An SDN-Based QoS Guaranteed Mechanism for Geospatial Flows

As a significant portion of big data, geospatial big data is experiencing extremely rapid growth, which brings tremendous pressure and great challenges on geospatial data servers. To provide better services for users, it is an important approach to enable the transmission of geospatial data with QoS, which mainly focus on bandwidth. Therefore, based on the software-defined network(SDN), this paper puts forward a bandwidth guaranteed mechanism for geospatial flows for the first time. In our proposed mechanism, taking advantage of the fuzzy analytic hierarchy process(FAHP), a prioritization model for geospatial flows is proposed to calculate the comprehensive priority. With the comprehensive priority, geospatial flows are mapped into slices with different QoS levels. And then, a dynamic routing algorithm and a bandwidth adjustment mechanism are proposed to guarantee the transmission of geospatial flows with high-priority in sliced SDN. Experimental results show that the proposed QoS guaranteed mechanism has the ability to provide differentiated services for geospatial flows, and it can help to alleviate congestion in the geospatial cloud data center as well.

I. Introduction

In recent years, with the rapid development of remote sensing [1], big data [2], Internet of Things(IoT) [3] and Smart Cities [4], geospatial data services are gradually integrating into various areas of social life [5] –[7]. Meanwhile, geospatial data volume is experiencing extremely fast growth as well. According to the statistics, NASA collected 27. 4PB of geospatial data from October 2017 to September 2018, its website requests were up to 3.1 million per minute, and the end user average daily distribution volume was 66. 8TB per day [8]. Due to the tremendous volume and the mass user requests, geospatial data servers are under great burden and pressure, which results in prominent performance problems, such as longer response time, data requests lost, and even denial of services. These problems not only limit the development of geospatial data services severely, but also affect the application of IoT and Smart Cities. To crack the hard nuts, it is an effective approach to provide QoS for geospatial data services.