Maintaining critical data access latency requirements and ensuring efficient energy consumption on the field devices are two important challenges of Industry 4.0. The traditional, centralized industrial networks which offer primitive data distribution functions, might be incapable of meeting such strict requirements. In this paper, in order to overcome this issue, we focus on a network of resource constraint IoT devices, and exploit the presence of a few more capable Edge nodes that act as distributed local data storing proxies for all IoT devices. We show that, given the proxy locations in the network, the initial energy supplies of the nodes, a pattern of data requests from IoT devices, and the maximum access latency that consumer nodes can tolerate, the problem of finding how to distribute the data on the Edge nodes maximizing the network lifetime is computationally hard. We design an offline centralized heuristic algorithm for identifying which paths in the network the data should follow and on which proxies they should be cached, in order to meet the data access latency constraint and to maximize the network lifetime. We implement the method and evaluate its performance using a testbed of IEEE 802.15.4-enabled network nodes. We demonstrate that the proposed heuristic (i) guarantees data access latency below the given threshold, and (ii) performs well in terms of network lifetime with respect to a theoretically optimal solution.