A Queueing Network Analysis of a Hierarchical Communication Architecture for Advanced Metering Infrastructure

Jin Seek Choi, Sunghwan Lee, Se Joon Chun

Research output: Contribution to journalArticlepeer-review


In the smart grid roadmap, the number of smart meter installations has grown at a large scale from the grid to homes. To accommodate such a large scale of smart meter installations, advanced metering infrastructure (AMI) facilitates a hierarchical communication architecture for bidirectional data flows between a meter data management system (MDMS) and smart meters via data concentrator units (DCUs). This paper evaluates the traffic load, throughput and end-to-end delay performance of meter data collection and processing and their communications in a hierarchical architecture. To evaluate the performance, we model the hierarchical AMI by a tree-structured feedforward open queueing network and approximately analyze the delay and system throughput using a parametric decomposition method. From the analytical results, we find that increasing the number of DCUs affects the levels of traffic and delay performance. However, the system throughput is bounded by the processing complexity of the MDMS, regardless of the number of DCUs. Based on these findings, we conclude that the number of smart meters and DCUs and the collection frequency can be properly arranged when designing a large-scale AMI. The performance results can be utilized for developing advanced metering applications without violating delay constraints.

Original languageEnglish
Article number9454069
Pages (from-to)4318-4326
Number of pages9
JournalIEEE Transactions on Smart Grid
Issue number5
StatePublished - 2021 Sep
Externally publishedYes


  • Advanced metering infrastructure
  • delay
  • performance
  • queueing analysis
  • system throughput
  • traffic load


Dive into the research topics of 'A Queueing Network Analysis of a Hierarchical Communication Architecture for Advanced Metering Infrastructure'. Together they form a unique fingerprint.

Cite this