ACM Transactions on Modeling and Performance Evaluation of Computing Systems

In data center networks, the reliability of Service Function Chain (SFC)—an end-to-end service presented by a chain of virtual network functions (VNFs)—is a complex and specific function of placement, configuration, and application requirements, both in hardware and software. Existing approaches to reliability analysis do not jointly consider multiple features of system components, including, (i) heterogeneity, (ii) disjointness, (iii) sharing, (iv) redundancy, and (v) failure interdependency. To this end, we develop a novel analysis of service reliability of the so-called generic SFC, consisting of n = k + r sub-SFCs, whereby k ≥ 1 and r ≥ 0 are the numbers of arbitrary placed primary and backup (redundant) sub-SFCs, respectively. Our analysis is based on combinatorics and a reduced binomial theorem—resulting in a simple approach, which, however, can be utilized to analyze rather complex SFC configurations. The analysis is practically applicable to various VNF placement strategies in arbitrary data center configurations, and topologies and can be effectively used for evaluation and optimization of reliable SFC placements.

Anna Engelmann

Admela Jukan

A Combinatorial Reliability Analysis of Generic Service Function Chains in Data Center Networks

数据中心网络中通用服务功能链的组合可靠性分析

Communication overhead is a main bottleneck in federated learning (FL) especially in the wireless environment due to the limited data rate and unstable radio channels. The communication challenge necessitates holistic selection of participating clients that accounts for both the computation needs and communication cost, as well as judicious allocation of the limited transmission resource. Meanwhile, the random unpredictable nature of both the training data samples and the communication channels requires an online optimization approach that adapts to the changing system state over time. In this work, we consider a general framework of online joint client sampling and power allocation for wireless FL under time-varying communication channels. We formulate it as a stochastic network optimization problem that admits a Lyapunov-typed solution approach. This leads to per-training-round subproblems with a special bi-convex structure, which we leverage to propose globally optimal solutions, culminating in a meta algorithm that provides strong performance guarantees. We further study three specific FL problems covering multiple scenarios, namely with IID or non-IID data, whether robustness against data drift is required, and with unbiased or biased client sampling. We derive detailed algorithms for each of these problems. Simulation with standard classification tasks demonstrate that the proposed communication-aware algorithms outperform their counterparts under a wide range of learning and communication scenarios.

Wen Xu

Ben Liang

Gary Boudreau

Hamza Sokun

Clipper: Online Joint Client Sampling and Power Allocation for Wireless Federated Learning

Clipper: 无线联邦学习的在线联合客户端采样与功率分配

Federated Learning (FL) is a promising technique for decentralized privacy-preserving Machine Learning (ML) with a diverse pool of participating devices with varying device capabilities. However, existing approaches to handle such heterogeneous environments do not consider “fairness” in model aggregation, resulting in significant performance variation among devices. Meanwhile, prior works on FL fairness remain hardware-oblivious and cannot be applied directly without severe performance penalties. To address this issue, we propose a novel hardware-sensitive FL method called \(\mathsf {FairHetero} \) that promotes fairness among heterogeneous federated clients. Our approach offers tunable fairness within a group of devices with the same ML architecture as well as across different groups with heterogeneous models. Our evaluation under MNIST , FEMNIST , CIFAR10 , and SHAKESPEARE datasets demonstrates that \(\mathsf {FairHetero} \) can reduce variance among participating clients’ test loss compared to the existing state-of-the-art (SOTA) techniques, resulting in increased overall performance.

Zahidur Talukder

Bingqian Lu

Shaolei Ren

Mohammad Atiqul Islam

Hardware-Sensitive Fairness in Heterogeneous Federated Learning

异构联邦学习中的硬件敏感公平性

A correct evaluation of scheduling algorithms and a good understanding of their optimization criteria are key components of resource management in HPC. In this work, we discuss bias and limitations of the most frequent optimization metrics from the literature. We provide elements on how to evaluate performance when studying HPC batch scheduling. We experimentally demonstrate these limitations by focusing on two use-cases: a study on the impact of runtime estimates on scheduling performance, and the reproduction of a recent high-impact work that designed an HPC batch scheduler based on a network trained with reinforcement learning. We demonstrate that focusing on quantitative optimization criterion (“our work improves the literature by X%”) may hide extremely important caveat, to the point that the results obtained are opposed to the actual goals of the authors. Key findings show that mean bounded slowdown and mean response time are hazardous for a purely quantitative analysis in the context of HPC. Despite some limitations, utilization appears to be a good objective. We propose to complement it with the standard deviation of the throughput in some pathological cases. Finally, we argue for a larger use of area-weighted response time, that we find to be a very relevant objective.

