Resource allocation (RA) is one of the most important stages of distributed query processing in Data Grid systems. Recently, a number of papers that propose different methods for RA were published. To deal with specific characteristics of the data grid systems, such as dynamicity, heterogeneity and large-scale, many studies extend classic methods from distributed and parallel databases domains. Others invite fundamentally different methods based on incentives for autonomous nodes. The present study provides a brief description, qualitative comparison and performance evaluation of the most interesting approaches (extended classic and incentive-based) for RA. Both approaches are promising and appropriate for successful data grid systems.