arXiv subjects

Saman Feghhi

Publications and source records attributed to Saman Feghhi.

5 recordsLinked to original sources

arXiv2021

A Safe Reinforcement Learning Architecture for Antenna Tilt Optimisation

Safe interaction with the environment is one of the most challenging aspects of Reinforcement Learning (RL) when applied to real-world problems. This is particularly important when unsafe actions have a high or irreversible negative impact on the environment. In the context of network management operations, Remote Electrical Tilt (RET) optimisation is a safety-critical application in which exploratory modifications of antenna tilt angles of base stations can cause significant performance degradation in the network. In this paper, we propose a modular Safe Reinforcement Learning (SRL) architecture which is then used to address the RET optimisation in cellular networks. In this approach, a safety shield continuously benchmarks the performance of RL agents against safe baselines, and determines safe antenna tilt updates to be performed on the network. Our results demonstrate improved performance of the SRL agent over the baseline while ensuring the safety of the performed actions.

cs.LG↗

arXiv2021

Remote Electrical Tilt Optimization via Safe Reinforcement Learning

Remote Electrical Tilt (RET) optimization is an efficient method for adjusting the vertical tilt angle of Base Stations (BSs) antennas in order to optimize Key Performance Indicators (KPIs) of the network. Reinforcement Learning (RL) provides a powerful framework for RET optimization because of its self-learning capabilities and adaptivity to environmental changes. However, an RL agent may execute unsafe actions during the course of its interaction, i.e., actions resulting in undesired network performance degradation. Since the reliability of services is critical for Mobile Network Operators (MNOs), the prospect of performance degradation has prohibited the real-world deployment of RL methods for RET optimization. In this work, we model the RET optimization problem in the Safe Reinforcement Learning (SRL) framework with the goal of learning a tilt control strategy providing performance improvement guarantees with respect to a safe baseline. We leverage a recent SRL method, namely Safe Policy Improvement through Baseline Bootstrapping (SPIBB), to learn an improved policy from an offline dataset of interactions collected by the safe baseline. Our experiments show that the proposed approach is able to learn a safe and improved tilt update policy, providing a higher degree of reliability and potential for real-world network deployment.

cs.LG↗

arXiv2018

An Efficient Web Traffic Defence Against Timing-Analysis Attacks

We introduce a new class of lower overhead tunnel that is resistant to traffic analysis. The tunnel opportunistically reduces the number of dummy packets transmitted during busy times when many flows are simultaneously active while maintaining well-defined privacy properties. We find that the dummy packet overhead is typically less than 20% on lightly loaded links and falls to zero as the traffic load increases i.e. the tunnel is capacity-achieving. The additional latency incurred is less than 100ms. We build an experimental prototype of the tunnel and carry out an extensive performance evaluation that demonstrates its effectiveness under a range of network conditions and real web page fetches.

cs.NI↗

arXiv2016

A Web Traffic Analysis Attack Using Only Timing Information

We introduce an attack against encrypted web traffic that makes use only of packet timing information on the uplink. This attack is therefore impervious to existing packet padding defences. In addition, unlike existing approaches this timing-only attack does not require knowledge of the start/end of web fetches and so is effective against traffic streams. We demonstrate the effectiveness of the attack against both wired and wireless traffic, achieving mean success rates in excess of 90%. In addition to being of interest in its own right, this timing-only attack serves to highlight deficiencies in existing defences and so to areas where it would be beneficial for VPN designers to focus further attention.

cs.NI↗

arXiv2016

Proportional Fair Rate Allocation for Private Shared Networks

In this paper, we consider fair privacy in a shared network subject to traffic analysis attacks by an eavesdropper. We initiate the study of the joint trade-off between privacy, throughput and delay in such a shared network as a utility fairness problem and derive the proportional fair rate allocation for networks of flows subject to privacy constraints and delay deadlines.

cs.NI↗