Optical transmission has emerged as the most cost-effective technology to implement high-bandwidth based communications and to transmit the huge volume of data with low latency. Fog computing extends cloud computing to improve efficiency and reduces the amount of data to be transferred to the cloud for data processing, analysis, and storage etc. In this paper, a new fog layer among optical elements is proposed that utilizes the resources of the optical network. It uses passive optical network (PON), optical line terminals (OLTs) and optical network units (ONUs) to deliver cloud-based services more effectively with minimum latency. A large number of jobs and limited resources in fog layer lead to the deadlock that affects the Quality of Service (QoS) and reliability in heterogeneous fog and cloud environment. Therefore, Social Network Analysis (SNA) based deadlock manager is proposed with a new concept of Free Resource Fog (FRF) that helps to remove deadlock by collecting available free resources from all running jobs. In order to utilize resources and minimize the response time of the submitted job, a rule-based algorithm is proposed that assigns priorities to the jobs and provides resources accordingly from fog and cloud. In addition, energy consumption and latency measure are presented those reflects the QoS as well as reliability to end users. Gephi is used for the implementation of SNA based deadlock management whereas Cloudsim is used to evaluate the utilization of fog and cloud computing resources using Resource Pool Manager (RPM). Finally, we conclude that optimum resource utilization and latency measures can enable future computing with optical fog systems.
Read full abstract