Pasar al contenido principal

GRAPH NEURAL NETWORKS FOR ROBUST AI/ML-DRIVEN NETWORK SECURITY APPLICATIONS

PCI2023-145974-2

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Afrontar las Prioridades de Nuestro Entorno
Subprograma Subprograma Estatal de Internacionalización
Convocatoria Proyectos Programación Conjunta Internacional
Año convocatoria 2023
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSITAT POLITECNICA DE CATALUNYA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 20
Encontrada(s) 1 página(s)

UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

Atomic-level mechanisms of abnormal activation in NRAS oncogenes from two dimensional free energy landscapes

  • Hu, Zheyao
  • Martí Rabassa, Jordi|||0000-0002-3721-9634
The NRAS-mutant subset of melanoma is one of the most aggressive and lethal types associated with poor overall survival. Unfortunately, a low understanding of the NRAS-mutant dynamic behavior has lead to the lack of clinically approved therapeutic agents able to directly target NRAS oncogenes. In this work, accurate local structures of NRAS and its mutants have been fully explored through the corresponding free energy surfaces obtained by microsecond scale well-tempered metadynamics simulations. Free energy calculations are crucial to reveal the precise mechanisms of Q61 mutations at the atomic level. Considering specific atom-atom distances and angles as appropriate reaction coordinates, we have obtained free energy surfaces revealing local and global minima together with their main transitions states, unveiling the mechanisms of abnormal NRAS activation from atomic-level and quantitatively analyzing the corresponding stable states. This will help to advance in our understanding of the basic mechanisms of NRAS mutations, offering new opportunities for the design of potential inhibitors., We are thankful for the financial support provided by the Spanish Ministry of Science, Innovation and Universities (funds MCIU/AEI/FEDER, EU). This publication is a part of the I + D + i project number PID2021-124297NB-C32, funded by MCIN/AEI/10.13039/501100011033/ and “FEDER Una manera de hacer Europa”. Zheyao Hu is a Ph.D. fellow from the Chinese Scholarship Council (grant 202006230070). J. M. thanks the Generalitat de Catalunya for the support through the grant Grup de Recerca SGR-Cat2021 Condensed, Complex and Quantum Matter Group reference 2021SGR-01411 and the Polytechnic University of Catalonia-Barcelona Tech through the funding AGRUPS. Computational resources awarded by the Barcelona Supercomputing Center-Spanish Supercomputing Network (grant BCV-2024-2-0005 “Simulation and modeling of biosystems with applications to medicine (1): drug design for the treatment of cancers of the RAS family”) are also acknowledged., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

Graph neural networks for graph-level anomaly detection in network intrusion detection

  • Anyfantis, Georgios|||0009-0002-3747-9935
  • Castell Uroz, Ismael
  • Barlet Ros, Pere|||0000-0001-7837-0886
Network attacks are a growing problem in today's world. Their number is growing exponentially, and they become more sophisticated every year, accentuating the need for robust Network Intrusion Detection Systems. Graph Neural Networks have been proven to be very capable in detecting and categorising network attacks, making them ideal for detection in real-life deployments. However, they are limited to training on a small number of outdated labelled datasets. We propose a new method for anomaly detection through a graph heterogeneous autoencoder as a way to reduce reliance on labelled datasets. Our graph autoencoder reconstructs the structure and features of a graph representing the traffic connections in a network. We train the autoencoder with normal graphs and then derive embeddings to use with traditional anomaly detection based on Machine Learning. Our results indicate that our solution can create distinct normal and anomalous graph embeddings that can then be used to detect attacks with a high degree of accuracy., This work was supported by the CHISTERA grant CHISTERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference nº PCI2023-145974-2) funded by the Agencia Estatal de Investigación through the PCI 2023 call. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

TSGFM - Graph neural networks for zero-shot time series forecasting in network monitoring

