Machine Learning Systems: A Survey from a Data-Oriented Perspective

Christian Cabrera; Andrei Paleyes; Pierre Thodoroff; Neil D. Lawrence

doi:doi.org/10.1145/3769292

Back to publications

Machine Learning Systems: A Survey from a Data-Oriented Perspective

Christian Cabrera, Andrei Paleyes, Pierre Thodoroff, Neil D. Lawrence

ACM Computing Surveys, 2026.

Abstract

Engineers are deploying ML models as parts of real-world systems with the upsurge of AI technologies. Real-world environments challenge the deployment of such systems because these environments produce large amounts of heterogeneous data, and users require increasingly efficient responses. These requirements push prevalent software architectures to the limit when deploying ML-based systems. Data-Oriented Architecture (DOA) is an emerging style that better equips systems to integrate ML models. Even though articles on deployed ML-based systems do not mention DOA, their authors make design decisions that implicitly follow DOA. Implicit decisions create a knowledge gap, limiting practitioners’ ability to implement ML-based systems. This article surveys why, how, and to what extent practitioners have adopted DOA to implement ML-based systems. We overcome the knowledge gap by answering these questions and explicitly showing the design decisions and practices behind these systems. The survey follows a well-known systematic and semi-automated methodology for reviewing articles in software engineering. The majority of reviewed works partially adopt DOA. Such an adoption enables systems to address big data management, low-latency processing, resource management, security, and privacy requirements. Based on these findings, we formulate practical advice to facilitate the deployment of ML-based systems.

Links

Cite this Paper

BibTeX


@Article{publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective,
  title = 	 {Machine Learning Systems: A Survey from a Data-Oriented Perspective},
  author = 	 {Cabrera, Christian and Paleyes, Andrei and Thodoroff, Pierre and Lawrence, Neil D.},
  journal =      {ACM Computing Surveys},
  year = 	 {2026},
  doi = 	 {doi.org/10.1145/3769292},
  pdf = 	 {https://doi.org/10.1145/3769292},
  url = 	 {/publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective.html},
  abstract = 	 {Engineers are deploying ML models as parts of real-world systems with the upsurge of AI technologies. Real-world environments challenge the deployment of such systems because these environments produce large amounts of heterogeneous data, and users require increasingly efficient responses. These requirements push prevalent software architectures to the limit when deploying ML-based systems. Data-Oriented Architecture (DOA) is an emerging style that better equips systems to integrate ML models. Even though articles on deployed ML-based systems do not mention DOA, their authors make design decisions that implicitly follow DOA. Implicit decisions create a knowledge gap, limiting practitioners’ ability to implement ML-based systems. This article surveys why, how, and to what extent practitioners have adopted DOA to implement ML-based systems. We overcome the knowledge gap by answering these questions and explicitly showing the design decisions and practices behind these systems. The survey follows a well-known systematic and semi-automated methodology for reviewing articles in software engineering. The majority of reviewed works partially adopt DOA. Such an adoption enables systems to address big data management, low-latency processing, resource management, security, and privacy requirements. Based on these findings, we formulate practical advice to facilitate the deployment of ML-based systems.}
}

Endnote

%0 Journal Article
%T Machine Learning Systems: A Survey from a Data-Oriented Perspective
%A Christian Cabrera
%A Andrei Paleyes
%A Pierre Thodoroff
%A Neil D. Lawrence
%J ACM Computing Surveys
%D 2026	
%F publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective
%R doi.org/10.1145/3769292
%U /publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective.html
%X Engineers are deploying ML models as parts of real-world systems with the upsurge of AI technologies. Real-world environments challenge the deployment of such systems because these environments produce large amounts of heterogeneous data, and users require increasingly efficient responses. These requirements push prevalent software architectures to the limit when deploying ML-based systems. Data-Oriented Architecture (DOA) is an emerging style that better equips systems to integrate ML models. Even though articles on deployed ML-based systems do not mention DOA, their authors make design decisions that implicitly follow DOA. Implicit decisions create a knowledge gap, limiting practitioners’ ability to implement ML-based systems. This article surveys why, how, and to what extent practitioners have adopted DOA to implement ML-based systems. We overcome the knowledge gap by answering these questions and explicitly showing the design decisions and practices behind these systems. The survey follows a well-known systematic and semi-automated methodology for reviewing articles in software engineering. The majority of reviewed works partially adopt DOA. Such an adoption enables systems to address big data management, low-latency processing, resource management, security, and privacy requirements. Based on these findings, we formulate practical advice to facilitate the deployment of ML-based systems.

RIS


TY  - JOUR
TI  - Machine Learning Systems: A Survey from a Data-Oriented Perspective
AU  - Christian Cabrera
AU  - Andrei Paleyes
AU  - Pierre Thodoroff
AU  - Neil D. Lawrence
DA  - 2026/09/27	
ID  - publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective
DO  - doi.org/10.1145/3769292
L1  - https://doi.org/10.1145/3769292
UR  - /publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective.html
AB  - Engineers are deploying ML models as parts of real-world systems with the upsurge of AI technologies. Real-world environments challenge the deployment of such systems because these environments produce large amounts of heterogeneous data, and users require increasingly efficient responses. These requirements push prevalent software architectures to the limit when deploying ML-based systems. Data-Oriented Architecture (DOA) is an emerging style that better equips systems to integrate ML models. Even though articles on deployed ML-based systems do not mention DOA, their authors make design decisions that implicitly follow DOA. Implicit decisions create a knowledge gap, limiting practitioners’ ability to implement ML-based systems. This article surveys why, how, and to what extent practitioners have adopted DOA to implement ML-based systems. We overcome the knowledge gap by answering these questions and explicitly showing the design decisions and practices behind these systems. The survey follows a well-known systematic and semi-automated methodology for reviewing articles in software engineering. The majority of reviewed works partially adopt DOA. Such an adoption enables systems to address big data management, low-latency processing, resource management, security, and privacy requirements. Based on these findings, we formulate practical advice to facilitate the deployment of ML-based systems.
ER  -

APA


Cabrera, C., Paleyes, A., Thodoroff, P. & Lawrence, N.D.. (2026). Machine Learning Systems: A Survey from a Data-Oriented Perspective. ACM Computing Surveys doi:doi.org/10.1145/3769292 Available from /publications/real-world-machine-learning-systems-a-survey-from-a-data-oriented-architecture-perspective.html.

Related People

Christian Cabrera Jojoa

Assistant Research Professor, Cambridge University

View Profile

Andrei Paleyes

Visiting Researcher, Cambridge University

View Profile

Pierre Thodoroff

PhD Student, Cambridge University

View Profile

Neil D. Lawrence

The DeepMind Professor of Machine Learning, Cambridge University

View Profile

Related Projects

AutoAI

While excitement about the potential of artificial intelligence (AI) technologies continues to build, a gap is emerging between our aspirations for the benefits of AI and our ability to deploy these technologies to tackle real-world challenges.

Data-Oriented Architectures for AI-based Systems

Data Oriented Architecture (DOA) is a software architecture pattern that creates data-driven, loosely coupled, decentralised, and open systems. DOA achieves these goals by exposing systems' data as first class citizen to distributed, stateless, and asynchronous systems’ components. These design decisions enable DOA-based systems to achieve desirable properties such as data availability, reusability, and monitoring, as well as systems adaptability, scalability, and autonomy.

Interfaces

The Interfaces research programme develops interpretable, self-sustaining multi-agent AI systems, treating software as the interface between socio-technical needs and AI capabilities, with validation in healthcare and other high-stakes settings.