Rui's Blog
  • Rui's Blog/Paper Reading Notes - Introduction
  • Personal Blog
    • Personal Blog - Index
      • How to Create Picture-in-Picture Effect / Video Overlay for a Presentation Video
      • How to Do Your Part to Protect the Environment in Wisconsin
      • How to Get a Driver's License in Wisconsin
      • How to Travel from the U.S. to China onboard AA127 in June 2021
      • How to Transfer Credits Back to UW-Madison
      • Resources on Learning Academic Writing (for Computer Science)
    • Towards applying to CS Ph.D. programs
  • Machine Learning Systems
    • Machine Learning Systems - Index
      • MLSys Papers - Short Notes
      • [2011 NSDI] Dominant Resource Fairness: Fair Allocation of Multiple Resource Types
      • [2014 OSDI] Scaling Distributed Machine Learning with the Parameter Server
      • [2018 OSDI] Gandiva: Introspective Cluster Scheduling for Deep Learning
      • [2018 SIGCOMM] Chameleon: Scalable Adaptation of Video Analytics via Temporal and Cross-camera ...
      • [2018 NIPS] Dynamic Space-Time Scheduling for GPU Inference
      • [2019 ATC] Analysis of Large-Scale Multi-Tenant GPU Clusters for DNN Training Workloads
      • [2019 NSDI] Tiresias: A GPU Cluster Manager for Distributed Deep Learning
      • [2019 SOSP] ByteScheduler: A Generic Communication Scheduler for Distributed DNN Training ...
      • [2019 SOSP] PipeDream: Generalized Pipeline Parallelism for DNN Training
      • [2019 SOSP] Parity Models: Erasure-Coded Resilience for Prediction Serving Systems
      • [2019 NIPS] GPipe: Efficient Training of Giant Neural Networks using Pipeline Parallelism
      • [2019 SC] ZeRO: memory optimizations toward training trillion parameter models
      • [2020 OSDI] Gavel: Heterogeneity-Aware Cluster Scheduling Policies for Deep Learning Workloads
      • [2020 OSDI] AntMan: Dynamic Scaling on GPU Clusters for Deep Learning
      • [2020 OSDI] BytePS: A High Performance and Generic Framework for Distributed DNN Training
      • [2020 SIGCOMM] Reducto: On-Camera Filtering for Resource-Efficient Real-Time Video Analytics
        • [2020 MLSys] Salus: Fine-Grained GPU Sharing Primitives for Deep Learning Applications
      • [2020 EuroSys] AlloX: Compute Allocation in Hybrid Clusters
      • [2020 VLDB] PyTorch Distributed: Experiences on Accelerating Data Parallel Training
      • [2020 NetAI] Is Network the Bottleneck of Distributed Training?
      • [2020 NSDI] Themis: Fair and Efficient GPU Cluster Scheduling
      • [2021 MLSys] Accordion: Adaptive Gradient Communication via Critical Learning Regime Identification
      • [2021 VLDB] Analyzing and Mitigating Data Stalls in DNN Training
      • [2021 FAST] CheckFreq: Frequent, Fine-Grained DNN Checkpointing
      • [2021 EuroMLSys] Interference-Aware Scheduling for Inference Serving
      • [2021 OSDI] Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning
      • [2021 MLSys] Wavelet: Efficient DNN Training with Tick-Tock Scheduling
      • [2021 NSDI] SwitchML: Scaling Distributed Machine Learning with In-Network Aggregation
    • Big Data Systems - Index
      • Big Data Systems Papers - Short Notes
      • [2003 SOSP] The Google File System
      • [2004 OSDI] MapReduce: Simplified Data Processing on Large Clusters
      • [2010 SIGMOD] Pregel: A System for Large-Scale Graph Processing
      • [2011 NSDI] Mesos: A Platform for Fine-Grained Resource Sharing in the Data Center
      • [2012 NSDI] Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster ...
      • [2012 OSDI] PowerGraph: Distributed Graph-Parallel Computation on Natural Graphs
      • [2019 FAST] DistCache: Provable Load Balancing for Large-Scale Storage Systems with Distributed...
      • [2021 HotOS] From Cloud Computing to Sky Computing
      • [2021 EuroSys] NextDoor: Accelerating graph sampling for graph machine learning using GPUs
  • Earlier Readings & Notes
    • High Performance Computing Course Notes
      • Lecture 1: Course Overview
      • Lecture 2: From Code to Instructions. The FDX Cycle. Instruction Level Parallelism.
      • Lecture 3: Superscalar architectures. Measuring Computer Performance. Memory Aspects.
      • Lecture 4: The memory hierarchy. Caches.