  • Latif Martínez, Hamid|||0000-0001-6006-3175
  • Vanerio, Juan
  • Casas Hernández, Pedro
  • Suárez Varela, José
  • Cabellos Aparicio, Alberto|||0000-0001-9329-7584
  • Barlet Ros, Pere|||0000-0001-7837-0886
We present TSGFM, a Time Series Graph Foundation Model for zero-shot network monitoring, leveraging spatiotemporal Graph Neural Networks (GNNs) to extract transferable representations across diverse multivariate time series (MTS) domains. Pretrained on heterogeneous time series datasets, TSGFM enables generalization without task-specific fine-tuning, addressing core challenges in dynamic network environments. TSGFM is benchmarked across five real-world MTS datasets and seven zero-shot forecasting scenarios, outperforming five state-of-the-art baselines in six out of seven tasks. Most notably, in zero-shot network monitoring analysis, TSGFM surpasses all competing models by at least 18%, even without any prior exposure to network monitoring data. We further compare TSGFM against leading Time Series Foundation Models (TSFMs), including TimeGPT and TimesFM. TSGFM achieves performance on par with TimeGPT, occasionally surpassing it, and consistently outperforms TimesFM, while using significantly less pretraining data and relying on a much simpler architecture. A detailed analysis of TSGFM’s learned spatial attention patterns reveals domain-specific connectivity structures. In particular, lower attention weights in network monitoring tasks suggest that dense spatial graphs may be unnecessary, opening opportunities for efficient spatial pruning without sacrificing accuracy. This challenges prevailing assumptions favoring fully connected spatiotemporal GNNs. To foster transparency and reproducibility, we release the complete implementation of TSGFM as open source, as well as the tested datasets., This work has been supported by the FWF Austrian Science Fund, Project reference I-6653, as well as by the Agencia Estatal de Investigación, Project reference PCI2023-145974-2, as part of the EU CHIST-ERA-2022-SPiDDS-02 project GRAPHS4SEC. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

TSGFM - Towards a graph foundation model for time series analysis in network monitoring

  • Latif Martínez, Hamid|||0000-0001-6006-3175
  • Casas Hernández, Pedro
  • Suárez Varela, José
  • Cabellos Aparicio, Alberto|||0000-0001-9329-7584
  • Barlet Ros, Pere|||0000-0001-7837-0886
We present TSGFM, a Time Series Graph Foundation Model for network monitoring data analysis, based on spatiotemporal Graph Neural Networks (GNN). Inspired by the success of foundation models in achieving generalization and adaptation, TSGFM leverages pretraining on diverse multivariate timeseries (MTS) data from multiple domains to enable effective zero-shot analysis in network monitoring tasks. We compare TSGFM performance against five state-of-the-art AI/ML models in seven zero-shot forecasting scenarios, using five MTS datasets from different domains. Evaluations demonstrate that TSGFM achieves superior performance in six out of seven zero-shot testing scenarios. Most notably, in zero-shot network monitoring analysis, TSGFM surpasses all competing models by at least 18%, even without training on any network monitoring data., This work has been supported by the FWF Austrian Science Fund, Project reference I-6653, as well as by the Agencia Estatal de Investigación, Project reference PCI2023-145974- 2, as part of the EU CHIST-ERA-2022-SPiDDS-02 project GRAPHS4SEC. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

RouteNet-Gauss: Hardware-enhanced network modeling with machine learning

  • Güemes Palau, Carlos|||0009-0001-8654-0512
  • Ferriol Galmés, Miquel|||0000-0002-7806-2979
  • Paillissé Vilanova, Jordi|||0000-0001-7733-9713
  • López Brescó, Albert|||0009-0005-0611-5516
  • Barlet Ros, Pere|||0000-0001-7837-0886
  • Cabellos Aparicio, Alberto|||0000-0001-9329-7584
Network simulation is pivotal in network modeling, assisting with tasks ranging from capacity planning to performance estimation. Traditional approaches such as Discrete Event Simulation (DES) face limitations in terms of computational cost and accuracy. This paper introduces RouteNet-Gauss, a novel integration of a testbed network with a Machine Learning (ML) model to address these challenges. By using the testbed as a hardware accelerator, RouteNet-Gauss generates training datasets rapidly and simulates network scenarios with high fidelity to real-world conditions. Experimental results show that RouteNet-Gauss significantly reduces prediction errors by up to 95% and achieves a 488x speedup in inference time compared to state-of-the-art DES-based methods. RouteNet-Gauss’s modular architecture is dynamically constructed based on the specific characteristics of the network scenario, such as topology and routing. This enables it to understand and generalize to different network configurations beyond those seen during training, including networks up to 10x larger. Additionally, it supports Temporal Aggregated Performance Estimation (TAPE), providing configurable temporal granularity and maintaining high accuracy in flow performance metrics. This approach shows promise in improving both simulation efficiency and accuracy, offering a valuable tool for network operators., This work was supported in part by the I+D+i Project titled Beyond 5G network infrastructure Leveraging deep AI acrOSS ultra-high capacity Optical data transport and quantum physical Substrates (BLOSSOMS) funded by Ministerio de Ciencia, Innovación y Universidades /Agencia Estatal de Investigación (MICIU/AEI)/10.13039/501100011033 and European Regional Development Fund/European Union (ERDF/EU), under Grant PID2024-158530OB-I00; in part by Catalan Institution for Research and Advanced Studies (ICREA); and in part by MICIU/AEI/10.13039/501100011033 and the European Union (GRAPHS4SEC project) under Grant PCI2023-145974-2. The work of Carlos Guemes-Palau was supported by the AGAUR-FI ajuts through the Joan Oró of the Secretariat of Universities and Research of the Department of Research and Universities of the Generalitat of Catalonia and European Social Plus Fund under Grant 2023 F-1 00083., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

