A Study on the Use of Runtime Files in Handling Crash Reports in a Large Telecom Company

Abstract

To minimize service downtime and ensure high system availability, Telecom companies must react quickly to failures and crashes of network services and applications. In this study, the focus is on the crash reporting process that normally goes through different levels of service support in a telecom company. To speed up this this repairing or recovering process, service support engineers and product developers rely on the analysis of runtime files (logs, traces, performance metrics, etc.) that are attached to crash reports submitted when an incident occurs. However, there is no clear understanding how these files are used and what their impact on the crash fixing time is. In this paper, we conduct an empirical study at Ericsson to study the use of runtime files in the crash resolution process. We tackle various research questions that revolve around the proportion of runtime files in a selected set of crash reports, the relationship between the severity of crashes and the type of files they contain, and the impact of different file types on the time to fix the crashes. We also study the prediction of the attachment of runtime files to crash reports during the creation of the reports. Our ultimate goal is to figure out how to collect enough information in the crash report system for support engineers and product developers to resolve the failures occurring in telecom network quickly and efficiently.