Modeling Users' Behavior for Testing the Performance of a Web Map Tile Service

Abstract

Web Map Tile Services (WMTS) are widely used in many fields to quickly and efficiently visualize geospatial data for public use. To ensure that a WMTS can successfully fulfill users' expectations and requirements, the performance of a service must be measured to track latencies and bottlenecks that may downgrade the overall quality of service (QoS). Traditional synthetic workloads used to evaluate WMTS applications are usually generated by repeated static URLs, through randomized requests, or by an access log replay. These three methods do not take request characteristics and users' behaviors into consideration, while access logs are not available for systems still under development. Thus, the evaluation outcomes obtained by these methods cannot represent the real performance of online WMTS applications.In this article a new workload model named HELP (Hotspot/think-timE/Length/Path) is proposed to measure the performance of a prototype WMTS. This model describes how users browse a WMTS map and statistically characterizes complete map navigation behaviors. Then, the HELP model is implemented in HP LoadRunner and used to generate a synthetic workload to evaluate the target WMTS. Experimental results illustrate that the performance representation of the HELP workload is more accurate than that of the other two models, and how a bottleneck in the target system was identified. Additional statistical analysis of request logs and "hotspots" visualizations further validate the proposed HELP workload.

Design and Evaluation of a High-concurrency Web Map Tile Service Framework on a High Performance Cluster

“…Three experiments were carried out to evaluate the performance and scalability of our proposed high-concurrency WMTS with a synthetic HELP workload [22]. This HELP workload can simulate how users browse a WMTS map and statistically characterizes complete map navigation behaviors.…”

Data-Driven Characterization and Modeling of Web Map System Workload

“…Closer to our work, Guan et al [5] proposed a theoretical workload model for WMSs called HELP. This model describes how users of desktop terminals browse an online Web map and statistically characterizes map navigation behavior in relation to the pan, zoom, routing, and search operations.…”

“…WMS users, on the other hand, navigate on the map by zooming and panning in the tile pyramid and, thus, access patterns for WMSs are different from those found in non-geospatial Web systems. As a consequence, conventional Web workload generators can not reproduce WMS workloads accurately [4], [5]. Traditionally, performance evaluation of WMSs is performed with stress testing tools (e.g., [16]), which implement a random tile request model (RTRM).…”

“…None of them, however, focus on WMSs. Yet, WMSs are important systems whose user navigation patterns are very different from those encountered in non-geospatial Web systems [4], [5]. For example, while navigation on the latter occurs mostly through hyperlinks, WMS users navigate by zooming and panning on the map.…”

“…Indeed, we are aware of only one work, the HELP model [5], that statistically characterizes how users of desktop terminals navigate during a WMS map navigation session. However, HELP is not entirely based on real-world data and lacks detailed knowledge of how WMSs are used in practice.…”