# MonARCh: an actor based architecture for dynamic linked data monitoring

## TL;DR

MonARCh is a system that monitors changes in linked data sources using SPARQL queries and actor-based architecture for scalability.

## Contribution

MonARCh introduces a novel actor-based architecture for dynamic linked data monitoring with concurrent computation and parallel join algorithms.

## Key findings

- MonARCh can monitor between 300 to 25,000 SPARQL queries in a two-node cluster.
- The system uses a publish/subscribe model with federated query execution for result change notifications.
- MonARCh outperforms existing systems in scalability and meets all required monitoring criteria.

## Abstract

Monitoring the data sources for possible changes is an important consumption requirement for applications running in interaction with the Web of Data. In this article, MonARCh which is an architecture for monitoring the result changes of registered SPARQL queries in the Linked Data environment, is proposed. MonARCh can be comprehended as a publish/subscribe system in the general sense. However, it differs in how communication with the data sources is realized. Data sources in the Linked Data environment do not publish the changes in the data. MonARCh provides the necessary communication infrastructure between the data sources and the consumers for the notification of changes. Users subscribe SPARQL queries to the system which are then converted to federated queries. MonARCh periodically checks for updates by re-executing SERVICE clauses and notifying users in case of any result change. In addition, to provide scalability, MonARCh takes the advantage of concurrent computation of the actor model. The parallel join algorithm utilized speeds up query execution and result generation processes. The design science methodology is used during the design, implementation and evaluation of the architecture. When compared to the literature MonARCh meets all the sufficient requirements from the linked data monitoring and state of the art perspectives while having many outstanding features from both points of view. The evaluation results show that even while working under the limited two-node cluster setting MonARCh could reach from 300 to 25,000 query monitoring capacity according to the diverse query selectivities executed within our test bench.