GAT-AD: Graph Attention Networks for contextual anomaly detection in network monitoring

  • Latif Martínez, Hamid|||0000-0001-6006-3175
  • Suárez Varela, José
  • Cabellos Aparicio, Alberto|||0000-0001-9329-7584
  • Barlet Ros, Pere|||0000-0001-7837-0886
Network anomaly detection is essential to promptly detect and fix issues in the network. Particularly, detecting traffic anomalies enables the early detection of configuration errors, malicious activities, or equipment malfunctions that could lead to severe impact on the network. In this paper, we present GAT-AD, a Deep Learning-based anomaly detection solution for network monitoring systems, which integrates a custom neural network model based on Graph Attention Networks (GAT). Our solution monitors aggregated traffic on origin–destination flows and automatically defines contexts that group flows with similar past activity. The neural network model within GAT-AD can be efficiently trained in a self-supervised manner. We evaluate our solution against two state-of-the-art anomaly detection baselines also based on graph representations and Deep Learning, in two different datasets: (i) WaDi, which is a well-known dataset for anomaly detection in a distributed sensor network, and (ii) Abilene, where we inject synthetically-generated anomalies into a dataset with real-world traffic from a large-scale backbone network. The results show that GAT-AD outperforms the two anomaly detection baselines: in WaDi by 14.1% in recall and 10.07% in F1-score, and in the Abilene dataset by ˜ 17.5% recall with respect to the best baseline., This publication is part of the Spanish I+D+i project TRAINER-A (ref. PID2020-118011GB-C21), funded by MCIN/AEI/10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA), Spain. This work was also supported by the CHISTERA grant CHIST-ERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference no PCI2023-145974-2) funded by the Agencia Estatal de Investigación, Spain through the PCI 2023 call. This work was also supported by the Spanish Ministry of Economic Affairs and Digital Transformation and the European Union – NextGeneration EU, in the framework of the Recovery Plan, Transformation and Resilience (PRTR) (Call UNICO I+D 5G 2021, refs. number TSI-063000-2021-3, TSI-063000-2021-38, and TSI-063000-2021-52). This work was partially supported by the ORIGAMI project, funded by the Smart Networks and Services Joint Undertaking (SNS JU), Spain under the European Union’s Horizon Europe research and innovation programme under Grant Agreement No 101139270., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

AST-GNN: A graph neural network for web tracking detection

  • Campeny Roig, Eloi
  • Castell Uroz, Ismael
  • Barlet Ros, Pere|||0000-0001-7837-0886
Web tracking technology is prevalent on the Internet today, with most websites employing user identification systems that can accurately identify users or devices behind browsers. While numerous works in literature attempt to create machine learning models for detecting these identification systems, many rely on features susceptible to obfuscation techniques and are only partially capable of identifying specific subsets of web tracking algorithms they were trained on. Additionally, classification is typically done over entire resources, making it difficult to distinguish between web tracking code and legitimate code within the same file. In this work, we propose AST-GNN, a graph neural network model applied to the abstract syntax tree structure of JavaScript files, which can predict portions of code used for tracking purposes. By focusing on the code structure, AST-GNN can detect various web tracking systems and it is robust against obfuscation techniques. Our results show that the system has an accuracy rate above 95% in identifying web tracking code snippets, with computation performance in the order of milliseconds, fast enough to be used in real-time., This work was supported by the CHISTERA grant CHIST-ERA-22- SPiDDS-02 corresponding to the GRAPHS4SEC project (reference nº PCI2023-145974-2) funded by the Agencia Estatal de Investigación through the PCI 2023 call. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

