空气质量监测大数据区间的统计问题

为降低客户端和服务端之间的远程过程调用（RPC）通讯，提高对存储空气质量监测数据的HBase表的区间统计效率，提出了一种基于协处理器的大数据区间统计方法。使用终端协处理器可以将区间统计过程放在服务端运行，通过特定的协议将区间统计所需的参数从客户端传递到服务端，协处理器调用结束后，将结果返回到客户端，客户端对返回的消息进行处理汇总，最终得到区间统计结果。实验证明，使用终端协处理器进行空气质量监测数据区间统计较使用客户端扫描方式至少快一个数量级，极大地提高了统计效率。

Big Data Interval Statistics for Air Quality Monitoring

To reduce the remote procedure call communications between the client and the server， and improve the interval statistical efficiency of the HBase table that stores air quality monitoring data， we proposed a big data interval statistics method based on co-processor. Interval statistics process was put into an endpoint co-processor running on server side， then the required parameters for interval statistics were transmitted from the client to the server through a specific protocol. The response message which was processed and summarized was returned to client when calling co-processor was finished， thus the interval statistics result was finally obtained. The experiments prove that using endpoint co-processor to do interval statistics for air quality monitoring data is at least one order of magnitude faster than using client scan method， so the statistics efficiency is promoted dramatically.