Robin Boëzennec

Fanny Dufossé

Guillaume Pallez

Qualitatively Analyzing Optimization Objectives in the Design of HPC Resource Manager

定性分析高性能计算资源管理器设计中的优化目标

Model pruning has been proposed as a technique for reducing the size and complexity of Federated learning (FL) models. By making local models coarser, pruning is intuitively expected to improve protection against privacy attacks. However, the level of this expected privacy protection has not been previously characterized, or optimized jointly with utility. In this paper, we first characterize the privacy offered by pruning. We establish information-theoretic upper bounds on the information leakage from pruned FL and we experimentally validate them under state-of-the-art privacy attacks across different FL pruning schemes. Second, we introduce PriPrune – a privacy-aware algorithm for pruning in FL. PriPrune uses defense pruning masks, which can be applied locally after any pruning algorithm, and adapts the defense pruning rate to jointly optimize privacy and accuracy. Another key idea in the design of PriPrune is Pseudo-Pruning : it undergoes defense pruning within the local model and only sends the pruned model to the server; while the weights pruned out by defense mask are withheld locally for future local training rather than being removed. We show that PriPrune significantly improves the privacy-accuracy tradeoff compared to state-of-the-art pruned FL schemes. For example, on the FEMNIST dataset, PriPrune improves the privacy of PruneFL by 45.5% without reducing accuracy.

Tianyue Chu

Mengwei Yang

Nikolaos Laoutaris

Athina Markopoulou

PriPrune: Quantifying and Preserving Privacy in Pruned Federated Learning

PriPrune：量化并保护隐私在修剪的联邦学习中的应用

Microservice applications are increasingly adopted in distributed data processing systems, such as in mobile edge computing and data mesh architectures. However, existing performance models of such systems fall short in providing comprehensive insights into the intricate interplay between data placement and data processing. To address these issues, this paper proposes a novel class of performance models that enables joint analysis of data storage access workflows, caching, and queueing contention. Our proposed models introduce a notion of access path for data items to model hierarchical data locality constraints. We develop analytical solutions to efficiently approximate the performance metrics of these models under different data caching policies, finding in particular conditions under which the underlying Markov chain admits a product-form solution. Extensive trace-driven simulations based on real-world datasets indicate that service and data placement policies based on our proposed models can respectively improve by up to 35% and 37% the average response time in edge and data mesh case studies compared to baseline resource allocation heuristics.

Yicheng Gao

Giuliano Casale

Rekha Singhal

Performance Modeling of Distributed Data Processing in Microservice Applications

微服务应用中分布式数据处理的性能建模

短名	ACM Trans. Model. Perform. Eval. Comput. Syst.
Journal Impact	0.72
国际分区	COMPUTER SCIENCE, INFORMATION SYSTEMS(Q4)

ISSN	2376-3639, 2376-3647
h-index	16

涉及主题	计算机科学工程类操作系统计算机网络数学程序设计语言经济物理分布式计算统计电信政治学法学万维网量子力学管理电气工程生物业务人工智能
出版信息	出版商: Association for Computing Machinery (ACM)，出版周期: ，期刊类型: journal，开源期刊: 否
基本数据	创刊年份: 2016，原创研究文献占比: 100.00%，自引率:14.30%， Gold OA占比: 0.00%

ACM Transactions on Modeling and Performance Evaluation of Computing Systems

期刊引文格式

有 1 位以上作者的期刊

有 2 位作者的期刊

有 3 位作者的期刊

有 5 位以上作者的期刊

书籍引用格式

学位论文引用格式

网页引用格式

专利引用格式

手工熬夜修改参考文献？研飞自动匹配期刊，一键轻松成稿，支持 Word/WPS

投稿经验分享

A Combinatorial Reliability Analysis of Generic Service Function Chains in Data Center Networks

Clipper: Online Joint Client Sampling and Power Allocation for Wireless Federated Learning

Hardware-Sensitive Fairness in Heterogeneous Federated Learning

Qualitatively Analyzing Optimization Objectives in the Design of HPC Resource Manager

PriPrune: Quantifying and Preserving Privacy in Pruned Federated Learning

Hardware-Sensitive Fairness in Heterogeneous Federated Learning

Clipper: Online Joint Client Sampling and Power Allocation for Wireless Federated Learning

PriPrune: Quantifying and Preserving Privacy in Pruned Federated Learning

Qualitatively Analyzing Optimization Objectives in the Design of HPC Resource Manager

Performance Modeling of Distributed Data Processing in Microservice Applications