BGP anomaly detection using the raw internet topology

  • Latif Martínez, Hamid|||0000-0001-6006-3175
  • Paillissé Vilanova, Jordi|||0000-0001-7733-9713
  • Barlet Ros, Pere|||0000-0001-7837-0886
  • Cabellos Aparicio, Alberto|||0000-0001-9329-7584
The Border Gateway Protocol (BGP) is central to the global connectivity of the Internet, enabling fast and efficient dissemination of routing information. Hence, detecting any anomaly concerning BGP announcements is of critical importance to ensure the continuous operation of Internet services. Typically, BGP anomaly detection algorithms have relied on features of the BGP messages, such as the average length of the AS_PATH attribute, the volume of messages, or the type of message (announcement or withdrawal). Even though these algorithms provide good performance, they do not take into account the Internet topology, that is, the graph of Autonomous Systems (AS) created by the BGP announcements. In addition, some of the existing algorithms can detect only specific types of anomalies, while others require retraining them to support new scenarios. In this paper we propose detecting BGP anomalies by leveraging the raw BGP topology graph, instead of manually curated features of the BGP messages. We implement a Machine Learning algorithm to process the entire BGP topology and evaluate it with real-world data from 4 well-known incidents. We compare our proposal against two state-of-the-art solutions and a classical method that use BGP features and features of the BGP topology, not the topology itself. Our results show that our solution obtains remarkable performance identifying the incidents. Finally, we test our model with regular data (non-anomalous) to prove that it can be used in a production scenario, with samples processed on the fly and guaranteeing a low false alarm rate., This work is part of the I+D+i project titled BLOSSOMS, grant PID2024-158530OB-I00, funded by MICIU/AEI/10.13039/501100011033/ and by ERDF/EU. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA). This work was also supported by Grant PCI2023-145974-2 funded by MICIU/AEI/10.13039/501100011033 and co-funded by the European Union (GRAPHS4SEC project). Jordi Paillissé is funded by European Union-Next Generation EU, Ministry of Universities and Recovery, Transformation and Resilience Plan, through a call from Universitat Politècnica de Catalunya (Grant Ref. 2022UPC-MSC-93871)., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Artículo científico (JournalArticle)

GraphCC: a practical graph learning-based approach to congestion control in datacenters

  • Bernárdez Gil, Guillermo
  • Suárez Varela, José
  • Shi, Xiang
  • Xiao, Shihan
  • Cheng, Xiangle
  • Barlet Ros, Pere|||0000-0001-7837-0886
  • Cabellos Aparicio, Alberto|||0000-0001-9329-7584
Congestion Control (CC) plays a fundamental role in optimizing traffic in Datacenter Networks (DCNs). Currently, DCNs implement two main CC protocols: DCTCP and DCQCN. Both protocols are based on Explicit Congestion Notification (ECN), where switches mark packets when they detect congestion. Nowadays, network experts carefully set ECN parameters to optimize the average network performance. However, today’s DCNs experience rapid and abrupt changes that severely affect the network state (e.g., dynamic workloads, incasts), which leads to under-utilization and sub-optimal performance. In this paper we present GraphCC, a framework for in-network CC optimization. GraphCC relies on Multi-agent Reinforcement Learning (MARL) and Graph Neural Networks (GNN), and is compatible with widely deployed ECN-based CC protocols. The proposed solution deploys distributed agents on switches that communicate with their neighbors to cooperate and optimize the global ECN configuration. In our evaluation, we test GraphCC with three real-world traffic workloads, focusing on its capability to accommodate scenarios unseen during training (e.g., traffic changes, failures). We compare GraphCC with a state-of-the-art MARL solution for ECN tuning, and observe that our method outperforms the state-of-the-art baseline in all evaluation scenarios, with improvements up to 20% in average Flow Completion Time, similar mean throughput (within 1%), and significant reductions in buffer occupancy (38.0-85.7%)., This publication is part of the Spanish I+D+i project TRAINER-A (reference PID2020-118011GB-C21), funded by MCIN/AEI/10.13039/ 501100011033, and the research has been partially funded by the Catalan Institution for Research and Advanced Studies (ICREA), the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia, and the European Social Fund. This work is also partially supported by the ORIGAMI project, funded by the Smart Networks and Services Joint Undertaking (SNS JU) under the European Union’s Horizon Europe research and innovation programme under Grant Agreement No 101139270, as well as by the CHISTERA grant CHIST-ERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference PCI2023-145974-2), funded by the Agencia Estatal de Investigación through the PCI 2023 call., Peer Reviewed
UPCommons. Portal del coneixement obert de la UPC
Publicación Contribución de congreso (ConferenceOutput)

A first look into Utiq: next-generation cookies at the ISP level

  • Castell Uroz, Ismael
  • Barlet Ros, Pere|||0000-0001-7837-0886
Online privacy has become increasingly important in recent years. While third-party cookies have been widely used for years, they have also been criticized for their poten-tial impact on user privacy. They can be used by advertisers to track users across multiple sites, allowing them to build detailed profiles of their behavior and interests. However, nowadays, many browsers allow users to block third-party cookies, which limits their usefulness for advertisers. In this paper, we take a first look at Utiq, a new way of user tracking performed directly by the ISP, to substitute the third-party cookies used until now. We study the main properties of this new identification methodology and their adoption on the 10K most popular websites. Our results show that, although still marginal due to the restrictions imposed by the system, between 0.7% and 1.2% of websites already include Utiq as one of their user identification methods., This work was supported by the CHISTERA grant CHIST-ERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference nº PCI2023-145974-2) funded by the Agencia Estatal de Investigación through the PCI 2023 call., Peer Reviewed
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