      • Lecture 5: Caches, wrap up. Virtual Memory.
      • Lecture 6: The Walls to Sequential Computing. Moore’s Law.
      • Lecture 7: Parallel Computing. Flynn's Taxonomy. Amdahl's Law.
      • Lecture 8: GPU Computing Intro. The CUDA Programming Model. CUDA Execution Configuration.
      • Lecture 9: GPU Memory Spaces
      • Lecture 10: GPU Scheduling Issues.
      • Lecture 11: Execution Divergence. Control Flow in CUDA. CUDA Shared Memory Issues.
      • Lecture 12: Global Memory Access Patterns and Implications.
      • Lecture 13: Atomic operations in CUDA. GPU ode optimization rules of thumb.
      • Lecture 14: CUDA Case Studies. (1) 1D Stencil Operation. (2) Vector Reduction in CUDA.
      • Lecture 15: CUDA Case Studies. (3) Parallel Prefix Scan on the GPU. Using Multiple Streams in CUDA.
      • Lecture 16: Streams, and overlapping data copy with execution.
      • Lecture 17: GPU Computing: Advanced Features.
      • Lecture 18: GPU Computing with thrust and cub.
      • Lecture 19: Hardware aspects relevant in multi-core, shared memory parallel computing.
      • Lecture 20: Multi-core Parallel Computing with OpenMP. Parallel Regions.
      • Lecture 21: OpenMP Work Sharing.
      • Lecture 22: OpenMP Work Sharing
      • Lecture 23: OpenMP NUMA Aspects. Caching and OpenMP.
      • Lecture 24: Critical Thinking. Code Optimization Aspects.
      • Lecture 25: Computing with Supercomputers.
      • Lecture 26: MPI Parallel Programming General Introduction. Point-to-Point Communication.
      • Lecture 27: MPI Parallel Programming Point-to-Point communication: Blocking vs. Non-blocking sends.
      • Lecture 28: MPI Parallel Programming: MPI Collectives. Overview of topics covered in the class.
    • Cloud Computing Course Notes
      • 1.1 Introduction to Clouds, MapReduce
      • 1.2 Gossip, Membership, and Grids
      • 1.3 P2P Systems
      • 1.4 Key-Value Stores, Time, and Ordering
      • 1.5 Classical Distributed Algorithms
      • 4.1 Spark, Hortonworks, HDFS, CAP
      • 4.2 Large Scale Data Storage
    • Operating Systems Papers - Index
      • CS 736 @ UW-Madison Fall 2020 Reading List
      • All File Systems Are Not Created Equal: On the Complexity of Crafting Crash-Consistent Applications
      • ARC: A Self-Tuning, Low Overhead Replacement Cache
      • A File is Not a File: Understanding the I/O Behavior of Apple Desktop Applications
      • Biscuit: The benefits and costs of writing a POSIX kernel in a high-level language
      • Data Domain: Avoiding the Disk Bottleneck in the Data Domain Deduplication File System
      • Disco: Running Commodity Operating Systems on Scalable Multiprocessors
      • FFS: A Fast File System for UNIX
      • From WiscKey to Bourbon: A Learned Index for Log-Structured Merge Trees
      • LegoOS: A Disseminated, Distributed OS for Hardware Resource Disaggregation
      • LFS: The Design and Implementation of a Log-Structured File System
      • Lottery Scheduling: Flexible Proportional-Share Resource Management
      • Memory Resource Management in VMware ESX Server
      • Monotasks: Architecting for Performance Clarity in Data Analytics Frameworks
      • NFS: Sun's Network File System
      • OptFS: Optimistic Crash Consistency
      • RAID: A Case for Redundant Arrays of Inexpensive Disks
      • RDP: Row-Diagonal Parity for Double Disk Failure Correction
      • Resource Containers: A New Facility for Resource Management in Server Systems
      • ReVirt: Enabling Intrusion Analysis through Virtual-Machine Logging and Replay
      • Scheduler Activations: Effective Kernel Support for the User-Level Management of Parallelism
      • SnapMirror: File-System-Based Asynchronous Mirroring for Disaster Recovery
      • The Linux Scheduler: a Decade of Wasted Cores
      • The Unwritten Contract of Solid State Drives
      • Venti: A New Approach to Archival Storage
    • Earlier Notes
      • How to read a paper
  • FIXME
    • Template for Paper Reading Notes
Powered by GitBook
On this page
  • One-line Summary
  • Paper Structure Outline
  • Background & Motivation
  • Design and Implementation
  • Ballooning
  • Idle Memory Tax
  • Content-Based Page Sharing
  • Hot I/O Page Remapping
  • Evaluation
  • New Vocabulary
  • Links

