PPoSS: Planning: Inflight Analytics to Control Large-Scale Heterogeneous Systems
PPoSS: Planning: Inflight Analytics to Control Large-Scale Heterogeneous Systems
批准号:
2029049
负责人:
Ravishankar Iyer
金额:
$25.0万
依托单位国家:
美国
项目类别:
Standard Grant
财政年份:
2020
资助国家:
美国
项目状态:
已结题
起止时间:
2020-10-01 至 2024-09-30
中文摘要
该项目的目标是从根本上重塑系统的设计,从硬件到应用程序,使用快速,新颖的飞行分析来控制和优化大规模异构计算机系统,以满足新兴应用程序的性能和弹性要求,如数据挖掘,人工智能和个性化医疗。为了实现这一目标,高级机器学习(ML)方法沿着领域知识将用于支持实时系统状态估计和决策,包括资源管理,拥塞/故障检测和缓解,抢先入侵检测和配置管理。需要在整个系统堆栈中进行创新,以充分利用从系统的多个层收集的上下文信息并快速适应部署环境、工作负载和应用程序要求,从而实现最佳结果。ML驱动的飞行分析方法,在这项工作中开发的,将被证明在异构的“机架规模”的计算系统,最终的未来目标是扩大框架的仓库规模的计算系统。该项目将围绕以下研究活动组织。(i)处理噪声和不完整的遥测数据(例如,硬件遥测、OS级日志和应用级跟踪)可从跨系统栈的监视器获得以执行系统状态估计(例如,资源利用)。遥测数据在语义、模态和时间粒度方面往往是嘈杂和不一致的,使得系统只能部分地被观察到。将开发贝叶斯深度学习模型,以准确捕获系统状态并科普数据噪声和不完整性。(ii)设计模型和算法,用于做出决策的实际飞行分析(例如,调度或故障缓解),以增强系统性能、可靠性和安全性。这样的框架将由基于部分可观察马尔可夫决策过程(POMDP)的相互依赖的ML模型的集合组成,该部分可观察马尔可夫决策过程(POMDP)由领域知识(例如,互连拓扑)并进行真实的实时训练。(iii)综合硬件加速器,实现快速、低成本的机上分析。为此,将开发一个编译器和一个运行时框架,它采用高级声明性概率程序(即,POMDP),自动将它们编译到加速器上,并计划它们在异构硬件(FPGA、ASIC和CPU/GPU)上的执行。(iv)评估飞行分析的可信度。为此,将创建一个信任评估框架,以评估由于异构组件的剩余缺陷,输入不确定性以及随机ML算法的使用而导致的故障和攻击的弹性。 在规划阶段,该项目将重点关注机架级系统背景下的飞行分析设计。开发的方法和算法将有助于资源有限的小规模站点更有效地管理其系统。参与该项目的学生将有难得的机会参与设计具有广泛适用性的异构ML驱动系统。将ML方法和算法集成到真实的系统中可以吸引各种各样的人,包括代表性不足的少数民族学生。其目标是提高人们对设计和部署下一代计算系统以支持新兴应用的科学和工程挑战的认识。该奖项反映了NSF的法定使命,并通过使用基金会的知识价值和更广泛的影响审查标准进行评估,被认为值得支持。
英文摘要
The goal of this project is to fundamentally reinvent the design of the system, from hardware to application, using fast, novel inflight analytics to control and optimize large-scale heterogeneous computer systems to meet the performance and resiliency requirements of emerging applications such as data mining, artificial intelligence, and individualized medicine. Towards that goal, advanced machine-learning (ML) methods along with domain knowledge will be employed to support real-time system-state estimation and decision-making, including resource management, congestion/failure detection and mitigation, preemptive intrusion detection, and configuration management. Innovations across the system stack will be needed to achieve optimal results by taking full advantage of contextual information collected from multiple layers of the system and adapting rapidly to the deployment environment, workloads, and application requirements. ML-driven inflight analytics methods, developed in this effort, will be demonstrated on a heterogeneous “rack-scale” computing system, with the ultimate future objective of scaling up the framework to a warehouse-scale computing system.The project will be organized around the following research activities. (i) Work with noisy and incomplete telemetry data (e.g., hardware telemetry, OS-level logs, and application-level traces) available from monitors across the system stack to perform system-state estimation (e.g., resource utilization). Telemetry data are often noisy and inconsistent in terms of semantics, modalities, and time granularities, making systems only partially observable. Bayesian deep-learning models will be developed to accurately capture system states and cope with data noise and incompleteness. (ii) Design models and algorithms for practical inflight analytics that make decisions (e.g., on scheduling or failure mitigation) based on the estimated system state to enhance system performance, reliability, and security. Such a framework will consist of an ensemble of interdependent ML models based on partially observable Markov decision processes (POMDPs) augmented with domain knowledge (e.g., interconnect topology) and trained in real time. (iii) Synthesize hardware accelerators for fast, low-cost inflight analytic. Toward that end, a compiler and a runtime framework will be developed that take high-level declarative probabilistic programs (i.e., the POMDPs), automatically compile them onto accelerators, and plan their execution across heterogeneous hardware (FPGAs, ASICs, and CPUs/GPUs). (iv) Assess the trustworthiness of inflight analytics. For that, a trust-assessment