Recently, the attention in automotive industry is focused on the potential for improving energy efficiency under the connected environment. For the connected hybrid electric vehicles (HEVs), the efficiency improvement might be further achieved by decision making of engine operating management not only depending on the real-time state of eco-vehicle but also the information of a cloud of vehicles on the traffic flow. This talk presents a framework for management of combustion engines in HEVs which consists of two layers. At the upper layer, a route-dependent decision problem for the engine ignition on/off mode selection is solved as a statistic optimization problem targeted on the statistical data of the vehicle cloud.  With the provided ignition on/off management strategy, at the lower layer, a real-time optimal control algorithm is constructed in the sense of receding horizon optimization and engine transient control, where the V2X-based prediction of the power demand is used as constraint. A benchmark case study with the presented design framework is introduced with demonstration of validation results conducted at a Carmaker-Simulink simulation platform.