Biblio

Network slicing qualitatively transforms network infrastructures such that they have maximum flexibility in the context of ever-changing service requirements. While the agility of cloud native network functions (CNFs) demonstrates significant promise, virtualization and softwarization severely degrade the performance of such network functions. Considerable efforts were expended to improve the performance of virtualized systems, and at this stage 10 Gbps throughput is a real target even for container/VM-based applications. Nonetheless, the current performance of CNFs with state-of-the-art enhancements does not meet the performance requirements of next-generation 6G networks that aim for terabit-class throughput. The present pace of performance enhancements in hardware indicates that straightforward optimization of existing system components has limited possibility of filling the performance gap. As it would be reasonable to expect a single silver-bullet technology to dramatically enhance the ability of CNFs, an organic integration of various data-plane technologies with a comprehensive vision is a potential approach. In this paper, we show a future vision of system architecture for terabit-class CNFs based on effective harmonization of the technologies within the wide-range of network systems consisting of commodity hardware devices. We focus not only on the performance aspect of CNFs but also other pragmatic aspects such as interoperability with the current environment (not clean slate). We also highlight the remaining missing-link technologies revealed by the goal-oriented approach.

Today's Internet is built on decades-old networking protocols that lack scalability, reliability, and security. In response, the networking community has developed path-aware Internet architectures that solve these issues while simultaneously empowering end hosts. In these architectures, autonomous systems construct authenticated forwarding paths based on their routing policies. Each end host then selects one of these authorized paths and includes it in the packet header, thus allowing routers to efficiently determine how to forward the packet. A central security property of these architectures is path authorization, requiring that packets can only travel along authorized paths. This property protects the routing policies of autonomous systems from malicious senders.The fundamental role of packet forwarding in the Internet and the complexity of the authentication mechanisms employed call for a formal analysis. In this vein, we develop in Isabelle/HOL a parameterized verification framework for path-aware data plane protocols. We first formulate an abstract model without an attacker for which we prove path authorization. We then refine this model by introducing an attacker and by protecting authorized paths using (generic) cryptographic validation fields. This model is parameterized by the protocol's authentication mechanism and assumes five simple verification conditions that are sufficient to prove the refinement of the abstract model. We validate our framework by instantiating it with several concrete protocols from the literature and proving that they each satisfy the verification conditions and hence path authorization. No invariants must be proven for the instantiation. Our framework thus supports low-effort security proofs for data plane protocols. The results hold for arbitrary network topologies and sets of authorized paths, a guarantee that state-of-the-art automated security protocol verifiers cannot currently provide.

The Named-Data Networking (NDN) has emerged as a clean-slate Internet proposal on the wave of Information-Centric Networking. Although the NDN's data-plane seems to offer many advantages, e.g., native support for multicast communications and flow balance, it also makes the network infrastructure vulnerable to a specific DDoS attack, the Interest Flooding Attack (IFA). In IFAs, a botnet issuing unsatisfiable content requests can be set up effortlessly to exhaust routers' resources and cause a severe performance drop to legitimate users. So far several countermeasures have addressed this security threat, however, their efficacy was proved by means of simplistic assumptions on the attack model. Therefore, we propose a more complete attack model and design an advanced IFA. We show the efficiency of our novel attack scheme by extensively assessing some of the state-of-the-art countermeasures. Further, we release the software to perform this attack as open source tool to help design future more robust defense mechanisms.