A node in a system of computers that have a common resource requirement can upload whatever part of the resource it has to other computers in the system even if it has not got the entire resource itself. This policy if followed by all systems will distribute network load over all the connections available instead of concentrating load on client-server connection in addition to improving the performance with respect to time. The same applies to nodes participating in a grid environment; however imagine each node in a grid as a computer and the grid itself as a local network for purpose of clarity and discussion. In this paper a number of techniques that measure the extent to which this concept can be applied and the optimization and efficiency increase each one provides are defined. The techniques are based on fairly simple sharing ideas based on real world occurrences. Our results, obtained while testing in heterogeneous networked systems, indicate that up to 66% (4 out of 6) of optimization features are redundant. Application of this one technique will ensure maximum utilization of all available shared communication media present forming the grid.