Low-Entropy Cloud Computing

Special Issue Banner

Scope

Traditional cloud computing application systems often suffer from the high-entropy problem. Three types of disorders exist, including workload interference, system jitter, and impedance mismatch. Such disorders make it difficult to simultaneously satisfy the requirements of high utilization and low latency. Applications often have to oversubscribe cloud resources, sometimes multiple times more than the actual need. Such oversubscribed resources often become stranded resources, i.e., resources that are allocated but unused. Although never used, stranded resources cannot be allocated to other applications. The five articles in this special issue present advances in low-entropy cloud computing, a new systems concept which addresses the above high-entropy and stranding problems. This set of techniques include a labeled processor architecture that enforces Differentiation, Isolation, and Prioritization (DIP) constraints, a labeled network stack offering a 10x goodput increase, a machine learning approach to configuration auto-tuning that significantly reduces tail latency, a Fractal Parallel Model (FPM) which limits the programming entropy with control pattern constraints, and a new benchmark suite called SDCBench. These articles involve systems benchmarking and entropy measurement, techniques in processor architecture, system software, application framework, and evaluation results for various workloads. Together, they show that low-entropy cloud computing techniques can simultaneously enhance user experience and resource utilization.

    Guest Editor

    Prof. Zhiwei Xu, Institute of Computing Technology, Chinese Academy of Sciences

      Table of Contents

      A Labeled Architecture for Low-Entropy Clouds: Theory, Practice and Lessons

      Chuanqi Zhang, Sa Wang, Zihao Yu, Huizhe Wang, Yinan Xu, Luoshan Cai, Dan Tang, Ninghui Sun, and Yungang Bao

      Intelligent Computing, vol. 2022, Article ID 9795476, 14 pages, 2022


      Queueing-Theoretic Performance Analysis of a Low-Entropy Labeled Network Stack

      Hongrui Guo, Wenli Zhang, Zishu Yu, and Mingyu Chen

      Intelligent Computing, vol. 2022, Article ID 9863054, 16 pages, 2022


      Resource Configuration Tuning for Stream Data Processing Systems via Bayesian Optimization

      Shixin Huang, Chao Chen, Gangya Zhu, Jinhan Xin, Zheng Wang, Kai Hwang, and Zhibin Yu

      Intelligent Computing, vol. 2022, Article ID 9820424, 16 pages, 2022


      Fractal Parallel Computing

      Yongwei Zhao, Yunji Chen, and Zhiwei Xu

      Intelligent Computing, vol. 2022, Article ID 9797623, 10 pages, 2022


      SDCBench: A Benchmark Suite for Workload Colocation and Evaluation in Datacenters

      Yanan Yang, Xiangyu Kong, Laiping Zhao, Yiming Li, Huanyu Zhang, Jie Li, Heng Qi, and Keqiu Li

      Intelligent Computing, vol. 2022, Article ID 9810691, 18 pages, 2022