Software realization of distributed computing systems is typically achieved through the component-based software development approach. Thus, software executing on such distributed systems consists of many autonomous components which interact with each other to coordinate their activities. The need for such coordination along with requirements such as heterogeneity, scalability, security, and availability, considerably increases the complexity of code in distributed systems. Simplifying distributed component-based software development requires abstractions at various levels that enable separate design and specification of component interactions and the components themselves. This paper depicts a method to specify component interactions involving temporal invocation constraints. Based on the knowledge of component interactions, two types of replaceability criteria are defined and empirically evaluated.