Increases in scale, complexity, dependency and security for networks have motivated increased automation of activities such as network monitoring. We have employed technology derived from active networking research to develop a series of network monitoring systems, but unlike most previous work, made application needs the priority over infrastructure properties. This choice has produced the following results: (1) the techniques for general infrastructure are both applicable and portable to specific applications such as network monitoring; (2) tradeoffs can benefit our applications while preserving considerable flexibility; and (3) careful engineering allows applications with open architectures to perform competitively with custom-built static implementations. These results are demonstrated via measurements of the lightweight active measurement environment (LAME), its successor, flexible LAME (FLAME), and their application to monitoring for performance and security.