Tight Load Balancing Via Randomized Local Search

Berenbrink, Petra; Kling, Peter; Liaw, Christopher; Mehrabian, Abbas

doi:10.1109/ipdps.2017.52

Cited by 4 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Upon activation, a token samples a random node and moves there if that node's load is smaller than the load at the token's current host node. It has recently been proved [2] that this process reaches perfect balance in O log n + log(n) · n 2 /m time (both in expectation and with high probability), which is asymptotically tight.…”

Section: Introductionmentioning

confidence: 88%

Tight & Simple Load Balancing

Berenbrink

Friedetzky

Kaaser

et al. 2019

2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

Self Cite

View full text Add to dashboard Cite

We consider the following load balancing process for m tokens distributed arbitrarily among n nodes connected by a complete graph: In each time step a pair of nodes is selected uniformly at random. Let 1 and 2 be their respective number of tokens. The two nodes exchange tokens such that they have ( 1 + 2 )/2 and ( 1 + 2 )/2 tokens, respectively. We provide a simple analysis showing that this process reaches almost perfect balance within O(n log n + n log ∆) steps, where ∆ is the maximal initial load difference between any two nodes. ACM Subject Classification Mathematics of computing → Probability and statistics → Stochastic processesKeywords and phrases load balancing, balls and bins, stochastic processes 1 The expression with high probability refers to a probability of 1 − n −Ω(1) . 2 We assume ∆ > 0 (such that log ∆ is well defined); otherwise the system is already perfectly balanced.

show abstract

Section: Introductionmentioning

confidence: 88%

Tight & Simple Load Balancing

Berenbrink

Friedetzky

Kaaser

et al. 2019

2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS)

Self Cite

View full text Add to dashboard Cite

show abstract

“…Secondly, for K ≥ 3 all entries above the diagonal are strictly positive, except for entry [[P ]] 0,K−1 = 0. This is verified in Appendix D. 4 The calculations for the analytical model of the transition matrix P have been carried out with Mathematica c and are exact. The empirical transition matrix for given values of N and K has been estimated by simulating, for each start state s 0 from 0 to K − 1, a number of one million one-step transitions and counting how often each possible successor state is assumed.…”

Section: Properties Of the Transition Matrixmentioning

confidence: 75%

“…Next, assuming that K and N are known, it will be important to identify good values for the parameters q O , q N and q I , so that in the case K ≤ N the average hitting time is minimized, whereas in the case K > N we want to allocate undisturbed transmission opportunities in a fair manner. It will also be interesting to transfer the algorithm to a scenario without central slotted time, where each node makes decisions asynchronously (see for example [4]), or to situations where a frequency can host a variable number of WBSNs subject to a constraint that the total load does not exceed a threshold.…”

Section: Discussionmentioning

confidence: 99%

“…We have validated the analytical model for the state transition matrix P against empirical transition matrices obtained from Monte-Carlo simulations. 4 In Figure 6 we compare the average hitting time for the simulation model (where the average is taken over 1,000,000 replications for given N and K) and the analytical model. In both cases we start with 0 settled balls.…”

Section: Markov Model For the Sticky Variantmentioning

confidence: 99%

See 1 more Smart Citation

Autonomous allocation of wireless body-sensor networks to frequency channels: modeling with repeated “balls-in-bins” experiments

Willig

2018

Wireless Netw

View full text Add to dashboard Cite

In this paper we model a situation where several wireless body sensor networks (WBSN) compete for occupation of a number of frequency channels. Each channel can host at most one WBSN with satisfactory performance and WBSNs have the ability to change their operating channel, subject to the constraint that they can only monitor the performance or occupancy of their current channel but not of any other channel. We consider a number of randomized schemes for changing the frequency channels and present and evaluate Markov chain models for these, building on a "balls-in-bins" approach.

show abstract

“…By default, the random sample size is D = 2 by using the Power of Two Choices (PoTC) algorithm [18]. Many network load balancing [19]- [21] and job scheduling [22] algorithms use PoTC for balanced resource distribution. The Good gateways are given priority for the selection (line 4).…”

Section: Step 3: Gateway Node Selectionmentioning

confidence: 99%

GateSelect: A novel Internet gateway selection algorithm for client nodes

Batbayar

Meseguer

Sadre

et al. 2020

2020 16th International Conference on Network and Service Management (CNSM)

View full text Add to dashboard Cite

The Internet gateway selection problem is becoming very important as the number of Internet-connected devices increases and stresses the limited number of Internet gateway nodes. The gateway nodes often experience frequent performance fluctuations, and the best gateway selection candidate changes frequently with growing network dynamics. We propose GateSelect, a customized selection algorithm for each client node that not only provides the best-effort selection candidate but also ensures the global, balanced distribution of the gateway nodes. We utilize over the counter, lightweight calculations to optimize the client-side selection algorithm by combining classification, short term performance prediction, and randomized selection. We compare our algorithm with several baseline algorithms, and the experiment results show that our proposal provides better performance and balanced distribution of gateway nodes.

show abstract

Tight Load Balancing Via Randomized Local Search

Cited by 4 publications

References 18 publications

Tight & Simple Load Balancing

Tight & Simple Load Balancing

Autonomous allocation of wireless body-sensor networks to frequency channels: modeling with repeated “balls-in-bins” experiments

GateSelect: A novel Internet gateway selection algorithm for client nodes

Contact Info

Product

Resources

About