Shuo Shen

A Storage Solution for Massive IoT Data Based on NoSQL

Storage is an important research direction of the data management of the Internet of Things. Massive and heterogeneous data of the Internet of Things brings the storage huge challenges. Based on the analysis of the IoT data characteristics, this paper proposed a storage management solution called IOTMDB based on NoSQL as current storage solutions are not well support storing massive and heterogeneous IoT data. Besides, the storage strategies about expressing and organizing IoT data in a uniform manner were proposed, some evaluations were carried out. Furthermore, we not only just concerned about descripting the data itself, but also cared for sharing of the data, so a data sharing mechanism based on ontology was proposed. Finally, index was studied and a set of query syntaxes to meet the needs of different kinds of IoT queries based NoSQL was given.

A User-Oriented Resource Management Model for the Internet of Things

With the rapid development of Internet of Things (IoT), vast amounts of heterogeneous resources have been connected to the network. How to effectively manage these resources will become a critical issue. This paper first summarizes the current status of the resource management in different scenarios, like WSN, RFID, and RSN. Then from a user perspective, based on the resource name service, we propose a generic resource management model, which is composed of users, devices and resources. Two actual cases are introduced to describe how users can obtain the logical information and control the entity object through this theoretical model. Finally, the analysis shows that this model can meet the needs of ordinary users in compatibility, stability, real-time and flexibility for IoT resources management.

A Server Friendly File Synchronization Mechanism for Cloud Storage

The rapid development of cloud computing has made traditional data centers energy hungry. Thus it is very important to reduce the amount of calculations on the cloud. In this paper we focus on the energy saving in cloud storage service. First we analyze rsync algorithm widely used in the cloud synchronization service and reveal how this algorithm will lead to high CPU load on cloud. Then we design and implement an improved algorithm that moves most calculations out of the cloud. The experiments show that the synchronization related CPU load in server dropped significantly when using our improved method. The improved mechanism can be applied not only in cloud storage service, but also in any centralized synchronization scenario to reduce CPU load on the central nodes.

EPP-IP: A flexible and automatic IP address registration mechanism

The inevitability of IPv6 has come to fore. How to resolve the problem of IP address registration has been a great challenge for network management. The existing solution uses manual registration mechanism, not suitable for dealing with massive IPv6 addresses. Based on the Extensible Provisioning Protocol, we propose an “EPP-IP” mechanism to achieve a flexible and automatic registration management for hierarchical IP addresses. It is compatible with the current address allocation strategy, and can be continued to extend as needed. Complete and consistent registration information will contribute greatly to other network security research.