| Michael Mitzenmacher, The Power of Two Choices in Randomized Load Balancing, Ph.D. thesis, UC Berkeley, 1996. |
....the cost of migrating tasks and maintaining load information, making these other policies suboptimal for such systems. Chapter 4 presents the CPU management policies that are tuned for large shared memory systems, and the experiments show that they perform close to optimal. Some researchers [1, 5, 44] have studied randomized algorithms for load balancing. These algorithms randomly select a small number of processors and choose the least loaded of these processors to run a task. Because these algorithms select processors randomly, they do not require any global load information, which can be ....
Michael Mitzenmacher. The power of two choices in randomized load balancing. Ph.D. Thesis, UC Berkeley, 1996.
....that the maximum load for d = 1 is about ln n= ln ln n (see e.g. Gon81] or [RS98] It came as quite a surprise when it was shown in [ABKU94] that for d = 2 the maximum load is exponentially smaller, namely about ln ln n= ln d. This phenomenon is often referred to as the power of two choices [Mit96b] There is a rather broad literature on this phenomenon. An early application of the power of two choices can be found in PRAM simulations on Distributed Memory Machines (see e.g. KLMadH92] DMadH93] Until now many di erent models for balls into bins games and related problems have been ....
Michael Mitzenmacher. On the Power of Two Choices in Randomized Load Balancing. PhD thesis, 1996.
.... is chosen [50, 78, 136] An alternative approach is to first probe a limited number of nodes at random, and then choose the best one [117, 136] Even probing only two nodes reduces the expected maximum load when mapping n threads from O(log n log log n) to O(log log n) 20] Mitzenmacher et al. [119] improvise on the work in [20] They consider the following dynamic model: customers arrive as a Poisson stream of rate at a collection of n servers. Each customer chooses some d servers independently and uniformly at random from the n servers, and waits for service at the one with the fewest ....
....process in finegrain systems. Executing a random function for every decision, and polling overheads due to the load probes pose the biggest challenge to make load balancing profitable in fine grain multithreaded systems such as EARTH. The basis for the Rand balancer is the supermarket model [119] in distributed computing. The resuts show that giving each ball two choices instead of just one leads to an exponential improvement in the maximum load on any node. The system considered has a very high number of queues to chose from, in the order of hundreds. While having more than a hundred ....
[Article contains additional citation context not shown here]
Michael David Mitzenmacher. The Power of Two Choices in Randomized Load Balancing. In Ph. D Thesis, University of California, Berkeley, California, 1996.
No context found.
Michael Mitzenmacher, The Power of Two Choices in Randomized Load Balancing, Ph.D. thesis, UC Berkeley, 1996.
No context found.
Michael Mitzenmacher. The Power of Two Choices in Randomized Load Balancing. IEEE Transactions on Parallel and Distributed Systems, 12(10):1094--1104, 2001. 9
No context found.
Michael Mitzenmacher. The power of two choices in randomized load balancing. IEEE Transactions on Parallel and Distributed Systems, 12, 2001. 10
No context found.
Michael Mitzenmacher. The Power of Two Choices in Randomized Load Balancing. PhD thesis, UC-Berkeley, 1996.
No context found.
Michael Mitzenmacher. The power of two choices in randomized load balancing. IEEE Transactions on Parallel and Distributed Systems, 12(10):1094--1104, 2001.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC