About Me

I am a Computer Science Ph.D. at Universidad Autónoma de Madrid (U.A.M) researching high performance data analysis solutions for network traffic analysis as well as anomaly detection methodologies. I received my B.Sc (2013), M.Sc (2014) and Ph.D. (2017) degrees in Computer Science Engineering from U.A.M.
In 2012 I joined the High Performance Computing and Networking Research Group at the same university in the Escuela Politécnica Superior (E.P.S), first as a student and later as a researcher as part of the Network of Excellence InterNet Science, where I collaborated on different European research projects. I continued my work as technical research staff for the project Fed4Fire, TRÁFICA, and dReDBox among others during my Ph.D.
Currently, I work for Naudit applying my research on projects with different enterprises. My current research topics include data analysis, log collection at high performance speeds and network traffic analysis for anomaly detection. I've also been a teaching assistant on different subjects at the E.P.S.

Contact Details

Carlos Vega
Lab. C-113
Universidad Autónoma de Madrid
C/ Francisco Tomas y Valiente, 11
28049 Madrid, Spain
(+34) 653 88 11 07


Universidad Autónoma de Madrid

Ph.D. in Computer Science 2014 - 2017 Excellent Cum Laude &Industrial Mention

My Ph.D. thesis revolves around the new challenges affecting the ETL stages involved in the network traffic analysis, contributing on each stage, from the data collection to the data visualization through the intermediate processing and analysis processes.
This thesis proposes a solution for HTTP traffic dissection at multi-Gb/s traffic rates [1] [2] and distribute the work load evenly for the sake of concurrency. In the same way, this work presents a centralized system for log collection [3], for the upcoming needs of the heterogeneous networks. Furthermore, a high level automated report system is proposed, along with new techniques for high performance processing of enriched records [4] and anomaly detection, among others. In this context, the implications and challenges of dissagregated architectures [5] have been studied together with their impact on data analysis tasks. Lastly, all this knowledge has been consolidated for the deployment of a real-time monitoring project on an enterprise network, developing techniques for anomaly detection and alarm generation.
This work can be found in Spanish and downloaded here. . The work is an extension of the papers published during my doctoral period, which can be found below.

M.Sc in Computer Science 2013 - 2014

My master's thesis addressed the problem of HTTP traffic dissection at high speeds such as 10 Gbps traffic rates. The project covered the design and development process of a web traffic processing tool which achieves aforementioned processing throughputs on commodity hardware. It has been used in different projects of our laboratory satisfying the needs of real problems in South America, Spain or Baréin, among others. The tool distinguish itself from others by the performance offered in the analysis speed, the resources management and the functionalities as well as the portability. Is the natural continuation of my previous work, adding comparatives against other SoA tools and making a deeper work on the abstract structures used for the achievement of high speed throughputs.
This work can be found in Spanish at the following site and downloaded here.

B.Sc in Computer Science 2009 - 2013

My degree project revolves around the performance evaluation of web traffic and gives both an example traffic processing tool for web traffic as well as a web interface for the representation of the analyzed traffic. The main motivation of this dissector’s development is being able to analize the delays between responses and their HTTP requests in a particular network, online or offline by using traffic files, and consecuently evaluate the behaviour of the analized network and make fixes to improve the web traffic performance. This degree project also assesses an specific representation of the gotten data by using maps and high performance graphs that are able to draw a huge number of points without an interface performance loss.
This work is available in Spanish at the following site and downloaded here.

Recent Work

Automated Traffic Analysis Reports

Work in Progress 2017

FERMIN (Factual Executive Report of a Monitored IP Network) is system for automatic reporting, whose design philosophy was to provide the network manager with a relevant report about the most interesting events in the least possible processing time, and with limited processing requirements. This novel traffic analysis solution for the generation of automated reports is aimed at improving proactive traffic monitoring in large IT infrastructures. More info.
See the note in the Conferences section for information about recent publications.

Multi-Gbps HTTP Traffic Analysis in commodity hardware based on local knowledge of TCP streams

Computer Networks September 2016 - January 2017

Novel techniques for performance evaluation of web traffic (response time, response code, etc.), with no reassembly of the underlying TCP connection, which severely restricts the traffic analysis throughput. Furthermore, sub-TCP connection load balancing techniques that significantly increase throughput. Such techniques provide performance evaluation statistics which are indistinguishable from the single-threaded alternative with full TCP connection reassembly. Paper available online at Computer Networks and arXiv .

Loginson: a transform and load system for very large scale log analysis in large IT infrastructures.

Journal of Supercomputing June 2016 - March 2017

A high performance log centralization system for large-scale log collection and processing. Besides log collection, Loginson provides high level analytics through a visual interface for the purpose of troubleshooting critical incidents. Paper available online at The Journal of Supercomputing and arXiv .


