Dynamic Scheduling for Stochastic Edge Cloud Computing Environments using A3C learning and Residual Recurrent Neural Networks
ABSTARCT :
The ubiquitous adoption of Internet-of-Things (IoT) based applications has resulted in the emergence of the Fog computing paradigm, which allows seamlessly harnessing both mobile-edge and cloud resources. Efficient scheduling of application tasks in such environments is challenging due to constrained resource capabilities, mobility factors in IoT, resource heterogeneity, network hierarchy, and stochastic behaviors. Existing heuristics and Reinforcement Learning based approaches lack generalize ability and quick adaptability, thus failing to tackle this problem optimally. They are also unable to utilize the temporal workload patterns and are suitable only for centralized setups. However, Asynchronous-Advantage-Actor-Critic (A3C) learning is known to quickly adapt to dynamic scenarios with less data and Residual Recurrent Neural Network (R2N2) to quickly update model parameters. Thus, we propose an A3C based real-time scheduler for stochastic Edge-Cloud environments allowing decentralized learning, concurrently across multiple agents. We use the R2N2 architecture to capture a large number of host and task parameters together with temporal patterns to provide efficient scheduling decisions. The proposed model is adaptive and able to tune different hyper-parameters based on the application requirements. We explicate our choice of hyper parameters through sensitivity analysis. The experiments conducted on Real world data set show a significant improvement in terms of energy consumption, response time, Service Level Agreement and running cost by 14.4%, 7.74%, 31.9%, and 4.64%, respectively when compared to the state-of-the-art algorithms.
EXISTING SYSTEM :
Task scheduling of complex systems having hierarchical topology is analogous to distribution of jobs or tasks of an application to a group of processors with heterogeneous processing capabilities for fulfilling the optimization goal of minimization of completion time. Therefore, a task graph and a process graph are feed as the inputs of task scheduling. The output is a schedule representing the assignment of tasks to processor.
DISADVANTAGE :
It still faces many challenges due to the limited processing capacity of the local computing resources.
The local computing resources require a lot of finance and human forces.
PROPOSED SYSTEM :
In this paper, we present a novel architecture, which is a collaboration of the computing resources on cloud provider side and the local computing resources (thick clients)on client side. In addition, the main factor of this framework is the dynamic genetic task scheduling to globally minimize the completion time in cloud service, while taking into account network condition and cloud cost paid by customers. Our simulation and comparison with other scheduling approaches show that the proposal produces a reasonable performance together with a noteworthy cost saving for cloud customers.
ADVANTAGE :
These modifications can generate to an actual multi face testate during genetic operators.
Thus, our way is to disregard the time frame through genetic manipulation and allot time slot to each assignment so as to achieve a reasonable schedule well ahead
|