Manufacturing Execution Systems (MES) are a bridge between the enterprise solutions and factory shop floors. MES allows the performance of contemporary factories by closer integration of the needs of enterprise stakeholders with the actual manufacturing hardware and software components. Considering different nature of the factories the application of MES has significant difference across industry. Such constrain requires a sophisticated solutions to contain the generality of application. Most of currently available solutions are proprietary and tightly coupled to the ecosystem of devices and enterprise resource planning systems. In current paper the architecture for the open, knowledge-driven MES is discussed. Authors argue that such MES will provide smart extensibility based on base-of-breed approach and hence will improve the quality of MES application, while reducing the introduction costs and system downtime due to reconfiguration.