Predominant resources for execution of any process are computational power and memory. In current Distributed system approaches, when the host system memory is overloaded and a process is executed on a remote system, then all the files, used by the process must also be transferred to the remote system. The host system may already be overloaded and it needs to transfer required files which in turn increase the load on the system. We have designed and implemented an algorithm to overcome this problem. Whenever we execute a new process and the system does not have enough memory for this process to execute, we first mount our file system on the remote system using Network File System (NFS) service, and then change the root of the remote system. Now, the process can be executed at remote system and it can create or modify any file on the host system directly without loading data corresponding to required file on host system's memory. This makes process run faster in remote system than in host system.