HPCC 2017

International Conferences on High Performance Computing and Communications December 2017 Bangkok, Thailand

An article about the dRedBox project has been presented (Dec., 2017) in the International Conferences on High Performance Computing and Communications (HPCC 2017) with FERMIN as use case of the proposed system. The paper is entitled "Diluting the Scalability Boundaries: Exploring the Use of Disaggregated Architectures for High-Level Network Data Analysis" and is authored by Carlos Vega, Jose Fernando Zazo, Hugo Meyer, Ferad Zyulkyarov, Sergio Lopez-Buedo and Javier Aracil. The accepted manuscript is available online at IEEE Xplore and arXiv .

JITEL 2017

XIII Jornadas de Ingeniería Telemática September 2017 Valencia, Spain

Presentation of the article "FlexiTop: sistema escalable y flexible de medidas de calidad para servicios Over-The-Top" authored by Daniel Perdices et al. during the "XIII Jornadas de Ingeniería Telemática" hosted in Valencia, Spain. This article is authored by: Daniel Perdices, Jorge E. Lopez de Vergara, Paula Roquero, Carlos Vega and Javier Aracil.

CEDI 2016

Jornadas Sarteco September 2016 Salamanca, Spain

Presentation of the article "HPSEngine: Motor de alto rendimiento y baja latencia para el procesamiento distribuido en tiempo real" by Rafael Leira et al. during the "Jornadas Sociedad de Arquitectura y Tecnología de Computadores" (SARTECO) celebrated in Salamanca as a part of the Spanish Conference of Computer Science (CEDI) 2016. This article is authored by: Rafael Leira, Paula Roquero, Carlos Vega, Iván González, Javier Aracil. [pp. 157-165]

Internet Conference EINS

2nd International Conference on Internet Science May 2015 Brussels, Belgium

Presentation of the Poster () from the project EINSrepo patronised by the European Comission and the European project EINS (European Network of Excellence in Internet Science) developed together with Sergio López-Buedo and with the collaboration of the Nexa Center for Internet & Society from the Politecnico di Torino.


MATLAB Academic Tour 2016

Machine Learning and Lowcost hardware with MATLAB March 2016 U.A.M.

Introduction to MATLAB and Machine Learning with MATLAB and Simulink with low cost hardware.


Applied Data Science with Python

Applied Plotting, Charting & Data Representation in Python April 2017 Coursera

Online Course given by the University of Michigan. 4 weeks of study, 3 - 5 hours/week. Course certificate can be downloaded here .

Coursera Big Data Analytics

Introduction to Big Data Analytics February 2016 Coursera

Online Course given by the University of California, San Diego. 5 weeks of study, 3 - 5 hours/week. Course certificate can be downloaded here .

Hadoop Platform and Application Framework January 2016 Coursera

Online Course given by the University of California, San Diego. 5 weeks of study, 1 - 2 hours/week. Course certificate can be downloaded here .

Introduction to Big Data December 2015 Coursera

Online Course given by the University of California, San Diego. 3 weeks of study, 5 - 6 hours/week. Course certificate can be downloaded here .

Introduction to M.L. with Functional Data

Aprendizaje Automático con Datos Funcionales July 2016 U.A.M.

Course given by Ph.D. José Luis Torrecilla from the Group of Machine Learning from the E.P.S.


Disaggregated Recursive Datacentre-in-a-Box

With Naudit HPCN Project duration: 2016-2018 H2020

dRedBox aims to deliver a full-fledged, vertically integrated datacentre-in-a-box prototype to showcase the superiority of disaggregation in terms of scalability, efficiency, reliability, performance and energy reduction. See the note in the Conferences section for information about recent publications.

Traffic analysis for operational intelligence

With Universidad Autónoma de Madrid Project duration: 2016 - 2019

The TRAFICA project aims at advancing the state of the art in the area of operational intelligence. On the one hand, novel virtualized probes will be designed and implemented as a substitute to the physical counterpart, for traffic capture scenarios which have virtualized servers at hand or network equipment with virtualization capabilities.

Network of Excellence in Internet Science

With Universidad Autónoma de Madrid Project duration: 2012 - 2016

The goal of EINS is coordinating and integrating European research aimed at achieving a deeper multidisciplinary understanding of the development of the Internet as a societal and technological artefact, whose evolution is increasingly interwined with that of human societies.

Federation for FIRE

With Universidad Autónoma de Madrid Project duration: 2012 - 2016

A federation of experimentation facilities will significantly accelerate Future Internet research. Fed4FIRE will deliver open and easily accessible facilities to the FIRE experimentation communities, which focus on fixed and wireless infrastructures, services and applications, and combinations thereof.

Get In Touch.

Through this form we can get in touch for whatever question you may have.

Error boy
Your message was sent, thank you!