Load balancing over heterogeneous networks with gossip-based algorithms

Franceschelli, Mauro; Giua, Alessandro; Seatzu, Carla

doi:10.1109/acc.2009.5160452

Cited by 16 publications

(6 citation statements)

References 26 publications

(37 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tsai et al [32] proposed a routing protocol for load balancing based on path energy and self-maintenance. Mauro [33] proposed a gossip-based distributed balancing algorithm for which the nodes need to have extensive information about their neighborhood. Most recently, Lee et al [34] applied DAG's (Direction Acyclic Graph) to different data collection cycles for load balancing so that energy efficiency was improved in both the time and the spatial domains.…”

Section: Related Workmentioning

confidence: 99%

Distributed probabilistic routing for sensor network lifetime optimization

Wang

Tan

2015

Wireless Netw

View full text Add to dashboard Cite

A probabilistic and distributed routing approach for multi-hop sensor network lifetime optimization is presented in this paper. In particular, each sensor self-adjusts their routing probabilities locally to their forwarders based on its neighborhood knowledge, while aiming at optimizing the overall network lifetime (defined as the elapsed time before the first node runs out of energy). The theoretical feasibility and a practical routing algorithm are presented. Specifically, a sufficient distributed condition regarding the neighborhood state for distributed probabilistic routing to achieve the optimal network lifetime is presented theoretically. Based on it, a distributed adaptive probabilistic routing (DAPR) algorithm, which considered both the transmission scheduling and the routing probability evolvement is developed. We prove quantitatively that DAPR could lead the routing probabilities of the distributed sensors to converge to an optimal state which optimizes the network lifetime. Further, when network dynamics happen, such as topology changes, DAPR can adjust the routing probabilities quickly to converge to a new state for optimizing the remained network lifetime. We presented the convergence speed of DAPR. Extensive simulations verified its convergence and near-optimal properties. The results also showed its quick adaptation to both the network topology and data rate dynamics.

show abstract

Section: Related Workmentioning

confidence: 99%

Distributed probabilistic routing for sensor network lifetime optimization

Wang

Tan

2015

Wireless Netw

View full text Add to dashboard Cite

show abstract

“…To reach the consensuswhich is described as a heuristic assuming knowledge on the weights of tasks-, nodes communicate within a homogeneous architecture via gossiping. This work was lately extended to heterogeneous architectures in [10], where the authors analyze the convergence time. The study assumes that all tasks are initially in the system.…”

Section: Related Workmentioning

confidence: 99%

The sandpile scheduler

et al. 2014

View full text Add to dashboard Cite

This paper studies a self-organized criticality (SOC) model called sandpile for dynamically load-balancing tasks arriving in the form of Bag-of-Tasks (BoTs) in large-scale decentralized system. The sandpile is designed as a decentralized agent system characterizing a cellular automaton, which works in a critical state at the edge of chaos. Depending on the state of the cellular automaton, rather different responses may occur when a new task is assigned to a resource: it may change nothing or generate avalanches that reconfigure the state of the system. The abundance of such avalanches is in power-law relation with their sizes, a scale-invariant behavior that emerges without requiring tuning or control parameters. That means that large-catastrophicavalanches are very rare but small ones occur very often. Such an emergent pattern can be efficiently adapted for non-clairvoyant scheduling, where tasks are load balanced into computing resources trying to maximize the performance but without assuming any knowledge on the tasks features. In order to validate the algorithm design, we conduct an empirical experimentation showing that the sandpile is able to find near-optimal schedules by reacting differently to different conditions of workloads and architectures.

show abstract

“…It has roots in parallel computation [2] and has been addressed in [3], [9], [14], and [13]. Load balancing is a constrained version of the consensus problem, and there exists many gossip-based algorithms for distributed load balancing over different types of graphs with various constraints and objectives [5], [7]. More recently there has been a focus on quantized consensus to address the problems of finite capacity of communication channels and finite precision in computation.…”

Section: Introductionmentioning

confidence: 99%

Quantized distributed load balancing with capacity constraints

Gravelle

Martínez

2014

53rd IEEE Conference on Decision and Control

View full text Add to dashboard Cite

Current research in the field of distributed consensus algorithms fails to adequately address physical limitations of real systems. This paper proposes a new algorithm for quantized distributed load balancing over a network of agents subject to upper-limit constraints. More precisely, loads are integer values, and nodes are constrained to remain under maximum load capacities at all times. Convergence to a set of desired states is proven for all connected graphs, any feasible initial load distribution, and separation and connectivity conditions on nodes with small maximum capacities. Simulations illustrate our results. 53rd IEEE Conference on Decision and Control

show abstract

Load balancing over heterogeneous networks with gossip-based algorithms

Cited by 16 publications

References 26 publications

Distributed probabilistic routing for sensor network lifetime optimization

Distributed probabilistic routing for sensor network lifetime optimization

The sandpile scheduler

Quantized distributed load balancing with capacity constraints

Contact Info

Product

Resources

About