector, high significance shows to the private sector. Same thing happens to reliability. It can be inferred that while the factors in the public sector are equally important in the private sector, different matrices of importance exist across the sector. Access, reliability, and availability may be important in operating the private sector, while security and usefulness may be the key criteria in providing for the public sector. In this light of difference, SN may also play a higher role in public than private.