TSGFM - Towards a graph foundation model for time series analysis in network monitoring

  • Latif Martínez, Hamid
  • Casas Hernández, Pedro
  • Suárez Varela, José
  • Cabellos Aparicio, Alberto
  • Barlet Ros, Pere
We present TSGFM, a Time Series Graph Foundation Model for network monitoring data analysis, based on spatiotemporal Graph Neural Networks (GNN). Inspired by the success of foundation models in achieving generalization and adaptation, TSGFM leverages pretraining on diverse multivariate timeseries (MTS) data from multiple domains to enable effective zero-shot analysis in network monitoring tasks. We compare TSGFM performance against five state-of-the-art AI/ML models in seven zero-shot forecasting scenarios, using five MTS datasets from different domains. Evaluations demonstrate that TSGFM achieves superior performance in six out of seven zero-shot testing scenarios. Most notably, in zero-shot network monitoring analysis, TSGFM surpasses all competing models by at least 18%, even without training on any network monitoring data., This work has been supported by the FWF Austrian Science Fund, Project reference I-6653, as well as by the Agencia Estatal de Investigación, Project reference PCI2023-145974- 2, as part of the EU CHIST-ERA-2022-SPiDDS-02 project GRAPHS4SEC. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

Atomic-level mechanisms of abnormal activation in NRAS oncogenes from two dimensional free energy landscapes

  • Hu, Zheyao
  • Martí Rabassa, Jordi
The NRAS-mutant subset of melanoma is one of the most aggressive and lethal types associated with poor overall survival. Unfortunately, a low understanding of the NRAS-mutant dynamic behavior has lead to the lack of clinically approved therapeutic agents able to directly target NRAS oncogenes. In this work, accurate local structures of NRAS and its mutants have been fully explored through the corresponding free energy surfaces obtained by microsecond scale well-tempered metadynamics simulations. Free energy calculations are crucial to reveal the precise mechanisms of Q61 mutations at the atomic level. Considering specific atom-atom distances and angles as appropriate reaction coordinates, we have obtained free energy surfaces revealing local and global minima together with their main transitions states, unveiling the mechanisms of abnormal NRAS activation from atomic-level and quantitatively analyzing the corresponding stable states. This will help to advance in our understanding of the basic mechanisms of NRAS mutations, offering new opportunities for the design of potential inhibitors., We are thankful for the financial support provided by the Spanish Ministry of Science, Innovation and Universities (funds MCIU/AEI/FEDER, EU). This publication is a part of the I + D + i project number PID2021-124297NB-C32, funded by MCIN/AEI/10.13039/501100011033/ and “FEDER Una manera de hacer Europa”. Zheyao Hu is a Ph.D. fellow from the Chinese Scholarship Council (grant 202006230070). J. M. thanks the Generalitat de Catalunya for the support through the grant Grup de Recerca SGR-Cat2021 Condensed, Complex and Quantum Matter Group reference 2021SGR-01411 and the Polytechnic University of Catalonia-Barcelona Tech through the funding AGRUPS. Computational resources awarded by the Barcelona Supercomputing Center-Spanish Supercomputing Network (grant BCV-2024-2-0005 “Simulation and modeling of biosystems with applications to medicine (1): drug design for the treatment of cancers of the RAS family”) are also acknowledged., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

AST-GNN: A graph neural network for web tracking detection

  • Campeny Roig, Eloi
  • Castell Uroz, Ismael
  • Barlet Ros, Pere
Web tracking technology is prevalent on the Internet today, with most websites employing user identification systems that can accurately identify users or devices behind browsers. While numerous works in literature attempt to create machine learning models for detecting these identification systems, many rely on features susceptible to obfuscation techniques and are only partially capable of identifying specific subsets of web tracking algorithms they were trained on. Additionally, classification is typically done over entire resources, making it difficult to distinguish between web tracking code and legitimate code within the same file. In this work, we propose AST-GNN, a graph neural network model applied to the abstract syntax tree structure of JavaScript files, which can predict portions of code used for tracking purposes. By focusing on the code structure, AST-GNN can detect various web tracking systems and it is robust against obfuscation techniques. Our results show that the system has an accuracy rate above 95% in identifying web tracking code snippets, with computation performance in the order of milliseconds, fast enough to be used in real-time., This work was supported by the CHISTERA grant CHIST-ERA-22- SPiDDS-02 corresponding to the GRAPHS4SEC project (reference nº PCI2023-145974-2) funded by the Agencia Estatal de Investigación through the PCI 2023 call. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

