Interprocess communication mechanisms, one of the effective factors for response time in High performance computing. Existing standards for interprocess communication between a heterogeneous operating systems such as MPI implementation at library level lead to lower performance and Increase response time compared to implementation at the operating system kernel level. We propose an approach to enable the use of IPC in heterogeneous distributed systems such as grid computing. This mechanism uses distributed shared memory for communication between Linux and Windows operating systems. Windows IPC mechanism uses shared memory and remote procedural call. Linux Inter-zprocess communication (IPC) converting Windows remote procedure call (RPC) to Proportional System V shared memory in kernel level. We propose wrapper for the use of our mechanism with distributed application.