The effectiveness of federated learning (FL) systems relies on extensive client participation. However, the existing incentive mechanisms often overlook the intrinsic coupling between participation and privacy strategies. This oversight leads to mismatches between incentive distributions and clients’ privacy costs and contribution quality, limiting the effectiveness of incentives for enhancing contribution quality and participation scale. To address this, DIM-NEG, a dynamic incentive model based on smart contracts and nested evolutionary games, is proposed. First, we construct a nested evolutionary game framework linking external participation with internal privacy strategies. By utilizing the internal equilibrium payoff as a feedback parameter for the external game, we achieve unified modeling of these coupled decision processes. Second, on the basis of this structure, a functionally decoupled dual-incentive lever mechanism is proposed. The server employs the strategy incentive to guide the internal strategy portfolio and the participation incentive to regulate the participation scale, enabling separate optimization of the strategy composition and overall participation level of the system via hierarchical control. Finally, utilizing blockchain-based smart contracts, we design an automated mechanism that encodes rules on-chain to resolve trust issues associated with centralized servers. A theoretical analysis and simulation results demonstrate that the DIM-NEG model achieves superior global accuracy, training efficiency, and communication cost-effectiveness, while exhibiting strong robustness in non-IID environments. The model adequately motivates users to participate in high-quality data sharing tasks and maintains system stability, thereby maximizing the overall effectiveness of the federated learning system.

We revisit the use of crowdsourced throughput measurements to infer and localize congestion on end-to-end paths, with particular focus on points of interconnections between ISPs. We analyze three challenges with this approach. First, accurately identifying which link on the path is congested requires fine-grained network tomography techniques not supported by existing throughput measurement platforms. Coarse-grained network tomography can perform this link identification under certain topological conditions, but we show that these conditions do not always hold on the global Internet. Second, existing measurement platforms provide limited visibility of paths to popular web content sources, and only capture a small fraction of interconnections between ISPs. Third, crowdsourcing measurements inherently risks sample bias: using measurements from volunteers across the Internet leads to uneven distribution of samples across time of day, access link speeds, and home network conditions. Finally, it is not clear how large a drop in throughput to interpret as evidence of congestion. We investigate these challenges in detail, and offer guidelines for deployment of measurement infrastructure, strategies, and technologies that can address empirical gaps in our understanding of congestion on the Internet.