A Flexible two-level mechanism in querying and presenting large-scale historical monitoring data

With the increasing scale of data transmitted over Internet, how to effectively query and present mass distributed monitoring data has become a serious problem in distributed monitoring field. Almost all users want to retrieve data as detail as possible; on the other hand, the practical bandwidth and latency limit the data scale in transmission. In order to solve this conflict, we design a flexible two-level mechanism. In our platform with this mechanism, users can customize various retrieving spaces by submitting their clients´ memories. Subsequently, after receiving some request of observing the performance of monitored nodes, the server-end will process the data queried from database according to some rules before responding clients with two-level monitoring data streams (compressed data level and original data level). Through carefully testing, we testify this mechanism is indeed valuable in querying and presenting large-scale monitoring data.