This document discusses using graphs to model relationships between cloud services from different providers to help with service selection for multi-cloud environments. It presents graphs as a way to capture compatibility, replaceability, and other conditional relationships between services. Centrality metrics, community detection algorithms, and graph pattern matching could then be applied to the modeled graphs to analyze service configurations and identify options that meet requirements for adaptability, quality, and cost. An example graph instantiation is provided to demonstrate compatibility modeling between specific database and application services from Amazon, Google, Microsoft, and other cloud providers.