Raft is a state-of-the-art consensus algorithm for state replication over a distributed system of nodes. According to Raft, all state updates occurring anywhere in the system are forwarded to the leader, which is elected among the system nodes to collect and replicate these updates to all other nodes. Thus, the time required for the state replication, named as system response time, depends on the delays between the leader and all other nodes. After multiple node failures and leadership transitions, each node can be leader with a probability that affects the expected response time. The leadership probabilities, in turn, are affected by the random intervals that nodes are waiting, after detecting a leader failure and before competing for the successive leadership. The Raft designers suggest the ranges of these intervals to be equal for all nodes. However, this may result in increased expected response time. In this paper, mathematical models are presented for estimating the ranges resulting in the desired leadership probabilities. The presented theoretical results are also confirmed by testbed experimentation with an open-source and widely used Raft implementation.
Software defined networking (SDN) complements low power Internet of Things (IoT), since the former offers dynamicity and the latter is susceptible to environmental changes. The SDN controller placement refers to the selection of the IoT sensors running the controllers, while the switch assignment is the process of mapping each sensor to a controller. Both choices affect the volume of the control traffic, a significant metric in low power wireless IoT networks where bandwidth is scarce or energy consumption is important. In this paper, we model an optimization problem for minimum control traffic, assess its complexity and devise a set of heuristic algorithms for expediting its solution. We initially present a fast and simple heuristic algorithm, which is then extended to two iterative algorithms with even better performance at the cost of time complexity. Our simulations and testbed experimentation reveal close to optimal performance of all heuristic solutions with significantly less computation time than explicitly solving the optimization problem. At the end, we provide insights for further enhancements of these heuristics with a bias for minimum control delay.
The adoption of the Software Defined Networking (SDN) paradigm is experiencing rapid growth, extending beyond typical datacenter networks to encompass 5G and IoT deployments. This expansion is primarily driven by the inherent flexibility offered by SDN in terms of network management and control. Mobile Ad-Hoc Networks (MANETs) have not yet fully embraced the advantages of the SDN approach due to their specific requirements arising from the dynamic nature of mobility and the unpredictable conditions of the wireless medium. The progress of adoption has been impeded by technical obstacles and the inherent conflict between the centralized nature of SDN and the decentralized approach of conventional MANET protocols. This paper introduces our implementation of a comprehensive SDN scheme that aims to address the divide between different networking approaches by integrating the advantages of both worlds. Our prototype has the capability to integrate with 802.11 Ad-Hoc networks. It provides several important features: a) fault tolerance for the SDN controller, which involves recovering from failures by electing a new controller, b) dynamic network topology discovery, and c) the ability to deploy globally optimal routing policies, based on the Multicommodity Flow Problem (MCFP), and following a cross-layer approach that utilizes the 802.11 MAC layer statistics. The applicability and performance benefits of our framework were assessed through the evaluation of our proposed scheme in contrast to a traditional MANET routing protocol.
Hyperledger Fabric (HLF) is a secure and robust blockchain (BC) platform that supports high-throughput and low-latency transactions. However, it encounters challenges in managing conflicting transactions that negatively affect throughput and latency. This paper proposes a novel solution to address these challenges and improve performance, especially in applications incorporating extensive volumes of highly conflicting transactions. Our solution involves reallocating the Multi-Version Concurrency Control (MVCC) of the validation phase to a preceding stage in the transaction flow to enable early detection of conflicting transactions. Specifically, we propose and evaluate two innovative modifications, called Orderer Early MVCC (OEMVCC) and OEMVCC with Execution Avoidance (OEMVCC-EA). Our experimental evaluation results demonstrate significant throughput and latency improvements, providing a practical solution for high-conflict applications that demand high performance and scalability.