*2.1. Methodologies Regarding Cloud Based Architecture*

A novel architecture for electric vehicles' charging and discharging is presented using public supply stations [15]. For electric vehicles' charging and discharging, two priority assignment algorithms: (1) random priority attribution and (b) calender priority attribution are developed in the cloud environment. These algorithms are used for monitoring the waiting time of electric vehicles in order to maintain grid stability in peak hours by setting the demand supply graph as a constraint. Smart phone has been considered as a component of the cyberphysical system for dynamic voltage scaling by minimizing the frequency of smart phone which leads to energy minimization [16]. Authors also develop an energy aware dynamic task scheduling algorithm for

minimizing the aggregated energy requirements of the running applications by considering two constraints: (1) time and (2) probability. This work has been restricted to the energy minimization in the smart phone system. It also lacks its applicability in SG energy management domain.

In [17], one new communication model is proposed which is composed from two models: the cloud based Demand Response (DR) model and distributed DR model for optimizing the communication delay. This model suffers from huge cost for the peak demand scenarios. Further, a cost computation model is also proposed for demand side consumers by the optimal use of the cloud resources. Moreover, this model provides an opportunity for the consumers regarding optimized resource availability which helps in cost minimization. They have designed the modified priority list algorithm which is used for optimal distribution of the cloud resources (instances): on-demand instances and reserved instances. It minimizes the total operational cost of the system.

The idea of nanogrids is presented in sustainable buildings for multi-tenant cloud environment in [18]. A game theory technique using the collaboration for power management in SG cyberphysical system is proposed in [19], where payoff function is formulated by investigating each player's information (i.e., transmission and service delay information) through conditional entropy. Moreover, the dynamic workflow management is designed for executing the tasks via virtual cloud platform. Various VMs are utilized in a coalition for running the jobs in an intelligent way for this scenario. In [20], the energy hubs are manoeuvred for storage of the consumers' data in the cloud environment for efficient DSM. Stochastic dynamic programming is used to manage the load in real time environment for reducing the expenses by incorporating the consumers' participation. This system is particularly designed for cloud environment and our system incorporates the fog in order to minimize the latency, PT and cost.

Some reviews about cloud computing technologies are presented in [13] along with their challenges. Some limitations of the existing power systems are also described in this work by the cloud computing technologies, for instance; the infrastructure of the power system is weak and there can be blackouts at any time due to uncertain catastrophes. For maintaining the resiliency, cloud computing facilitates the consumers by providing computing, storage and security features.
