ase261004

Read full text Summarise with VIRAT
Authors: anonymous
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

# Building a Cross-Protocol Anomaly Analysis Pipeline for Mobile Core Networks This repository contains the source code, additional implementation details, and supplementary material accompanying the anonymous submission of our paper. To preserve anonymity, this repository intentionally omits author names, affiliations, proprietary data, and identifying metadata. --- ## Repository structure - `source code` Python code used to: - fuse SS7, Diameter, and GTP signaling into per-subscriber, per-minute fused records - construct synthetic anomalies through field-group swaps across fused records - embed fused records using multiple embedding models - apply unsupervised anomaly detectors per embedding model - compute consensus scores as the number of embedding models flagging a record as anomalous - generate the statistics and tables reported in the paper See the docstrings and inline comments in the individual files for script-level details and expected inputs and outputs. - `supplementary details` Additional documentation that complements the main paper, including the detailed descriptions of the field groups used for synthetic anomaly construction and the resulting cross-protocol inconsistencies. The main supplementary file is: - `Supplementary_details__Synthetic_Anomaly_Construction.pdf` --- ## Data availability This repository contains the source code, additional implementation details, and supplementary material used in this study. The underlying signaling data are not included. The dataset consists of operational mobile-core signaling records containing sensitive subscriber and operator information. These data cannot be released because they are subject to operator confidentiality, contractual restrictions, and privacy constraints. As a result, we cannot redistribute the raw dataset or derived per-subscriber fused records. The code and supplementary material are provided to document the methodology and to support adaptation to other datasets under appropriate agreements. --- ## Code usage The code is intended for researchers and practitioners with access to their own mobile-core signaling data. Typical usage pattern: 1. Prepare fused records Export SS7, Diameter, and GTP signaling records from your environment and construct per-subscriber, per-minute fused records matching the schema expected by the scripts. 2. Generate synthetic anomalies Use the mutation scripts to create synthetic anomalies through field-group swaps, consistent with the supplementary document. 3. Embed fused records Serialize each fused record to text and obtain vector representations using the supported embedding models. 4. Run unsupervised anomaly detectors Apply the anomaly detection scripts to obtain per-model anomaly decisions. 5. Compute consensus scores and statistics Aggregate per-model decisions into a consensus score per record and reproduce tables and analyses analogous to those reported in the paper. Because the dataset is not distributed, paths, environment variables, and model choices may need to be adapted to your local environment.

Fields of Study & Subjects

Network Security and Intrusion Detection Anomaly Detection Techniques and Applications Software System Performance and Reliability

Publication details

Journal / source
Zenodo (CERN European Organization for Nuclear Research)
Publication date
30 March 2026
Language
EN