Fog computing is a recent research trend to bring cloud computing services to network edges. EDCs are deployed to decrease the latency and network congestion by processing data streams and user requests in near real time. EDC deploy- ment is distributed in nature and positioned between cloud data centers and data sources. Load balancing is the process of redistributing the work load among EDCs to improve both resource utilization and job response time. Load balanc- ing also avoids a situation where some EDCs are heavily loaded while others are in idle state or doing little data processing. In such scenarios, load balancing between the EDCs plays a vital role for user response and real-time event detec- tion. As the EDCs are deployed in an unattended environment, secure authentication of EDCs is an important issue to address before performing load balancing. This article proposes a novel load balancing technique to authenticate the EDCs and find less loaded EDCs for task allocation. The proposed load balancing technique is more effi- cient than other existing approaches in finding less loaded EDCs for task allocation. The pro- posed approach not only improves efficiency of load balancing; it also strengthens the security by authenticating the destination EDCs.
Based on the fog computing architecture, all the data are stored and processed at the cloud, where EDCs work as the intermediate data cen- ters to reduce the latency of user requests. Cloud is always deployed in the secure environment, so we have considered cloud to initiate the authen- tication process. Cloud initiates the process to assign initial ID (E i ) associated with the key (K i ) and shared key (K c ) for the individual EDCs during the EDCs� deployment. EDCs use trusted modules (e.g., Trusted Platform Module, TPM) to store the secret information from the cloud and the rekeying process . After initialization of the EDCs, each individual EDC starts to authen- ticate the EDCs in the region. This helps in the future to avoid malicious EDCs participating in load balancing.
This article follows the Breadth First Search (BFS) method to design the proposed load balancing technique. We have used two parameters, m and n, to maintain the load of all the EDCs, where m is the current load and n is the maximum capacity to process the tasks. In order to compute the cur- rent load states, we use a parameter p, where p = m/n. Individual EDCs get load balancing requests from other EDCs to process their tasks. If EDC-I is overloaded, EDC-I broadcasts a con- trol packet by sending requests to other EDCs in the region with its own ID and the received load information, that is, (E i , L i ). The ID of EDC-I is defined as E i , whereas the received load infor- mation is defined as L i . The neighbor EDC (named as EDC-J) checks the received ID and compares it with its own database. In the case of a match, EDC-J then looks for the load information from the control packets; otherwise, it ignores the control packet to avoid a denial of service (DoS) attack.
With the great advancements in computing environment and the availability of EDC services in fog computing, the problem of load balancing of EDCs has gained great attention and impor- tance. There are numerous research works that have been conducted to solve the load balanc- ing problem. However, none of them adequately address the EDC authentication issue.
As EDCs are deployed in the network edges in an unattend- ed scenario, authentication of EDCs has become a key factor before load balancing. All the EDCs are deployed in a distributed environment, so load balancing should work in a distributed sce- nario. Load balancing in distributed environments are divided into two main approaches: static load balancing and dynamic load balancing
� The proposed approach presents an adap- tive EDC authentication technique with the help of a centralized cloud data center. This authentication is initiated by the cloud and then all EDCs authenticate each other by fol- lowing cloud credentials. � The proposed approach brings a sustainable and dynamic load balancing technique by considering the load of the destination EDCs. This load information is shared during the authentication process, so individual EDCs do not need additional communication to get the load information from others.
� Finally, the proposed approach combines both the authentication and load balanc- ing techniques to apply in the EDCs. The proposed approach also evaluates the per- formance by validating the efficiency and scalability.
WSharing information with unknown OSN user is a risk which leads being a target by the attackers. In order to address this issue, we studied the existing FRA methods and identified their limitations. Based on those, we proposed a reliable method for decision making of accepting friend request on OSNs. More in detail, we designed an automated filtering
algorithm that estimated a matching factor, which utilized for making reliable decision on accepting a friend request. It is evident from the findings that the proposed method is more reliable compared to the existing FRA methods, which validated the hypothesis of this study.