How does mobility help distributed systems compute?

Brains are composed of connected neurons that compute by transmitting signals. The neurons are generally fixed in space, but the communication patterns that enable information processing change rapidly. By contrast, other biological systems, such as ant colonies, bacterial colonies, slime moulds and immune systems, process information using agents that communicate locally while moving through physical space. We refer to systems in which agents are strongly connected and immobile as solid , and to systems in which agents are not hardwired to each other and can move freely as liquid . We ask how collective computation depends on agent movement. A liquid cellular automaton (LCA) demonstrates the effect of movement and communication locality on consensus problems. A simple mathematical model predicts how these properties of the LCA affect how quickly information propagates through the system. While solid brains allow complex network structures to move information over long distances, mobility provides an alternative way for agents to transport information when long-range connectivity is expensive or infeasible. Our results show how simple mobile agents solve global information processing tasks more effectively than similar systems that are stationary. This article is part of the theme issue ‘Liquid brains, solid brains: How distributed cognitive architectures process information’.

A Communication Probability-Based Mapping Algorithm for Mesh-Based Network-on-Chip Systems

Network-on-chip (NoC) mapping algorithms significantly affect NoC system performance in terms of communication cost and energy consumption. For a specific application represented by a task graph, this paper proposes an energy-efficient mapping algorithm that searches for the mapping decision with best communication locality and therefore lowest energy consumption. To this end, we formulate the concerned mapping problem as an optimization model, and propose an effective meta-heuristic algorithm to solve the formulated optimization model. During the mapping procedure, we employ a simulation-free, communication probability-based energy model to evaluate the quality of each candidate mapping. By iteratively updating the best explored mapping decision using a meta-heuristic search strategy, the mapping procedure can eventually identify an mapping decision with optimal energy efficiency in the search space. The proposed mapping algorithm has been verified on NoC systems of different sizes using a variety of benchmark applications. Simulation results demonstrate that the mapping decision produced by this algorithm achieves an up to 23% energy reduction compared with the traditional round-robin strategy.

Optimized Virtual Machine Placement with Traffic-Aware Balancing in Data Center Networks

Virtualization has been an efficient method to fully utilize computing resources such as servers. The way of placing virtual machines (VMs) among a large pool of servers greatly affects the performance of data center networks (DCNs). As network resources have become a main bottleneck of the performance of DCNs, we concentrate on VM placement with Traffic-Aware Balancing to evenly utilize the links in DCNs. In this paper, we first proposed a Virtual Machine Placement Problem with Traffic-Aware Balancing (VMPPTB) and then proved it to be NP-hard and designed a Longest Processing Time Based Placement algorithm (LPTBP algorithm) to solve it. To take advantage of the communication locality, we proposed Locality-Aware Virtual Machine Placement Problem with Traffic-Aware Balancing (LVMPPTB), which is a multiobjective optimization problem of simultaneously minimizing the maximum number of VM partitions of requests and minimizing the maximum bandwidth occupancy on uplinks of Top of Rack (ToR) switches. We also proved it to be NP-hard and designed a heuristic algorithm (Least-Load First Based Placement algorithm, LLBP algorithm) to solve it. Through extensive simulations, the proposed heuristic algorithm is proven to significantly balance the bandwidth occupancy on uplinks of ToR switches, while keeping the number of VM partitions of each request small enough.

Virtual Machine Migration in Cloud Computing Environments

Recent developments in virtualization and communication technologies have transformed the way data centers are designed and operated by providing new tools for better sharing and control of data center resources. In particular, Virtual Machine (VM) migration is a powerful management technique that gives data center operators the ability to adapt the placement of VMs in order to better satisfy performance objectives, improve resource utilization and communication locality, mitigate performance hotspots, achieve fault tolerance, reduce energy consumption, and facilitate system maintenance activities. Despite these potential benefits, VM migration also poses new requirements on the design of the underlying communication infrastructure, such as addressing and bandwidth requirements to support VM mobility. Furthermore, devising efficient VM migration schemes is also a challenging problem, as it not only requires weighing the benefits of VM migration, but also considering migration costs, including communication cost, service disruption, and management overhead. This chapter provides an overview of VM migration benefits and techniques and discusses its related research challenges in data center environments. Specifically, the authors first provide an overview of VM migration technologies used in production environments as well as the necessary virtualization and communication technologies designed to support VM migration. Second, they describe usage scenarios of VM migration, highlighting its benefits as well as incurred costs. Next, the authors provide a literature survey of representative migration-based resource management schemes. Finally, they outline some of the key research directions pertaining to VM migration and draw conclusions.