Аннотация
The article examines the possibility of reducing the time spent on performing tasks in distributed data processing systems, which are necessary for remote access to information resources. For this, the content and volume of annotation data of resources are analyzed; the main methods of scheduling task processing in distributed systems are identified and analyzed: Least Recently Used Node/Server, Resource consumption prediction, Random Selection, Queue metrics based and Elapsed-time prediction; SJF, SPN and SRTF; Time-sharing and Least Attained Service; FIFO, LIFO, RSS. The revealed techniques are analyzed under conditions of incompleteness of information about the resource requirements of distributed systems.