Computational Models for Defenses against Internet-based Attacks

 

(Dissertation research page)

Last revision: 05/07/2005

 

 

Li-Chiou Chen

 

Proposal (NSF ITR 0218466):  Project summary

Li-Chiou Chen. (2003). “Computational Models for Defenses against Internet-based Attacks,” unpublished PhD dissertation, August 2003, Department of Engineering and Public Policy, Carnegie Mellon University, PDF(901KB).

 

Abstract

Internet-based attacks have become an important concern to the government and business since more systems are reliant upon the Internet to exchange information.  In particular, distributed denial of service (DDOS) attacks have been used as a prevalent way to compromise the availability of networks or information services. The economic incentives of Internet Service Providers (ISPs) to provide DDOS defenses and the public policy concerns to deploy these defenses have not been formally investigated previously. 

Security services, such as Virtual Private Networks, have been provided by ISPs as optional network services to deal with the secrecy of data transportation. In the case of DDOS attacks, ISPs provide DDOS defenses that ensure the availability of the subscribers’ online services. This dissertation proposes that ISPs provide DDOS defenses on their network as security services to their subscribers and studies the service models for providing the defenses and the public policies needed to facilitate the provision of the defenses. The focus will be on the DDOS defenses that actively filter out ongoing attack traffic.

This dissertation analyzes how the side effects of defenses influence the provision of the defenses and investigates the economic incentives for the service provision. The contributions of this dissertation are as follows: First, this dissertation categorizes the current defenses that actively respond against DDOS attacks at network routers. The characterization is based on attack detection algorithms and attack responses.  Secondly, the service provision model is analyzed based on the performance efficiency of DDOS defenses under various network topologies and various settings in the technology. When providing defenses which are congestion-based and are dynamically enforced, ISPs should design services that focus on adjusting the filtering rate of the attack traffic to meet the needs of different subscribers. When providing defenses that are anomaly-based and are statically enforced, ISPs should design services that focus on the false positive rate of attack detection. Next, the economic incentives for ISPs to offer defense services are then analyzed based on empirical data. To operate the DDOS defense services cost effectively, ISPs should set the filter location closer to the attack sources and price subscribers based on their willingness to pay.  Finally, cooperation among multiple ISPs on providing the defenses is analyzed. In order to improve the quality of the defenses when attacks are distributed, ISPs should cooperate with other highly influential ISPs.  Public policies should encourage source filtering and provide incentives for highly influential ISPs to deploy DDOS defenses.

 

Chapter Outlines

 

 

Acknowledgement

This work is supported in part by the National Science Foundation ITR 0218466, the National Science Foundation IGERT 9354995 and the Pennsylvania Infrastructure Technology Alliance, a partnership of Carnegie Mellon, Lehigh University, and the Commonwealth of Pennsylvania's Department of Economic and Community Development. Additional support was provided by ICES (the Institute for Complex Engineered Systems) and CASOS – the Center for Computational Analysis of Social and Organizational Systems at Carnegie Mellon University.  The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of the National Science Foundation, the Commonwealth of Pennsylvania or the U.S. government.

Committee Members

Prof. Kathleen Carley (chair), CS/EPP/CASOS

Prof. Benoit Morel, Engineering and Public Policy

Prof. David Krackhardt, Heinz School of Public Policy and Management

Dr. Thomas Longstaff, Software Engineering Institute