Cloud Application Monitoring for Efficient Network Management in Public Schools
John Pickard, Dale Drummond, Philip Lunsford, Ciprian Popoviciu · 2020
Abstract With the increased movement of administrative, operations, and educational tools from on-premise resources to hosted or Software as a Service (SaaS) offerings, K-12 and post-secondary schools and school administrators are placing increasing demands on the network, the applications being used, and the staff supporting those applications. As the importance of the network’s role in day-to-day education operations increases, new ways to understand how the network itself and the services provided over the network are performing are needed. This paper will present the findings and quantifiable benefits observed by shifting from a traditional network monitoring approach to an application performance or User Experience approach. This novel approach aids school administrators and IT support staff in reducing the time needed to identify, isolate, and resolve performance issues experienced by end users. Distributed network performance monitoring agent platforms are used to collect extensive User Experience focused data for external web resources identified by the school administrators and IT Support staff as being vital to the delivery of their curriculum. Measurements for DNS resolution time, TCP/IP communications response time, and HTTP load times (User Experience) are captured by each monitoring agent, recorded and then analyzed. The monitoring agents, both wired and wireless, are distributed at 36 individual schools: 20 elementary schools, 7 middle schools, and 9 high schools. In addition, monitoring agents are placed at the school district Central Office (CO) and the school district’s Internet Service Provider (ISP). Identical measurements are recorded from each agent at a polling rate ranging from 10 minutes to 1-hour dependent on the importance of the service being monitored. Higher importance services are monitored more frequently than lower importance services. Data analysis is applied to establish a performance baseline for each service, at each agent location, over the same time window. The data collected and presented enables the limited IT staff of the school district to identify performance issues in a timely manner and automatically narrow down the fault domain. The methodologies employed and shared in this study will allow for automatic and proactive notification of the support staff enabling issues to be addressed before being reported by end users.