Green clouds optimally use energy resources in large-scale distributed computing environments. Large scale industries such as smart grids are adopting green cloud paradigm to optimize energy needs and to maximize lifespan of smart devices such as smart meters. Both, energy consumption and lifespan of smart meters are critical factors in smart grid applications where performance of these factors decreases with each cycle of grid operation such as record reading and dispatching to the edge nodes. Also, considering large-scale infrastructure of smart grid, replacing out-of-energy and faulty meters is not an economical solution. Therefore, to optimize the energy consumption and lifespan of smart meters, we present a knowledge-based usage strategy for smart meters in this paper. Our proposed scheme is novel and generates custom graph of smart meter tuple datasets and fetches the frequency of lifespan and energy consumption factors. Due to very large-scale dataset graphs, the said factors are fine-grained through R3F filter over modified Hungarian algorithm for smart grid repository. After receiving the exact status of usage, the grid places smart meters in logical partitions according to their utilization frequency. The experimental evaluation shows that the proposed approach enhances lifespan frequency of 100 smart meters by 72% and optimizes energy consumption at an overall percentile of 21% in the green cloud-based smart grid.

Employee churn analytics is the process of assessing employee turnover rate and predicting churners in a corporate company. Due to the rapid requirement of experts in the industries, an employee may switch workplaces, and the company then has to look for a substitute with the training to deal with the tasks. This has become a bottleneck and the corporate sector suffers with additional cost overheads to restore the work routine in the organization. In order to solve this issue in a timely manner, we identify several ML techniques that examine an employee’s record and assess factors in generalized ways to assess whether the resource will remain to continue working or may leave the workplace with the passage of time. However, sensor-based information processing is not much explored in the corporate sector. This paper presents an IoT-enabled predictive strategy to evaluate employee churn count and discusses the factors to decrease this issue in the organizations. For this, we use filter-based methods to analyze features and perform classification to identify firm future churners. The performance evaluation shows that the proposed technique efficiently identifies the future churners with 98% accuracy in the IoT-enabled corporate sector organizations.

Big data analytics has simplified the processing complexity of extremely large data sets through ecosystems, such as Hadoop, MapR, and Cloudera. Apache Hadoop is an open-source ecosystem that manages large data sets in a distributed environment. MapReduce is a programming model that processes massive amount of unstructured data sets over Hadoop cluster. Recently, Hadoop enhances its homogeneous storage function to heterogeneous storage and stores data sets into multiple storage media, i.e., SSD, RAM, and DISK. This development increases the performance of data block placement strategy and allows a client to store large data sets into multiple storage media efficiently than homogeneous storage. However, this evolution increases the consumption of computing capacity and memory usage over MapReduce job scheduling. The scheduler processes MapReduce job into homogeneous container having configuration of CPU, memory, DISK volume, and network I/O, and accesses, processes, and stores data sets over heterogeneous storage media. This produces a processing latency of locating and pairing source data set to MapReduce tasks and results an abnormal high consumption of computing capacity and memory usage in a Datanode. Similarly, when scheduler assigns MapReduce jobs to multiple Datanodes, the same processing latency can severely affect the performance of whole cluster. In this paper, we present Storage-Tag-Aware Scheduler (STAS) that reduces processing latency by scheduling MapReduce jobs into heterogeneous storage containers, i.e., SSD, DISK, and RAM container. STAS endorses job with a tag of storage media, such as JobSSD, JobDISK, and JobRAM and parses them into heterogeneous shared-queues, which assign processing configuration to enlist jobs. STAS manager then schedules shared-queue jobs into heterogeneous MapReduce containers and generates an output into storage media of the cluster. The experimental evaluation shows that STAS optimizes the consumption of computing capacity and memory usage efficiently than available schedulers in a Hadoop cluster.

A software product line is used for the development of a family of products utilizing the reusability of existing resources with low costs and time to market. Feature Model (FM) is used extensively to manage the common and variable features of a family of products, such as Internet of Things (IoT) applications. In the literature, the binary pattern for nested cardinality constraints (BPNCC) approach has been proposed to compute all possible combinations of development features for IoT applications without violating any relationship constraints. Relationship constraints are a predefined set of rules for the selection of features from an FM. Due to high probability of relationship constraints violations, obtaining optimum features combinations from large IoT-based FMs are a challenging task. Therefore, in order to obtain optimum solutions, in this paper, we have proposed multi-objective optimum-BPNCC that consists of three independent paths (first, second, and third). Furthermore, we applied heuristics on these paths and found that the first path is infeasible due to space and execution time complexity. The second path reduces the space complexity; however, time complexity increases due to the increasing group of features. Among these paths, the performance of the third path is best as it removes optional features that are not required for optimization. In experiments, we calculated the outcomes of all three paths that show the significant improvement of optimum solution without constraint violation occurrence. We theoretically prove that this paper is better than previously proposed optimization algorithms, such as a non-dominated sorting genetic algorithm and an indicator-based evolutionary algorithm.