Load balancing in large networks: a comparative study

Lüling, Reinhard; Monien, Burkhard; Ramme, Friedhelm

doi:10.1109/spdp.1991.218196

Cited by 44 publications

(18 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Almost all research has focused on algorithms for specific topologies andÂor rely on global routing phases. A class of such research has involved performance analysis of load balancing algorithms by simulations [LMR91]. Among analytical results, load balancing on specific topologies under statistical assumptions on input load distributions has been studied [HCT89].…”

Section: Past Workmentioning

confidence: 99%

Dynamic Load Balancing by Random Matchings

Ghosh

Muthukrishnan

1996

Journal of Computer and System Sciences

View full text Add to dashboard Cite

The fundamental problems in dynamic load balancing and job scheduling in parallel and distributed networks involve moving load between processors. In this paper we consider a new model for load movement in synchronous machines. In each step of our model, load can be moved across only a matching set of communication links but across each link any amount of load can be moved. We present an efficient local algorithm for the dynamic load balancing problem under our model of load movement. Our algorithm works on networks of arbitrary topology under possible failure of links. The running time of our algorithm is related to the eigenstructure of the underlying graph. We also present experimental results analyzing issues in load balancing related to our algorithms. ]

show abstract

Section: Past Workmentioning

confidence: 99%

Dynamic Load Balancing by Random Matchings

Ghosh

Muthukrishnan

1996

Journal of Computer and System Sciences

View full text Add to dashboard Cite

show abstract

“…In one traffic micro simulation [19], the implemented method corresponds, to a local decision, local migration ( L D L M s , see [22]) strategy applied to the network nodes. Incident edges are transferred or split accordingly.…”

Section: Dynamic Load-balancingmentioning

confidence: 99%

Large scale traffic simulations

Nagel¹,

Rickert

Barrett

1997

View full text Add to dashboard Cite

DECLAIMERThis report was prepared as an account of work spotrsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, make any warranty, express or implied, or as~mes any legal Iiability or responsibility for the accuracy, ampleten-or usefulness of any information, apparatus, product, or process disdosed, or represents that its use would not infringe privateIy. Abstract. Large scale microscopic (i.e. vehicle-based) traffic simulations pose high demands on computational speed in at least two application areas: (i) real-time traffic forecasting, and (E) long-term planning applications (where repeated "looping" between the microsimulation and the simulated planning of individual person's behavior is necessary). As a rough number, a real-time simulation of an area such as Los Angeles (ca. 1 million travellers) will need a computational speed of much higher than 1 million aparticle" (= vehicle) updates per second. This paper reviews how this problem is approached in different projects and how these approaches are dependent both on the specific questions and on the prospective user community. The approaches reach from highly parallel and vectorizable, single-bit implementations on parallel supercomputers for Statistical Physics questions, via more realistic implementations on coupled workstations, to more complicated driving dynamics implemented again on parallel supercomputers.

show abstract

“…We use Poisson-distributed synthetic loads and compare the performance with other methods such as Random [7], Gradient [14,15], Sender Initiated [8], Receiver Initiated [8], Adaptive Contracting [8], and Tree Walking [20], as summarized in Section 2. Our experiments demonstrate that the quality of load balancing achieved by the SBN approach compares favorably with respect to four metrics: (i) message traffic per processor, (ii) total jobs transferred, (iii) maximum variance in processor idle time, and (iv) total completion time.…”

Section: Introductionmentioning

confidence: 99%