A generalization of multiple choice balls-into-bins

Park, Gahyun

doi:10.1145/1993806.1993862

Cited by 16 publications

(12 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Large jobs: For jobs that need multiple RUs, Tarcil uses batch sampling [28,29]. For m requested units, the scheduler samples R · m RUs and selects the m best among them as shown in Figure 5a.…”

Section: Scheduling With Guaranteesmentioning

confidence: 99%

“…As long as these agents rarely attempt to assign work to the same servers (infrequent conflicts), they proceed concurrently without delays. Sampling-based scheduling: Based on results from randomized load balancing [25,29], we can design samplingbased cluster schedulers [8,14,28]. Sampling the state of just a few servers reduces the latency of each scheduling decision and the probability of conflicts between concurrent agents, and is likely to find available resources in non heavily-loaded clusters.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Tarcil

Delimitrou

Sánchez

Kozyrakis

2015

Proceedings of the Sixth ACM Symposium on Cloud Computing

113

View full text Add to dashboard Cite

Scheduling diverse applications in large, shared clusters is particularly challenging. Recent research on cluster scheduling focuses either on scheduling speed, using sampling to quickly assign resources to tasks, or on scheduling quality, using centralized algorithms that search for the resources that improve both task performance and cluster utilization.We present Tarcil, a distributed scheduler that targets both scheduling speed and quality. Tarcil uses an analytically derived sampling framework that adjusts the sample size based on load, and provides statistical guarantees on the quality of allocated resources. It also implements admission control when sampling is unlikely to find suitable resources. This makes it appropriate for large, shared clusters hosting shortand long-running jobs. We evaluate Tarcil on clusters with hundreds of servers on EC2. For highly-loaded clusters running short jobs, Tarcil improves task execution time by 41% over a distributed, sampling-based scheduler. For more general scenarios, Tarcil achieves near-optimal performance for 4× and 2× more jobs than sampling-based and centralized schedulers respectively.

show abstract

Section: Scheduling With Guaranteesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Tarcil

Delimitrou

Sánchez

Kozyrakis

2015

Proceedings of the Sixth ACM Symposium on Cloud Computing

113

View full text Add to dashboard Cite

show abstract

“…Just as M.Mitzenmacher researched, with high probability, the maximum load in any node is approximately log n log log n . Suppose that each node is placed sequentially into the least full of d nodes chosen independently and uniformly at random, the maximum load is then only log log n log d + O(1) with high probability [12]. The key challenge with a decentralized design is providing a perfect policy to avoid confliction, given that concurrently load balancing scheme may make conflicting task scheduling decisions among nodes.…”

Section: B Self-similaritymentioning

confidence: 99%

DLBS: Decentralized load balancing scheme for event-driven cloud frameworks

Zhou

Sun

et al. 2014

2014 20th IEEE International Conference on Parallel and Distributed Systems (ICPADS)

View full text Add to dashboard Cite

With the development of cloud computing, more and more applications are moving to a distributed fashion to solve problems. These applications usually contain complex iterative or incremental procedures and have a more urgent requirement on low-latency. Thus many event-driven cloud frameworks are proposed. To optimize this kind of frameworks, an efficient strategy to minimize the execution time by redistributing workloads is needed. Nowadays, load balance is a critical issue for the efficient operation of cloud platforms and many centralized schemes have already been proposed. However, few of them have been designed to support event-driven frameworks. Besides, as the cluster size and volume of tasks increases, centralized scheme will lead to a bottleneck of master node. In this paper, we demonstrate a decentralized load balancing scheme named DLBS for event-driven cloud frameworks and present two technologies to optimize it. In our design, schedulers are placed in every node for independently load-monitoring, autonomous decision-making and parallel task-scheduling. With the help of DLBS, master frees from the burden and tasks are executed with lower latency. We analyze the excellence of DLBS theoretically and proof it through simulation. At last, we implement and deploy it on a 64-machine cluster and demonstrate that it performs within 20% of an ideal scheme, which are consistent with simulation results.

show abstract

“…In our prior work [13] 2 , we posed the following questions: If we place 2 balls at a time into 2 least loaded out of 3 bins chosen i.u.r., is the maximum load still Θ(ln ln n)?. What would occur if 5 balls are placed into 8 bins or 99 balls into 100 bins?…”

Section: Introductionmentioning

confidence: 99%

A Generalization of Multiple Choice Balls-into-Bins: Tight Bounds

Park

2016

Algorithmica

Self Cite

View full text Add to dashboard Cite

This paper investigates a general version of the multiple choice model called the (k, d)-choice process in which n balls are assigned to n bins. In the process, k < d balls are placed into the k least loaded out of d bins chosen independently and uniformly at random in each of n k rounds. The primary goal is to derive tight bounds on the maximum bin load for (k, d)-choice for any 1 ≤ k < d ≤ n. Our results enable one to choose suitable parameters k and d for which the (k, d)-choice process achieves the optimal tradeoff between the maximum bin load and message cost: a constant maximum load and O(n) messages. The maximum load for a heavily loaded case where m > n balls are placed into n bins is also presented for the case d ≥ 2k. Potential applications are discussed such as distributed storage as well as parallel job scheduling in a cluster.

show abstract

A generalization of multiple choice balls-into-bins

Cited by 16 publications

References 10 publications

Tarcil

Tarcil

DLBS: Decentralized load balancing scheme for event-driven cloud frameworks

A Generalization of Multiple Choice Balls-into-Bins: Tight Bounds

Contact Info

Product

Resources

About