Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems最新文献_第3页

Analyzing Location-Based Advertising for Vehicle Service Providers Using Effective Resistances 基于有效阻力的车辆服务提供商位置广告分析

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331484

Haoran Yu, Ermin Wei, R. Berry

引用次数: 3

Quantifying Data Locality in Dynamic Parallelism in GPUs gpu动态并行量化数据局部性

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331473

Xulong Tang, Ashutosh Pattnaik, Onur Kayiran, Adwait Jog, M. Kandemir, C. Das

引用次数: 3

Global Attraction of ODE-based Mean Field Models with Hyperexponential Job Sizes 具有超指数作业规模的基于ode的平均场模型的全局吸引力

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331476

B. Van Houdt

{"title":"Global Attraction of ODE-based Mean Field Models with Hyperexponential Job Sizes","authors":"B. Van Houdt","doi":"10.1145/3309697.3331476","DOIUrl":"https://doi.org/10.1145/3309697.3331476","url":null,"abstract":"Mean field modeling is a popular approach to assess the performance of large scale computer systems. The evolution of many mean field models is characterized by a set of ordinary differential equations that have a unique fixed point. In order to prove that this unique fixed point corresponds to the limit of the stationary measures of the finite systems, the unique fixed point must be a global attractor. While global attraction was established for various systems in case of exponential job sizes, it is often unclear whether these proof techniques can be generalized to non-exponential job sizes. In this paper we show how simple monotonicity arguments can be used to prove global attraction for a broad class of ordinary differential equations that capture the evolution of mean field models with hyperexponential job sizes. This class includes both existing as well as previously unstudied load balancing schemes and can be used for systems with either finite or infinite buffers. The main novelty of the approach exists in using a Coxian representation for the hyperexponential job sizes and a partial order that is stronger than the componentwise partial order used in the exponential case.","PeriodicalId":268462,"journal":{"name":"Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125517693","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 14

TeksDB: Weaving Data Structures for a High-Performance Key-Value Store TeksDB:用于高性能键值存储的编织数据结构

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331490

Youil Han, B. Kim, Jeseong Yeon, Sungjin Lee, Eunji Lee

引用次数: 4

Securing Distributed Gradient Descent in High Dimensional Statistical Learning 高维统计学习中的分布式梯度下降保护

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331499

Lili Su, Jiaming Xu

{"title":"Securing Distributed Gradient Descent in High Dimensional Statistical Learning","authors":"Lili Su, Jiaming Xu","doi":"10.1145/3309697.3331499","DOIUrl":"https://doi.org/10.1145/3309697.3331499","url":null,"abstract":"We consider unreliable distributed learning systems wherein the training data is kept confidential by external workers, and the learner has to interact closely with those workers to train a model. In particular, we assume that there exists a system adversary that can adaptively compromise some workers; the compromised workers deviate from their local designed specifications by sending out arbitrarily malicious messages. We assume in each communication round, up to q out of the m workers suffer Byzantine faults. Each worker keeps a local sample of size n and the total sample size is $N=nm$. We propose a secured variant of the gradient descent method that can tolerate up to a constant fraction of Byzantine workers, i.e., $q/m = O(1)$. Moreover, we show the statistical estimation error of the iterates converges in $O(łog N)$ rounds to $O(sqrtq/N + sqrtd/N )$, where d is the model dimension. As long as $q=O(d)$, our proposed algorithm achieves the optimal error rate $O(sqrtd/N )$. Our results are obtained under some technical assumptions. Specifically, we assume strongly-convex population risk. Nevertheless, the empirical risk (sample version) is allowed to be non-convex. The core of our method is to robustly aggregate the gradients computed by the workers based on the filtering procedure proposed by Steinhardt et al. citeSteinhardt18. On the technical front, deviating from the existing literature on robustly estimating a finite-dimensional mean vector, we establish a em uniform concentration of the sample covariance matrix of gradients, and show that the aggregated gradient, as a function of model parameter, converges uniformly to the true gradient function. To get a near-optimal uniform concentration bound, we develop a new matrix concentration inequality, which might be of independent interest.","PeriodicalId":268462,"journal":{"name":"Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131917008","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 7

Random Walk Based Sampling for Load Balancing in Multi-Server Systems 基于随机漫步的多服务器系统负载均衡抽样

