Most of security measures for network computing are provided by service provider itself. However, this kind of methods cannot be trusted radically by the user for lacking the ability to control the resource directly. The computing environment for the user is unable to know what software will be provided. To address this problem, we present a security scheme, named Cleanroom Monitoring System (CMS), to monitor the computing environment provided by server. The CMS provides a mechanism for the user to control the runtime system. Specifically, a secure server is added in CMS as remote trusted base, to provide clean software for the user. A kernel-based monitoring is designed into client OS for illegal procedure detecting, and it is unable to be bypassed. To protect the monitoring module from malicious attacking, the CMS creates the trust chain from the secure server to the client with a lightweight verifiable computing mechanism. This paper presents the implementation and evaluation of the CMS which has gone through rigorous and thorough evaluation of effectiveness and performance. It is currently deployed on the local server and personal computers for a real scenario. The evaluation result shows that the CMS performs effectively with an acceptable overhead.