framework will be created to evaluate resiliency to failures and attacks due to residual imperfections of heterogeneous components, input uncertainty, and the use of stochastic ML algorithms. While in the planning stage, this project will focus on design of inflight analytics in the context of rack-scale systems. The methods and algorithms developed will be useful in helping smaller-scale sites with limited resources manage their systems more efficiently. Students involved in this project will have a rare opportunity to participate in the design of heterogeneous ML-driven systems with broad applicability. The integration of ML methods and algorithms into real systems can be attractive to a diverse range of individuals, including underrepresented minority students. The goal is to raise awareness of scientific and engineering challenges in design and deployment of next-generation computing systems to support emerging applications.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
期刊论文(15)
专著(0)
科研奖励(0)
会议论文
登录
查看更多内容
DOI:
10.1109/micro50266.2020.00017
发表时间:
2020-10
期刊:
2020 53rd Annual IEEE/ACM International Symposium on Microarchitecture (MICRO)
影响因子:
--
作者:
[Dimitrios Skarlatos;Qingrong Chen;Jianyan Chen;Tianyin Xu;J. Torrellas]
通讯作者:
Dimitrios Skarlatos;Qingrong Chen;Jianyan Chen;Tianyin Xu;J. Torrellas
DOI:
10.1145/3517207.3526971
发表时间:
2022-04
期刊:
Proceedings of the 2nd European Workshop on Machine Learning and Systems
影响因子:
--
作者:
[Haoran Qiu;Weichao Mao;Archit Patke;Chen Wang;H. Franke;Z. Kalbarczyk;T. Başar;R. Iyer]
通讯作者:
Haoran Qiu;Weichao Mao;Archit Patke;Chen Wang;H. Franke;Z. Kalbarczyk;T. Başar;R. Iyer
Delay sensitivity-driven congestion mitigation for HPC systems
HPC 系统的延迟敏感度驱动的拥塞缓解
DOI:
10.1145/3447818.3460362
发表时间:
2021
期刊:
Proceedings of The 35th ACM International Conference on Supercomputing (ICS ‘21
影响因子:
--
作者:
[Patke, Archit, Jha, Saurabh, Qiu, Haoran, Brandt, Jim, Gentile, Ann, Greenseid, Joe, Kalbarczyk, Zbigniew, Iyer, Ravishankar K.]
通讯作者:
Iyer, Ravishankar K.
DOI:
10.1145/3458336.3465276
发表时间:
2021-06
期刊:
Proceedings of the Workshop on Hot Topics in Operating Systems
影响因子:
--
作者:
[Xudong Sun;L. Suresh;Aishwarya Ganesan;Ramnatthan Alagappan;Michael Gasch;Lilia Tang;Tianyin Xu]
通讯作者:
Xudong Sun;L. Suresh;Aishwarya Ganesan;Ramnatthan Alagappan;Michael Gasch;Lilia Tang;Tianyin Xu
DOI:
10.1145/3542929.3563475
发表时间:
2022-11
期刊:
Proceedings of the 13th Symposium on Cloud Computing
影响因子:
--
作者:
[Haoran Qiu;Weichao Mao;Archit Patke;Chen Wang;H. Franke;Z. Kalbarczyk;T. Başar;R. Iyer]
通讯作者:
Haoran Qiu;Weichao Mao;Archit Patke;Chen Wang;H. Franke;Z. Kalbarczyk;T. Başar;R. Iyer
共 13 条
CPS: Breakthrough:Towards Resiliency in Cyber-physical Systems for Robot-assisted Surgery
-
批准号:1545069
-
项目类别:Standard Grant
-
资助金额:$50.0万
-
财政年份:2016
-
负责人:Ravishankar Iyer
-
依托单位:
I/UCRC: Phase I: Computing and Genomics-An Essential Partnership for Biology Breakthroughs (CCBGM)
-
批准号:1624790
-
项目类别:Continuing Grant
-
资助金额:$75.0万
-
财政年份:2016
-
负责人:Ravishankar Iyer
-
依托单位:
CI: NEW: Collaborative Research: Computer System Failure Data Repository to Enable Data-Driven Dependability
-
批准号:1513051
-
项目类别:Standard Grant
-
资助金额:$37.09万
-
财政年份:2015
-
负责人:Ravishankar Iyer
-
依托单位:
I/UCRC Planning Grant: Computing and Genomics - An Essential Partnership for Biology Breakthroughs
-
批准号:1439719
-
项目类别:Standard Grant
-
资助金额:$1.63万
-
财政年份:2014
-
负责人:Ravishankar Iyer
-
依托单位:
Adaptive Software Implemented Fault-Tolerance for Networked Systems
-
批准号:9902026
-
项目类别:Standard Grant
-
资助金额:$25.0万
-
财政年份:1999
-
负责人:Ravishankar Iyer
-
依托单位:
Academic Research Infrastructure: Acquisition of Research Equipment for High-Speed Computing and Networking Initiative
-
批准号:9601631
-
项目类别:Standard Grant
-
资助金额:$72.49万
-
财政年份:1996
-
负责人:Ravishankar Iyer
-
依托单位:
Engineering Research Equipment Grant: Investigation of LISPMachine Architecture Reliability and Performance
-
批准号:8604893
-
项目类别:Standard Grant
-
资助金额:$12.9万
-
财政年份:1986
-
负责人:Ravishankar Iyer
-
依托单位:
海外基金