CODECO Media Consumption Traffic for Edge-Cloud Video Streaming Datasets
Authors/Creators
Description
Overview
| Field | Value |
|---|---|
| Dataset Name | Media Consumption Traffic for Edge-Cloud Video Streaming |
| Dataset Unique ID | HEU-101092696-CODECO-P3-MDS-MediaConsumptionTraffic |
| Project | CODECO — COgnitive DECentralized Orchestration |
| Grant Agreement Number | 101092696 |
| Topic | HORIZON-CL4-2022-DATA-01 |
| Related WP / Task | WP5 |
| Version | 1.0 |
| Release Date | 2026 |
| License | ODC Open Database License (ODbL) |
Dataset Description
This dataset contains HTTP request/response traces generated during video streaming sessions in the context of the P3 use case of the CODECO project.
The data was extracted from HAR (HTTP Archive) network captures obtained from browser developer tools during demonstrations with CDN platform.
In this scenario, the central CDN origin server was deployed in a data centre located in Greece, hosting the master copy of the media content. In addition, an edge cache instance was deployed in Madrid (Spain) on a smaller compute cluster orchestrated through the CODECO platform.
When a user accessed the streaming service, the request was initially handled by the central CDN endpoint. The client was then redirected to the edge cache running in the Madrid cluster, allowing the media segments to be served locally from the edge rather than from the central site.
The resulting network traces in this dataset therefore reflect traffic delivered by both the central CDN node (Greece) and the edge cache node (Madrid) during streaming sessions.
Each row represents a single HTTP transaction associated with media streaming activity, typically including:
- Media segment requests (e.g.,
.ts,.m4s) - Streaming manifest files (e.g.,
.m3u8,.mpd) - Associated request metadata and network performance metrics.
The dataset enables analysis of content delivery behaviour across distributed edge/cloud nodes, supporting research on:
- media streaming performance
- edge caching efficiency
- distributed CDN architectures
- QoE monitoring for video delivery
Use Cases and Intended Audience
This dataset is intended for researchers and engineers working on:
- Edge-cloud media delivery
- Content Delivery Networks (CDN)
- Edge caching and orchestration
- Network performance analysis
Typical analyses include:
- chunk delivery latency
- throughput estimation
- edge cache node comparison
- streaming session reconstruction
Data Origin and Generation Methodology
- Origin: Network traffic captured during CODECO MDS demonstrations.
- Capture method: Browser developer tools network inspector (HAR export).
- Processing pipeline:
- Streaming sessions were executed using the CODECO media delivery service.
- Browser network traffic was recorded as HAR files.
- A Python extraction pipeline parsed the HAR entries and anonymized IPs.
- Relevant fields were extracted and flattened into a CSV dataset.
The resulting dataset represents real HTTP traffic observed during streaming playback.
File Structure
Smart Cities MDS
- README.md
- HEU-101092696-CODECO-P3-MDS-MediaConsumptionTraffic.csv
- anonymize-raw-traffic-file-madrid-cluster.har
CSV Schema
File: HEU-101092696-CODECO-P3-MDS-MediaConsumptionTraffic.csv
| Column | Source | Description |
|---|---|---|
startedDateTime |
HAR | Timestamp when the request started |
url |
HAR | Full request URL |
hostname |
Derived | Host extracted from URL |
path |
Derived | URL path |
file_ext |
Derived | File extension |
asset_type |
Derived | Resource type (manifest, chunk, other) |
chunk_id |
Derived | Media segment identifier extracted from filename |
status |
HAR | HTTP response status code |
statusText |
HAR | HTTP response message |
serverIPAddress |
HAR | IP address of the server/cache node |
connection |
HAR | Connection identifier |
content_size |
HAR | Decoded content size |
content_encoding |
HAR | Encoding used for stored payload |
bodySize |
HAR | Actual transferred body size |
response_headersSize |
HAR | Size of response headers |
request_headersSize |
HAR | Size of request headers |
time_ms |
HAR | Total request duration |
blocked_ms |
HAR | Time waiting for network slot |
connect_ms |
HAR | TCP connection time |
wait_ms |
HAR | Server processing time (TTFB) |
receive_ms |
HAR | Response download time |
throughput_Bps |
Derived | Estimated throughput (bytes/sec) |
is_success |
Derived | 1 if 200 ≤ status < 300 |
is_redirect |
Derived | 1 if 300 ≤ status < 400 |
is_error |
Derived | 1 if status ≥ 400 |
is_media |
Derived | 1 if asset is a streaming manifest or media chunk |
Asset Type Classification
| Extension | Type |
|---|---|
.m3u8, .mpd |
manifest |
.ts, .m4s, .mp4 |
media chunk |
| others | other |
Data Format
- Format: CSV (Comma-Separated Values)
- Encoding: UTF-8
- Header row: Included
- Time format: ISO 8601 timestamps
Metadata and Standards
- Capture format: HAR (HTTP Archive)
- Network protocol: HTTP/HTTPS
- Time units: milliseconds for timing metrics
License
This dataset is released under the ODC Open Database License (ODbL).
You are free to:
- Share — copy and redistribute the database
- Create — produce works from the database
- Adapt — modify and build upon the database
Under the conditions that you:
- Attribute the source
- Share derived databases under the same license
- Keep the data open
Full license text:
https://opendatacommons.org/licenses/odbl/
Citation
If you use this dataset in your research, please cite the CODECO project:
CODECO — COgnitive DECentralized Orchestration
Horizon Europe Grant Agreement No. 101092696
Topic: HORIZON-CL4-2022-DATA-01
Dataset:
HEU-101092696-CODECO-P3-MDS-MediaConsumptionTrafficContact
For questions regarding this dataset, please contact the CODECO consortium or open an issue in the repository.
Project website:
https://www.codeco-project.eu