Enterprises must maintain and improve the reliability of their networks. To do this at reasonable expense, many enterprises choose to outsource the management of their network to third parties, who care for large number of networks for a variety of customers. We obtain and analyze almost a year's worth of failure data for thousands of enterprise networks from such an enterprise service provider. Our data includes device logs, network probes, call center tickets, and incident reports. We study not just the causes and frequency of failures, but also their perceived impact and how failures are resolved. We also characterize what types of failure cause end users to complain. Finally, we show that low-level data may be insufficient to accurately characterize user-visible failures.
|Original language||English (US)|
|Journal||HP Laboratories Technical Report|
|State||Published - May 24 2012|
ASJC Scopus subject areas
- Hardware and Architecture
- Computer Networks and Communications