UPC research activities at CONTENT
The CONTENT Network of Excellence targets a key area of Information Society Technologies, namely Content Delivery Networks for Home Users, as an integral part of Networked Audio-Visual Systems and Home Platforms.
CONTENT aims to build the European Research Area in this important communication topic by integrating a group of experts with the purpose of taking forward the state of the art and increasing European leadership in Content Networks. The overall goal of the CONTENT Network-of-Excellence is to integrate the research efforts of the members to address the technical challenges at the different system levels to enable easy-to-install and easy-to-use AV services in and between homes. In particular, the main technical objective will be to boost the potential of European Community Networking by improving Content Distribution infrastructures for the delivery of live (streaming) content and interactive stored content, and by integrating, in an open way, tools and mechanisms that would enable the curation of multimedia assets and their subsequent access for the benefit of the communities of users, producing a set of appropriate services for them, both in the context of the “long tail” or applied to (re-purposed) assets created by traditional broadcasters.
We seek to integrate our research expertise on the different levels of (overlay) networks and services for AV content and put special emphasis on the new challenges that arise from community networks and future end-user services for AV content. This leads to three system planes that need to be addressed concurrently: (a) community networks: covering the home networking infrastructure and the access network; (b) overlay networks: proving and end-to-end view for AV services, and (c) content service networks providing a set of services and the composition of simple services to more complex ones. Furthermore, we address the fundamental tradeoffs between non-functional requirements, like efficiency vs. resilience within the planes and between the planes. This requires a well designed coordination between the system planes.In parallel, the consortium will perform exchange, education and training of personnel inside and outside the network, dissemination of research results and, in general, spreading of excellence.
The joint programme of activities aims at the creation of a self-standing virtual centre of excellence that should result in developing the necessary means for achieving a durable integration of the research capacities. The joint programme of activities includes those aimed at integration, as well as activities related to the spreading of excellence and dissemination of results outside the network. CONTENT will continue the idea initiated in E-NEXT of investing in the education of young researchers through the European Doctoral School of Advanced Topics In Networking (SATIN) to prepare the next generation of researchers for the European Research Area.
CONTENT members draw considerable experience from the current FP-6 NoE E-NEXT, and will achieve these objectives through its focused Joint Research Workpackage, harmonized by a network coordination committee, and interacting with network coordination, integration, spreading excellence and dissemination workpackages, with guidance from an Industry Advisory Board.
Research carried out at UPC
Modeling Botnet Spread
What are the critical topological parameters that determine the network robustness, i.e. the goal is to gain new insights in the form of quantitative measures that could be applied to improve the robustness of current complex networks. We are especially interested in malware behavior and protection strategies for computer networks
A botnet is a set of computers compromised by malicious code and under the control of the attacker - botmaster. The size of a botnet can range from several hundreds to 50 thousands computers in a botnet. Most of the botnets are hierarchically constructed having one command and control (C&C) server bot controller, usually Internet Relay Chat (IRC), which distributes botmaster’s commands. The problems with centralized communication lay in the central point of failure - a bot controller hijacking or disabling. The malware writers started to recognize the necessity of deploying a new distributed architecture. Several bots have appeared which use peer-to-peer (P2P) networks as a communication channel. Without a centralized point of failure, different strategies need to be applied in order to stop the bot from spreading and attacking. One of the possible strategies is index-poisoning attack. Defense consists of insertion of massive numbers of bogus records into the index for a set of targeted files.
The model of index poisoning defense strategy is established. The data that will be used to estimation the efficiency of defense strategy are being gathered.
1) We are going to investigate what is the influence of index poisoning attack on the spread of a botnet.
2) We will propose several models and defense techniques and to verify them with simulations of bot spread. Joined research related to distributed defense techniques (distributed Intrusion Detection Systems and Intrusion Prevention Systems) is to be discussed.
3) A plan for future joined research related to IDS and IPS is to be made
Contact Person at UPC: Loránd Jakab
UPC and TUDelft
On the Transition to the Mobile Internet
Our primary goal is to analyze the transition to the Mobile Internet, identify potential issues and propose solutions. This is transition is related with the technical evolution of the current Internet architecture to a fully enabled mobile Internet. With the Mobile Internet end hosts can change its point of attachment seamlessly without disrupting its network connections.
1.- We aim to apply the P2P-based solution to other problems. Specifically we are focusing our research on other service discovery mechanisms such as NAT traversal, servers and/or LISP. We plan to adapt our P2P technology to these issues and evaluate its performance.
2.- We aim to process the collected data traces to identify which network flows should be forwarded through the Home Agent, characterize the load of a generic Home Agent and evaluate the performance of novel Home Agent architectures using a realistic load.
3.- We aim to analyze and study specific applications of the available bandwidth and evaluate how this metric can improve other research topics. For instance we aim to study which would be the improvement on overlay routing or congestion control when the available bandwidth is known.
UPC, UC, Agilent Technologies (Agilent Laboratories Scotland) and UC3M
Security Threats in Peer-to-Peer Content Distribution Networks
Botnets are one of the biggest current threats to the security of the Internet, and CDNs are also affected. Recent advances in botnet technology point to an evolution towards peer-to-peer structures, and the ramifications of this for CDNs are still unknown. We propose to investigate botnets in the context of CDNs, starting with a recount of attacks possible in/from these networks.
We reviewed literature in several overlapping areas that are relevant to the security of peer-to-peer content distribution networks to identify known attacks to and from these type of platforms. These attacks were classified in a taxonomy based on several different criteria, such as the attackers awareness, attack source and motivation, affected target, policy, etc. This effort was a starting point to help understand the challenges in the security of CDNs and help research in this area. The resulting paper was submitted to a journal, but unfortunately not accepted, and we are currently working on restructuring it based on the reviewer's comments.
There is a planned short scientific mission planned with the following objectives:
• Review and discuss key related work, and reviewers comments; decide how to focus the rewritten paper to match the target journal
• Prepare the structure of the joint publication, write some parts and divide work for after the mission
• Familiarize myself with the laboratory environment and tools developed in-house at UoN, exploring possible future collaboration
UPC and CINI/UoN
Multihoming smart routing for bringing end-to-end QoS
Study of end-to-end performance/QoS benefits through Intelligent (or smart) Route Control (IRC)
Study of scalable and stable route control strategies for competitive environments – Cooperative Route Control – Social Route Control 2009 Joint Research Goal Basically, our main goal is to give further details about an IRC solution: Offering the insights of a step-by-step design of an IRC solution.
New evaluation results showing the IRC solution feasibility in terms of performance/QoS benefits and route stability for other performance metrics (e.g., jitter, path available bandwidth)
Understanding of the potential interactions between Smart Routing (SR) and interdomain Traffic Engineering (TE)
Enrichment of the study about the interactions between SR and TE, more specifically: Conclusion of the implementation of a steady-state simulation model that will be able to simulate more realistically the behavior of a large scale network composed of ISP and stubs that employ TE and IRC techniques. This model will use different Internet Topologies inferred from BGP dumps of GEANT Network, depending of the fraction of traffic under control, and will include genetic TE algorithms and different IRC algorithms.
Investigation of how the granularity and timescale of traffic rule these interactions;
Integration of a Utility-based Smart Routing scheme into our cooperative framework proposed to tackle the interactions between SR and TE, in which IP Network Providers advertise agnostic routing preferences to Smart Routing, observing the TE objectives;
Evaluation of our cooperative framework using our steady-state simulator.
UPC and UC
General Project Website
More information can be found at http://www.ist-content.org/