The application of the multistage interconnection networks (MINs) in systems-on-chip (SoC) and networks-on-chip (NoC) is hottest since year 2002. Nevertheless, nobody used them practically for parallel communication. However, to overcome all the previous problems, a new method is proposed that uses MIN to provide intra-(global) communication among application-specific NoCs in networks-in-package (NiP). For this, four O(n) 2 fault-tolerant parallel algorithms are proposed. It allows different NoCs to communicate in parallel using either fault-tolerant irregular Penta multistage interconnection network (PNN) or fault-tolerant regular Hexa multistage interconnection network (HXN). These two are acting as an interconnects-on-chip (IoC) in NiP. Both IoC use packet switching and wormhole switching to route packets from source NoC to destination NoC. The results are compared in terms of packet losses and wormhole switching which comes out to be better than packet switching. The comparison of IoC on cost and MTTR concluded that the HXN has the higher cost than the PNN, but MTTR values of the HXN are low in comparison to the PNN. This signifies that the ability to tolerate faults and online repairing of the HXN is higher and faster than the PNN.