More for Less - Getting More Clients by Broadcasting Less Data

Broadcasting is scalable in terms of served users but not in terms of served data volume. Additionally, waiting time deadlines may be difficult to uphold due to the data clutter, forcing the clients to flee the system. This work proposes a way of selecting subsets of the original data that ensure near-optimal service ratio. The proposed technique relies on the novel data compatibility distance, which is introduced herein. Clustering techniques are then used for defining optimal data subsets. Comparison with related work and brute force-derived solutions yielded superior and near-optimal service ratios in all test cases. Thus, it is demonstrated that a system can attract more clients by using just a small portion of the available data pool.