✅ Перевірена відповідь на це питання доступна нижче. Наші рішення, перевірені спільнотою, допомагають краще зрозуміти матеріал.
A team is working on an architecture for distributed computation amongst many thousands of compute nodes in a large cluster. Initial tasks are distributed by a central 'master' node to the compute nodes. All the compute nodes are known before the computation begins. Once an individual computation is completed, the output must be communicated to several other compute nodes. Eventually, when computation reaches an end state, the final results from every compute node must be communicated back to the master. What kind of architecture is suitable here?