Okta Nurika

Workability review of genetic algorithm approach in networks

In this paper, we surveyed the implementations of genetic algorithm within networks, whether it is computer network, transportation network, and other fields that have networking context. Their feasibilities are discussed along with our suggestions for potential improvements. Genetic algorithm can also be an alternative to other optimization methods/algorithms. In some cases, it even outperforms other methods. However, the choice of genetic algorithm might be influenced by some concerns, such as execution time and problem size. Generally, genetic algorithm process will accomplish according to its parameters sizes. Finally, the success stories prove the applicability, adaptability, and scalability of genetic algorithm, specifically for almost-any network optimization.

Mathematical models for network card simulation and their empirical validations

Network simulations mostly focus on protocol observation, development, and prediction. However, newly developed protocol will take time to become standard and be implemented on network devices. A faster way of network improvement is to optimize existing network card parameters, because they are generic and protocol-independent. In this paper, we propose mathematical models to represent network card parameters, thus their optimal configuration can be predicted, which then will accelerate the reduced network speed caused by unoptimized network hardware. The simulation gaps that we fulfill are the inclusion of transmission size, packet polling based on timer, and packet polling based on watermark or packet capture threshold. The results generated and their validations show the accuracy of our models.

Implementation of Network Cards Optimizations in Hadoop Cluster Data Transmissions

In this paper, the previously invented new methods of network card optimization are applied in a Hadoop cluster, where data transfers occur from the Master to the slave node. The slave nodes network card setting is optimized subjective to the characteristics of the incoming data transmissions, which are indicated by the overall transmission size and packet size. The throughput comparisons between the optimized network card settings and the default setting conclude that the optimized versions always generate higher throughputs. Synchronously, the optimized settings also minimize CPU cycles utilization as they deploy timer-based polling (passive wait mode), in order to process the received data packets. This novel practice within Hadoop cluster may be replicated by other data cluster vendors, thus improving their data transfers throughput and efficiency.