Conference paper
2023 19th International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT), 2023, pp. 429-436
APA
Click to copy
Zafeiropoulos, A., Fotopoulou, E., Vassilakis, C., Tzanettis, I., Lombardo, C., Carrega, A., & Bruschi, R. (2023). Intent-Driven Distributed Applications Management Over Compute and Network Resources in the Computing Continuum. In 2023 19th International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT) (pp. 429โ436). https://doi.org/10.1109/DCOSS-IoT58021.2023.00074
Chicago/Turabian
Click to copy
Zafeiropoulos, Anastasios, Eleni Fotopoulou, Constantinos Vassilakis, Ioannis Tzanettis, Chiara Lombardo, Alesssandro Carrega, and Roberto Bruschi. โIntent-Driven Distributed Applications Management Over Compute and Network Resources in the Computing Continuum.โ In 2023 19th International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT), 429โ436, 2023.
MLA
Click to copy
Zafeiropoulos, Anastasios, et al. โIntent-Driven Distributed Applications Management Over Compute and Network Resources in the Computing Continuum.โ 2023 19th International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT), 2023, pp. 429โ36, doi:10.1109/DCOSS-IoT58021.2023.00074.
BibTeX Click to copy
@inproceedings{zafeiropoulos2023a,
title = {Intent-Driven Distributed Applications Management Over Compute and Network Resources in the Computing Continuum},
year = {2023},
pages = {429-436},
doi = {10.1109/DCOSS-IoT58021.2023.00074},
author = {Zafeiropoulos, Anastasios and Fotopoulou, Eleni and Vassilakis, Constantinos and Tzanettis, Ioannis and Lombardo, Chiara and Carrega, Alesssandro and Bruschi, Roberto},
booktitle = {2023 19th International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT)}
}
๐ฌ Intent-Driven Distributed Application Management: Coordinating Computing and Networking Resources Across the Computing Continuum
The evolution toward highly distributed computing environments introduces significant challenges for application deployment, infrastructure coordination, and lifecycle management.
Applications increasingly interact with heterogeneous compute and network resources distributed across IoT devices, edge infrastructures, and centralized cloud platforms.
Traditional management approaches often require detailed infrastructure configuration and coordination, creating additional complexity as applications and operational requirements evolve.
Our research examines intent-driven orchestration as an approach to increasing automation, integration, and intelligence in the management of distributed applications and their underlying computing and networking resources.
The fundamental principle is to decouple high-level application objectives from the detailed infrastructure operations required to satisfy them.
The technical infographic illustrates several complementary concepts.
๐น 1. Intent Abstraction
Users and application operators express desired objectives and policies without necessarily specifying the complete sequence of infrastructure operations.
Intent-driven management provides a higher level of abstraction between service requirements and resource coordination.
๐น 2. Compute and Network Resource Coordination
Distributed applications rely on heterogeneous infrastructure resources that must be coordinated across administrative and technological boundaries.
Orchestration mechanisms manage the interplay between application requirements and available infrastructure capabilities.
๐น 3. Computing Continuum
The computing continuum encompasses interconnected computing environments with different characteristics, including IoT, edge, and cloud resources.
Coordinating these resources introduces challenges involving heterogeneity, resource availability, and application management.
๐น 4. Application Lifecycle Management
Distributed applications require coordinated management throughout their lifecycle.
The growing diversity of platforms and resources increases the complexity of managing application operations and responding to changing requirements.
๐น 5. Policy-Driven Resource Management
High-level policies provide guidance for infrastructure management decisions.
The orchestration system is responsible for coordinating and optimizing resources in accordance with the expressed objectives.
๐น 6. Adaptation to Changing Conditions
Intent-driven orchestration aims to support adjustments in resource management as application needs and environmental conditions change.
This capability is particularly relevant for distributed systems with heterogeneous and evolving resource requirements.
A possible intent-driven management workflow involves five stages:
1. Intent Definition: Specify high-level application objectives and relevant policies.
2. Requirement Interpretation: Translate those objectives into infrastructure-related management requirements.
3. Resource Coordination: Determine how available computing and networking resources should support the distributed application.
4. Application Management: Coordinate the necessary deployment and lifecycle operations.
5. Adaptation: Reconsider resource management decisions when requirements or operating conditions change.
This sequence is an explanatory architectural interpretation rather than a verified reproduction of the paper's complete implementation workflow.
โ Management Abstraction: Reduce the need for users to describe low-level infrastructure operations.
โ Operational Automation: Support increasingly automated coordination of distributed resources.
โ Infrastructure Integration: Manage resources spanning heterogeneous computing and networking environments.
โ Application Flexibility: Facilitate adaptation to changing service requirements and operating conditions.
โ Resource Coordination: Align infrastructure decisions with application objectives and policies.
โ Lifecycle Management: Address the growing complexity of operating distributed applications throughout the computing continuum.
Intent-driven orchestration represents an important direction for managing distributed applications across heterogeneous computing and networking infrastructures.
By shifting from detailed configuration instructions toward high-level objectives and policies, orchestration systems can provide a foundation for more integrated, adaptable, and automated management.
The broader research challenge is to translate these objectives into reliable infrastructure operations while coordinating resources across the computing continuum.

๐๐ What If Applications Could Tell Networks What They Need, Instead of How to Configure Everything? โ๏ธโก
Imagine a world where applications seamlessly operate across IoT devices, edge servers, and cloud platforms without requiring developers to manually coordinate every computing and networking resource.
As distributed systems become increasingly heterogeneous, managing their deployment, configuration, and adaptation introduces significant operational complexity.
What if we could simply express our objectives and let an intelligent orchestration system coordinate the underlying infrastructure?
Our research investigates this vision through intent-driven management of distributed applications across the computing continuum.
The central idea is straightforward: instead of specifying every technical operation, users define high-level goals and policies, while orchestration mechanisms coordinate resources to satisfy those requirements.
๐ฏ Intent-Based Management: Express application objectives and operational policies at an abstract level.
โ๏ธ Computing Continuum: Coordinate heterogeneous computing and networking resources across distributed environments.
โ๏ธ Automated Orchestration: Reduce the complexity of resource coordination, application deployment, and configuration.
๐ Dynamic Adaptation: Adjust infrastructure management decisions according to changing application requirements and environmental conditions.
๐งฉ Application Lifecycle Management: Address the complexity of managing distributed applications throughout their operational lifecycle.
๐ก Why does this matter?
โ Less dependence on manual infrastructure configuration.
โ Greater flexibility in managing heterogeneous computing environments.
โ Opportunities for more efficient resource coordination.
โ Improved alignment between application requirements and infrastructure management.
โ A foundation for increasingly automated distributed application management.
Our manga-inspired infographic illustrates the journey from complex infrastructure management toward a more abstract, policy-driven orchestration paradigm.
๐ฏ The message: Describe WHAT your application needs. Let orchestration determine HOW computing and networking resources should be coordinated.
