Resumen de Novel architectures and strategies for security offloading
Internet has become an indispensable and powerful tool in our modern society. Its ubiquitousness, pervasiveness and applicability have fostered paradigm changes around many aspects of our lives. This phenomena has positioned the network and its services as fundamental assets over which we rely and trust. However, Internet is far from being perfect. It has considerable security issues and vulnerabilities that jeopardize its main core functionalities with negative impact over its players. Furthermore, these vulnerabilities¿ complexities have been amplified along with the evolution of Internet user mobility.
In general, Internet security includes both security for the correct network operation and security for the network users and endpoint devices. The former involves the challenges around the Internet core control and management vulnerabilities, while the latter encompasses security vulnerabilities over end users and endpoint devices. Similarly, Internet mobility poses major security challenges ranging from routing complications, connectivity disruptions and lack of global authentication and authorization.
The purpose of this thesis is to present the design of novel architectures and strategies for improving Internet security in a non-disruptive manner. Our novel security proposals follow a protection offloading approach. The motives behind this paradigm target the further enhancement of the security protection while minimizing the intrusiveness and disturbance over the Internet routing protocols, its players and users. To accomplish such level of transparency, the envisioned solutions leverage on well-known technologies, namely, Software Defined Networks, Network Function Virtualization and Fog Computing.
From the Internet core building blocks, we focus on the vulnerabilities of two key routing protocols that play a fundamental role in the present and the future of the Internet, i.e., the Border Gateway Protocol (BGP) and the Locator-Identifier Split Protocol (LISP). To this purpose, we first investigate current BGP vulnerabilities and countermeasures with emphasis in an unresolved security issue defined as Route Leaks. Therein, we discuss the reasons why different BGP security proposals have failed to be adopted, and the necessity to propose innovative solutions that minimize the impact over the already deployed routing solution. To this end, we propose pragmatic security methodologies to offload the protection with the following advantages: no changes to the BGP protocol, neither dependency on third party information nor on third party security infrastructure, and self-beneficial. Similarly, we research the current LISP vulnerabilities with emphasis on its control plane and mobility support. We leverage its by-design separation of control and data planes to propose an enhanced location-identifier registration process of end point identifiers. This proposal improves the mobility of end users with regards on securing a dynamic traffic steering over the Internet.
On the other hand, from the end user and devices perspective we research new paradigms and architectures with the aim of enhancing their protection in a more controllable and consolidated manner. To this end, we propose a new paradigm which shifts the device-centric protection paradigm toward a user-centric protection. Our proposal focus on the decoupling or extending of the security protection from the end devices toward the network edge. It seeks the homogenization of the enforced protection per user independently of the device utilized. We further investigate this paradigm in a mobility user scenario. Similarly, we extend this proposed paradigm to the IoT realm and its intrinsic security challenges. Therein, we propose an alternative to protect both the things, and the services that leverage from them by consolidating the security at the network edge. We validate our proposal by providing experimental results from prof-of-concepts implementations.
