MODEL OF COMPUTATIONAL RESOURCE METRIC COLLECTION FOR A BLOCKCHAIN-ORIENTED INFORMATION SYSTEM IN A CLOUD ENVIRONMENT

Abstract

The paper addresses the problem of formalizing the process of collecting metrics of computational resources in a blockchain-oriented information system deployed in a cloud environment. It is shown that existing approaches to monitoring cloud and microservice systems focus on telemetry collection and SLA monitoring tools, but do not provide a consistent description of multi-level architectures and heterogeneous metric streams. The aim of the study is to develop a generalized mathematical model that represents the multi-level structure of the system and provides a unified representation of resource metrics for further analysis and optimisation. Within the proposed approach, a formal model of resources of a single cloud environment is constructed, sets of levels and nodes of the blockchain-oriented information system are introduced, and metric vectors for individual nodes are defined. On this basis, a tensor representation of the metric collection process is proposed, which simultaneously takes into account the architectural level, a particular node and the chosen set of basic indicators. This representation allows system telemetry to be treated as a single structured object, simplifies the selection of data subsets and the construction of aggregated state indicators. The adequacy of the model is verified using structural and informational criteria, and its applicability is demonstrated on a simulation example of a three-level architecture. A comparison with the traditional analysis of separate time series confirms a reduction in the effort required to localise bottlenecks and to prepare decisions on scaling and resource reallocation. The proposed model can be integrated with common monitoring stacks and used as a foundation for further research on workload prediction, adaptive resource reservation and configuration management policies in multi-cloud and heterogeneous environments.