GAT-AD: Graph Attention Networks for contextual anomaly detection in network monitoring

  • Latif Martínez, Hamid
  • Suárez Varela, José
  • Cabellos Aparicio, Alberto
  • Barlet Ros, Pere
Network anomaly detection is essential to promptly detect and fix issues in the network. Particularly, detecting traffic anomalies enables the early detection of configuration errors, malicious activities, or equipment malfunctions that could lead to severe impact on the network. In this paper, we present GAT-AD, a Deep Learning-based anomaly detection solution for network monitoring systems, which integrates a custom neural network model based on Graph Attention Networks (GAT). Our solution monitors aggregated traffic on origin–destination flows and automatically defines contexts that group flows with similar past activity. The neural network model within GAT-AD can be efficiently trained in a self-supervised manner. We evaluate our solution against two state-of-the-art anomaly detection baselines also based on graph representations and Deep Learning, in two different datasets: (i) WaDi, which is a well-known dataset for anomaly detection in a distributed sensor network, and (ii) Abilene, where we inject synthetically-generated anomalies into a dataset with real-world traffic from a large-scale backbone network. The results show that GAT-AD outperforms the two anomaly detection baselines: in WaDi by 14.1% in recall and 10.07% in F1-score, and in the Abilene dataset by ˜ 17.5% recall with respect to the best baseline., This publication is part of the Spanish I+D+i project TRAINER-A (ref. PID2020-118011GB-C21), funded by MCIN/AEI/10.13039/501100011033. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA), Spain. This work was also supported by the CHISTERA grant CHIST-ERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference no PCI2023-145974-2) funded by the Agencia Estatal de Investigación, Spain through the PCI 2023 call. This work was also supported by the Spanish Ministry of Economic Affairs and Digital Transformation and the European Union – NextGeneration EU, in the framework of the Recovery Plan, Transformation and Resilience (PRTR) (Call UNICO I+D 5G 2021, refs. number TSI-063000-2021-3, TSI-063000-2021-38, and TSI-063000-2021-52). This work was partially supported by the ORIGAMI project, funded by the Smart Networks and Services Joint Undertaking (SNS JU), Spain under the European Union’s Horizon Europe research and innovation programme under Grant Agreement No 101139270., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

A first look into Utiq: next-generation cookies at the ISP level

  • Castell Uroz, Ismael
  • Barlet Ros, Pere
Online privacy has become increasingly important in recent years. While third-party cookies have been widely used for years, they have also been criticized for their poten-tial impact on user privacy. They can be used by advertisers to track users across multiple sites, allowing them to build detailed profiles of their behavior and interests. However, nowadays, many browsers allow users to block third-party cookies, which limits their usefulness for advertisers. In this paper, we take a first look at Utiq, a new way of user tracking performed directly by the ISP, to substitute the third-party cookies used until now. We study the main properties of this new identification methodology and their adoption on the 10K most popular websites. Our results show that, although still marginal due to the restrictions imposed by the system, between 0.7% and 1.2% of websites already include Utiq as one of their user identification methods., This work was supported by the CHISTERA grant CHIST-ERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference nº PCI2023-145974-2) funded by the Agencia Estatal de Investigación through the PCI 2023 call., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

