Leveraging Quantum Computing for Optimal Data Allocation in Distributed Systems

We show how to use quantum computing to optimize data allocation in distributed systems.Our problem model considers storage constraints (modeling limitations in storage capacity on each node), k-safety constraints (ensuring that a user-defined number of copies of each data partition is distributed in the system), and performance (calculating overheads due to the need for sending data for remote accesses over the network).Specifically, our approach leverages quantum annealers, providing high-quality solutions to instances of quadratic unconstrained binary optimization (QUBO) problems quickly.We show how to transform data allocation instances into equivalent QUBO instances.Also, we analyze the number of qubits needed to represent the resulting instances.More precisely, we determine the asymptotic number of required qubits as a function of the dimensions of the input problem.Our analysis takes into account different annealer topologies, ranging from fully connected qubits to sparse connection structures.Finally, we outline future work for this project.