Modeling and simulation of the redundant array of inexpensive/independent disks storage for internet of things monitoring servers

摘要

Internet of things environment consists of a large number of small devices that have various sensing capabilities with wireless network and computing power. Internet of things nodes send data to a server for analysis. In the current days, column-wise compressed storages are mounted to the big data servers due to its space efficiency and flash drive has become an important storage in building internet data center due to its superior performance and power-economic nature. We will introduce a short overview on sensor networks and will propose a new storage management scheme for sensor databases. The core idea of our scheme is based on a new RAID5 storage model using column-wise segmentation and compression to improve space efficiency without sacrificing input/output (I/O) performance. In addition, we developed a simulation system using CSim application program interface (API). Finally, we show that proposed storage control scheme outperforms the original RAID5 scheme by performance evaluation. The proposed scheme could be used in various industrial applications such as smart factory and self-drive car databases which handle massive amounts of 1 PB data from internet of things sensors. The aim of this study is to provide an effective skill of internet of things storage programming to graduate and undergraduate students in a computer software department. This study assists the students to understand the availability issue of disk storages and the space-efficiency issue of redundant array of inexpensive/independent disks (RAID) storages. This study also provides the open-source-based programming skill, CSim simulation skill and API programming skill, the RAID setup skill, performance test skill, and result analysis skill.