Special Session 4： Data Spaces across the Cloud-Edge Continuum: Architectures for Decentralized Data Sharing and Distributed Computation

Special Session Chair: Cansu Tanrikulu, Fraunhofer Institute for Production Systems and Design Technology

Email: cansu.tanrikulu@ipk.fraunhofer.de

Data-intensive computing is increasingly distributed across organizations, data centers, high-performance computing facilities, and edge nodes. Centralized architectures, such as data lakes, warehouses, and platform-based aggregation, offer simplicity and high throughput but can introduce latency, data-movement costs, energy overhead, regulatory constraints, and loss of control over data. Data spaces and federated data ecosystems provide an alternative by enabling interoperable exchange, machine-readable usage policies, and computation close to where data is generated or maintained.

This Special Session will examine architectures, technologies, and methods for data sharing and distributed computation across the cloud-edge continuum. It will cover data-space connectors, data placement and replication, federated query processing, near-data computing, cross-domain workflow orchestration, metadata and provenance, runtime policy enforcement, and high-performance data transfer. Contributions from computational science, research-data infrastructures, digital twins, and data-intensive engineering are particularly encouraged. The session aims to compare centralized, decentralized, and hybrid approaches through measurable criteria such as latency, throughput, scalability, cost, energy consumption, resilience, and interoperability, while advancing reproducible methods, benchmarks, and practical implementations.

Topics of interest include, but are not limited to:

1. Data-space architectures, protocols, connectors, and interoperability standards
2. Cloud-edge-device continuum architectures and distributed computing infrastructures
3. Centralized, decentralized, federated, and hybrid data-sharing models
4. Data placement, partitioning, caching, replication, and storage tiering
5. Federated query processing, in-situ analytics, and near-data computing
6. Distributed workflow orchestration, scheduling, and resource management
7. High-performance data transfer and networking across administrative domains
8. Data sovereignty, usage control, access management, trust, and auditability
9. Metadata, semantic interoperability, catalogs, provenance, and data lineage
10. Cloud-edge integration for computational and data-intensive science
11. Research-data infrastructures, FAIR data, simulation workflows, and digital twins
12. Performance, scalability, resilience, cost, and energy-efficiency evaluation
13. Testbeds, benchmarks, reference implementations, and reproducible experiments
14. Applications in manufacturing, mobility, energy, healthcare and smart cities

Special session proposals will be evaluated based on the timeliness of the topic, its uniqueness, and qualifications of the proposers. Accepted special sessions will be listed on the ICCAI 2027 website. To avoid multiple special sessions covering similar topics, it is likely that an accepted proposal will be combined with similar proposals. Organizers should not contribute more than one paper for the special session.

Key Dates

              ●  Deadline for Special Session Proposals submission: Nov. 20, 2026
              ●  Notification of acceptance of the proposal: Nov. 30, 2026
              ●  Submission of papers for Special Sessions: Dec. 10, 2026
              ●  Notification of papers acceptance: Jan. 10, 2027
              ●  Deadline for camera-ready papers: Feb. 10, 2027

The papers should be formatted following the submissionn instructions. They should be submitted through the regular online submission system and will undergo the same peer-review process as regular papers. Proponents of special sessions are invited to submit their proposals in pdf attached to an email to the address: iccai@vip.163.com, with the Subject line: ‘Special Session Proposal’.

We are looking forward to your proposals!