In the era of cloud services, there is a strong desire to improve the elasticity and reliability of applications in the cloud. The standard way of achieving these goals is to decouple the life-cycle of important application states from the life-cycle of individual application instances: states, and data in general, are written to and read from cloud databases, deployed close to the application code. The high performance requirements on the application impose strict latency limits on these storage solutions for state access. Cloud database instances are therefore distributed on multiple hosts in order to strive to ensure data locality for all functions. However, the shared nature of certain states, and the inevitable dynamics of the application workload necessarily lead to inter-host data access within the data center (or even across data centers, if the application requires a multi-data center setup). In order to minimize the inter-host communication due to state externalization, we propose an advanced cloud scheduling algorithm that places functions' states across the hosts of a data center. We create a model for the state placement with the aim of minimizing state access latency, and we prove that it is a complex problem. We therefore propose heuristics for fast and efficient placement methods and we evaluate those across realistic scenarios. We show that our approximations are close to the optimal placement, and in large-scale settings the algorithms take only a few minutes to yield good placement results.