BGP anomaly detection using the raw internet topology

  • Latif Martínez, Hamid
  • Paillissé Vilanova, Jordi
  • Barlet Ros, Pere
  • Cabellos Aparicio, Alberto
The Border Gateway Protocol (BGP) is central to the global connectivity of the Internet, enabling fast and efficient dissemination of routing information. Hence, detecting any anomaly concerning BGP announcements is of critical importance to ensure the continuous operation of Internet services. Typically, BGP anomaly detection algorithms have relied on features of the BGP messages, such as the average length of the AS_PATH attribute, the volume of messages, or the type of message (announcement or withdrawal). Even though these algorithms provide good performance, they do not take into account the Internet topology, that is, the graph of Autonomous Systems (AS) created by the BGP announcements. In addition, some of the existing algorithms can detect only specific types of anomalies, while others require retraining them to support new scenarios. In this paper we propose detecting BGP anomalies by leveraging the raw BGP topology graph, instead of manually curated features of the BGP messages. We implement a Machine Learning algorithm to process the entire BGP topology and evaluate it with real-world data from 4 well-known incidents. We compare our proposal against two state-of-the-art solutions and a classical method that use BGP features and features of the BGP topology, not the topology itself. Our results show that our solution obtains remarkable performance identifying the incidents. Finally, we test our model with regular data (non-anomalous) to prove that it can be used in a production scenario, with samples processed on the fly and guaranteeing a low false alarm rate., This work is part of the I+D+i project titled BLOSSOMS, grant PID2024-158530OB-I00, funded by MICIU/AEI/10.13039/501100011033/ and by ERDF/EU. This work is also partially funded by the Catalan Institution for Research and Advanced Studies (ICREA). This work was also supported by Grant PCI2023-145974-2 funded by MICIU/AEI/10.13039/501100011033 and co-funded by the European Union (GRAPHS4SEC project). Jordi Paillissé is funded by European Union-Next Generation EU, Ministry of Universities and Recovery, Transformation and Resilience Plan, through a call from Universitat Politècnica de Catalunya (Grant Ref. 2022UPC-MSC-93871)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

RouteNet-Gauss: Hardware-enhanced network modeling with machine learning

  • Güemes Palau, Carlos
  • Ferriol Galmés, Miquel
  • Paillissé Vilanova, Jordi
  • López Brescó, Albert
  • Barlet Ros, Pere
  • Cabellos Aparicio, Alberto
Network simulation is pivotal in network modeling, assisting with tasks ranging from capacity planning to performance estimation. Traditional approaches such as Discrete Event Simulation (DES) face limitations in terms of computational cost and accuracy. This paper introduces RouteNet-Gauss, a novel integration of a testbed network with a Machine Learning (ML) model to address these challenges. By using the testbed as a hardware accelerator, RouteNet-Gauss generates training datasets rapidly and simulates network scenarios with high fidelity to real-world conditions. Experimental results show that RouteNet-Gauss significantly reduces prediction errors by up to 95% and achieves a 488x speedup in inference time compared to state-of-the-art DES-based methods. RouteNet-Gauss’s modular architecture is dynamically constructed based on the specific characteristics of the network scenario, such as topology and routing. This enables it to understand and generalize to different network configurations beyond those seen during training, including networks up to 10x larger. Additionally, it supports Temporal Aggregated Performance Estimation (TAPE), providing configurable temporal granularity and maintaining high accuracy in flow performance metrics. This approach shows promise in improving both simulation efficiency and accuracy, offering a valuable tool for network operators., This work was supported in part by the I+D+i Project titled Beyond 5G network infrastructure Leveraging deep AI acrOSS ultra-high capacity Optical data transport and quantum physical Substrates (BLOSSOMS) funded by Ministerio de Ciencia, Innovación y Universidades /Agencia Estatal de Investigación (MICIU/AEI)/10.13039/501100011033 and European Regional Development Fund/European Union (ERDF/EU), under Grant PID2024-158530OB-I00; in part by Catalan Institution for Research and Advanced Studies (ICREA); and in part by MICIU/AEI/10.13039/501100011033 and the European Union (GRAPHS4SEC project) under Grant PCI2023-145974-2. The work of Carlos Guemes-Palau was supported by the AGAUR-FI ajuts through the Joan Oró of the Secretariat of Universities and Research of the Department of Research and Universities of the Generalitat of Catalonia and European Social Plus Fund under Grant 2023 F-1 00083., Peer Reviewed, Postprint (author's final draft)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

TSGFM - Graph neural networks for zero-shot time series forecasting in network monitoring

  • Latif Martínez, Hamid
  • Vanerio, Juan
  • Casas Hernández, Pedro
  • Suárez Varela, José
  • Cabellos Aparicio, Alberto
  • Barlet Ros, Pere
We present TSGFM, a Time Series Graph Foundation Model for zero-shot network monitoring, leveraging spatiotemporal Graph Neural Networks (GNNs) to extract transferable representations across diverse multivariate time series (MTS) domains. Pretrained on heterogeneous time series datasets, TSGFM enables generalization without task-specific fine-tuning, addressing core challenges in dynamic network environments. TSGFM is benchmarked across five real-world MTS datasets and seven zero-shot forecasting scenarios, outperforming five state-of-the-art baselines in six out of seven tasks. Most notably, in zero-shot network monitoring analysis, TSGFM surpasses all competing models by at least 18%, even without any prior exposure to network monitoring data. We further compare TSGFM against leading Time Series Foundation Models (TSFMs), including TimeGPT and TimesFM. TSGFM achieves performance on par with TimeGPT, occasionally surpassing it, and consistently outperforms TimesFM, while using significantly less pretraining data and relying on a much simpler architecture. A detailed analysis of TSGFM’s learned spatial attention patterns reveals domain-specific connectivity structures. In particular, lower attention weights in network monitoring tasks suggest that dense spatial graphs may be unnecessary, opening opportunities for efficient spatial pruning without sacrificing accuracy. This challenges prevailing assumptions favoring fully connected spatiotemporal GNNs. To foster transparency and reproducibility, we release the complete implementation of TSGFM as open source, as well as the tested datasets., This work has been supported by the FWF Austrian Science Fund, Project reference I-6653, as well as by the Agencia Estatal de Investigación, Project reference PCI2023-145974-2, as part of the EU CHIST-ERA-2022-SPiDDS-02 project GRAPHS4SEC. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Artículo científico (JournalArticle)

