无代理技术的设计与实现

Design and Implementation of Agentless Technology

代理的大量使用会导致额外资源消耗增加的问题,同时当今云计算领域无代理技术因多次I/O操作的模拟而存在一定的处理时延。为了解决这些问题,设计了一种高效的无代理技术的实现架构,从减少代理数量和降低处理时延两个方面进行优化改进。研究了用户请求捕获响应、虚拟磁盘多机同时访问、操作隔离和降低处理时延等设计实现难点问题的解决方案。实验表明设计的无代理技术,能够避免传统代理模式导致的额外资源消耗问题;读写操作效率,分别是一般无代理模式的6.3倍和4.1倍,传统代理模式的9.7倍和5.7倍,可以大幅提高处理效率。

The heavy use of agents will lead to the increase of additional resource consumption,and the agentless technology in the cloud computing field has a certain processing delay due to the simulation of multiple I/O operations.In order to solve these problems,an efficient implementation architecture of agentless technology is designed to optimize and improve both the number of agents and the processing delay.The solutions of user request capture response,virtual disk simultaneous access,operation isolation and processing delay reduction are studied.Experiments show that the new agentless technology can avoid the additional resource consumption caused by the traditional agent mode.The read-write operation efficiency is 6.3 times and 4.1 times of the general agentless mode,and 9.7 times and 5.7 times of the traditional agent mode,respectively,which can greatly improve the processing efficiency.