An Automata-Theoretic Approach to Reasoning About Parameterized Systems and Specifications We introduce {\em generalized register automata} (GRAs) and study their properties and applications in reasoning about systems and specifications over infinite domains. We show that GRAs can capture both {\em VLTL} -- a logic that extends LTL with variables over infinite domains, and {\em abstract systems} -- finite state systems whose atomic propositions are parameterized by variable over infinite domains. VLTL and abstract systems naturally model and specify infinite-state systems in which the source of infinity is the data domain (c.f., range of processes id, context of messages). Thus, GRAs suggest an automata-theoretic approach for reasoning about such systems. We demonstrate the usefulness of the approach by pushing forward the known border of decidability for the model-checking problem in this setting. From a theoretical point of view, GRAs extend register automata and are related to other formalisms for defining languages over infinite alphabets.