GraphCC: a practical graph learning-based approach to congestion control in datacenters

  • Bernárdez Gil, Guillermo
  • Suárez Varela, José
  • Shi, Xiang
  • Xiao, Shihan
  • Cheng, Xiangle
  • Barlet Ros, Pere
  • Cabellos Aparicio, Alberto
Congestion Control (CC) plays a fundamental role in optimizing traffic in Datacenter Networks (DCNs). Currently, DCNs implement two main CC protocols: DCTCP and DCQCN. Both protocols are based on Explicit Congestion Notification (ECN), where switches mark packets when they detect congestion. Nowadays, network experts carefully set ECN parameters to optimize the average network performance. However, today’s DCNs experience rapid and abrupt changes that severely affect the network state (e.g., dynamic workloads, incasts), which leads to under-utilization and sub-optimal performance. In this paper we present GraphCC, a framework for in-network CC optimization. GraphCC relies on Multi-agent Reinforcement Learning (MARL) and Graph Neural Networks (GNN), and is compatible with widely deployed ECN-based CC protocols. The proposed solution deploys distributed agents on switches that communicate with their neighbors to cooperate and optimize the global ECN configuration. In our evaluation, we test GraphCC with three real-world traffic workloads, focusing on its capability to accommodate scenarios unseen during training (e.g., traffic changes, failures). We compare GraphCC with a state-of-the-art MARL solution for ECN tuning, and observe that our method outperforms the state-of-the-art baseline in all evaluation scenarios, with improvements up to 20% in average Flow Completion Time, similar mean throughput (within 1%), and significant reductions in buffer occupancy (38.0-85.7%)., This publication is part of the Spanish I+D+i project TRAINER-A (reference PID2020-118011GB-C21), funded by MCIN/AEI/10.13039/ 501100011033, and the research has been partially funded by the Catalan Institution for Research and Advanced Studies (ICREA), the Secretariat for Universities and Research of the Ministry of Business and Knowledge of the Government of Catalonia, and the European Social Fund. This work is also partially supported by the ORIGAMI project, funded by the Smart Networks and Services Joint Undertaking (SNS JU) under the European Union’s Horizon Europe research and innovation programme under Grant Agreement No 101139270, as well as by the CHISTERA grant CHIST-ERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference PCI2023-145974-2), funded by the Agencia Estatal de Investigación through the PCI 2023 call., Peer Reviewed, Postprint (published version)
Recercat. Dipósit de la Recerca de Catalunya
Publicación Contribución de congreso (ConferenceOutput)

Graph neural networks for graph-level anomaly detection in network intrusion detection

  • Anyfantis, Georgios
  • Castell Uroz, Ismael
  • Barlet Ros, Pere
Network attacks are a growing problem in today's world. Their number is growing exponentially, and they become more sophisticated every year, accentuating the need for robust Network Intrusion Detection Systems. Graph Neural Networks have been proven to be very capable in detecting and categorising network attacks, making them ideal for detection in real-life deployments. However, they are limited to training on a small number of outdated labelled datasets. We propose a new method for anomaly detection through a graph heterogeneous autoencoder as a way to reduce reliance on labelled datasets. Our graph autoencoder reconstructs the structure and features of a graph representing the traffic connections in a network. We train the autoencoder with normal graphs and then derive embeddings to use with traditional anomaly detection based on Machine Learning. Our results indicate that our solution can create distinct normal and anomalous graph embeddings that can then be used to detect attacks with a high degree of accuracy., This work was supported by the CHISTERA grant CHISTERA-22-SPiDDS-02 corresponding to the GRAPHS4SEC project (reference nº PCI2023-145974-2) funded by the Agencia Estatal de Investigación through the PCI 2023 call. This work is also supported by the Catalan Institution for Research and Advanced Studies (ICREA Academia)., Peer Reviewed, Postprint (published version)