Data-stream Query Optimization Across System Boundaries

这个子项目的工作假设仍然是，分散的数据流处理有其好处，因为它减少了从移动节点到地面节点、在地面节点之间以及从地面节点到中心节点的数据传输。生物学家的要求现在可以用形式语言和机器可处理的形式作为查询。这些查询已经可以计算，但只能在中央节点和人工数据上进行。进一步的数据流将导致新的查询，为生物学家提供额外的信息，并具有分散执行的选项。这使用来自各个节点的状态消息进行了扩展，以允许系统监视。它将用于识别关键系统状态、评估重新配置选项以及启动重新配置操作。后者可能意味着设置几个参数(例如传感器的采样率)，但也意味着将代码从地面节点部署到移动节点(例如用于预聚传感器数据的操作员)。

The working hypothesis of this sub-project remains that decentralized data-stream processing has its benefit, because it reduces data transmission from the mobile nodes to the ground nodes, among the ground nodes, and from the ground nodes to the central node. The requirements of the biologists are now available as queries in formal language and in machine-processible form. These queries can already be evaluated, but only on the central node and on artifical data. Further data streams will lead to new queries with additional information for the biologists and with the option of decentralized execution. This is extended with status messages from the various nodes to allow for system monitoring. It will be used for the recognition of critical system states, for the assessment of reconfiguration options, and for the initiation of reconfiguration actions. The latter can mean setting a few parameters (e.g. the sampling rate of a sensor), but also the deployment of code from the ground nodes to the mobile nodes (e.g. operators for the pre-aggregation of sensor data).