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331506

Dengwang Tang, V. Subramanian

{"title":"Random Walk Based Sampling for Load Balancing in Multi-Server Systems","authors":"Dengwang Tang, V. Subramanian","doi":"10.1145/3309697.3331506","DOIUrl":"https://doi.org/10.1145/3309697.3331506","url":null,"abstract":"In multi-server systems, a classical job assignment algorithm works as follows: at the arrival of each job, pick d servers independently and uniformly at random and send the job to the least loaded server among the d servers. This model is known as the power-of-d choices algorithm. In this paper, we analyze a variant of this algorithm, where d servers are sampled through d independent non-backtracking random walks on a k-regular graph. The random walkers are periodically reset to independent uniform random positions. Under some assumptions on the underlying graph, we show that the system dynamics under this new algorithm converges to the solution of a deterministic ordinary differential equation (ODE), which is the same ODE as the classical power-of-d choices. We also show that the new algorithm stablizes the system, and the stationary distribution of the system converges to the stationary solution of the ODE. The new scheme can be considered as a derandomized version of power-of-d choices as it reduces the use of randomness while maintaining the performance of power-of-d choices.","PeriodicalId":268462,"journal":{"name":"Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114405386","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

What-If Analysis of Page Load Time in Web Browsers Using Causal Profiling 使用因果分析分析Web浏览器中的页面加载时间

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331483

Behnam Pourghassemi, Ardalan Amiri Sani, Aparna Chandramowlishwaran

{"title":"What-If Analysis of Page Load Time in Web Browsers Using Causal Profiling","authors":"Behnam Pourghassemi, Ardalan Amiri Sani, Aparna Chandramowlishwaran","doi":"10.1145/3309697.3331483","DOIUrl":"https://doi.org/10.1145/3309697.3331483","url":null,"abstract":"Web browsers have become one of the most commonly used applications for desktop and mobile users. Despite recent advances in network speeds and several techniques to speed up web page loading, browsers still suffer from relatively long page load time (PLT). Particularly, web applications need browsers to have a higher performance to compete with native applications. Recent studies show that network connection is not the bottleneck of the browser's performance anymore, however, no subsequent analysis has been conducted to inspect which parts of the browser's computation contribute to the performance overhead. In this paper, we apply a comprehensive and quantitative what-if analysis on the web browser's page loading process. Unlike conventional profiling methods, we apply causal profiling to precisely determine the impact of each computation stage such as HTML parsing and Layout on PLT. For this purpose, we develop COZ+1 is available at https://gitlab.com/coz-plus/coz-plus, a high-performance causal profiler capable of analyzing large software systems such as the Chromium browser. COZ+ highlights the most influential spots for further optimization. For example, it shows that optimizing JavaScript by 40% is expected to improve the Chromium page loading performance by more than 8.5% under typical network conditions.","PeriodicalId":268462,"journal":{"name":"Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114656323","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 21

Two Extensions of Kingman's GI/G/1 Bound Kingman的GI/G/1界的两个扩展

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331494

F. Ciucu, Felix Poloczek

引用次数: 1

Learning to Control Renewal Processes with Bandit Feedback 学习用强盗反馈控制更新过程

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331515

Semih Cayci, A. Eryilmaz, R. Srikant

{"title":"Learning to Control Renewal Processes with Bandit Feedback","authors":"Semih Cayci, A. Eryilmaz, R. Srikant","doi":"10.1145/3309697.3331515","DOIUrl":"https://doi.org/10.1145/3309697.3331515","url":null,"abstract":"We consider a bandit problem with K task types from which the controller activates one task at a time. Each task takes a random and possibly heavy-tailed completion time, and a reward is obtained only after the task is completed. The task types are independent from each other, and have distinct and unknown distributions for completion time and reward. For a given time horizon τ, the goal of the controller is to schedule tasks adaptively so as to maximize the reward collected until τ expires. In addition, we allow the controller to interrupt a task and initiate a new one. In addition to the traditional exploration-exploitation dilemma, this interrupt mechanism introduces a new one: should the controller complete the task and get the reward, or interrupt the task for a possibly shorter and more rewarding alternative? We show that for all heavy-tailed and some light-tailed completion time distributions, this interruption mechanism improves the reward linearly over time. From a learning perspective, the interrupt mechanism necessitates implicitly learning statistics beyond the mean from truncated observations. For this purpose, we propose a robust learning algorithm named UCB-BwI based on the median-of-means estimator for possibly heavy-tailed reward and completion time distributions. We show that, in a K-armed bandit setting with an arbitrary set of L possible interrupt times, UCB-BwI achieves O(Kłog(τ)+KL) regret. We also prove that the regret under any admissible policy is Ømega(Kłog(τ)), which implies that UCB-BwI is order optimal.","PeriodicalId":268462,"journal":{"name":"Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125798469","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 14

Understanding the Networking Performance of Wear OS 了解Wear OS的组网性能

Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer Systems Pub Date : 2019-06-20 DOI: 10.1145/3309697.3331475

Xiao Zhu, Y. Guo, Ashkan Nikravesh, Feng Qian, Z. Morley Mao

引用次数: 1