Was this helpful?

  1. Earlier Readings & Notes
  2. Operating Systems Papers - Index

Memory Resource Management in VMware ESX Server

One-line Summary

This paper introduced some mechanisms and policies for better memory management in ESX Servers.

Paper Structure Outline

  1. Introduction

  2. Memory Virtualization

  3. Reclamation Mechanisms

    1. Page Replacement Issues

    2. Ballooning

    3. Demand Paging

  4. Sharing Memory

    1. Transparent Page Sharing

    2. Content-Based Page Sharing

    3. Implementation

  5. Shares cs. Working Sets

    1. Share-Based Allocation

    2. Reclaiming Idle Memory

    3. Measuring Idle Memory

    4. Experimental Results

  6. Allocation Policies

    1. Parameters

    2. Admission Control

    3. Dynamic Reallocation

  7. I/O Page Remapping

  8. Related Work

  9. Conclusions

Background & Motivation

This paper attempts to resolve some issues in existing Virtual Machine Managers (VMMs) by introducing several novel mechanisms and policies to manage memory.

Design and Implementation

Several novel techniques and mechanisms are presented.

Ballooning

This technique allows memories to be reclaimed by ESX Servers from the guest OS. When a VM is claimed memory, the ballooning technique persuades the guest OS into cooperating with it. A balloon module is loaded into the guest OS as a pseudo-device driver or kernel service. When the balloon is inflated (by pinning physical pages), the guest OS gets pressured and invokes its own memory management algorithms. Based on the amount of remaining memory, the guest OS either returns pages from its free list or pages out some particular pages to its own virtual disk. In contrast, when a balloon is deflated, the guest OS gets to use more memory.

Idle Memory Tax

Idle clients can hoard memory while active clients without enough memory suffer severe memory pressure. This technique basically charges a client more for an idle page than for one it is actively using, preventing the aforementioned scenario from happening when memory is scarce. To estimate the fraction of active memory of each virtual machine, ESX Server uses a statistical sampling approach to obtain estimates.

Content-Based Page Sharing

Transparent page sharing is pioneered by Disco, but in order to achieve page sharing, Disco needs to make interface changes in the guest OS. In this paper, ESX Server finds identical pages between VMs by scanning for shared opportunities and hashing the contents for easy comparison. If a match is found, a standard copy-on-write (COW) technique is used to share the pages.

Hot I/O Page Remapping

For high-memory systems, modern processors support a physical address extension (PAE) mode that allows the hardware to address up to 64 GB of memory with 36-bit addresses. However, many devices that use direct memory access (DMA) for I/O transfers can only address a subset of this memory. The existing solution is to copy the data in high memory through a temporary bounce buffer into low memory. This creates significant overhead. ESX Server addresses this issue by maintaining statistics to track "hot" pages in high memory that are involved in repeated I/O operations. When the page is too hot (count exceeding a threshold), the page is remapped into low memory.

Evaluation

Unlike a traditional paper, this paper does not have a section dedicated to evaluations. Instead, the evaluations are broken down, following techniques of which they analyze. See the previous section for the evaluations.

New Vocabulary

  • ESX Server: An enterprise-level virtualization tool. ESX Server runs on "bare-metal". This means that the ESX Server software is directly installed into the computer w/o an operating system for it to run on top of.

Links

PreviousLottery Scheduling: Flexible Proportional-Share Resource ManagementNextMonotasks: Architecting for Performance Clarity in Data Analytics Frameworks

Last updated 4 years ago

Was this helpful?

by , a UW-Madison alumni, now a Ph.D. student at UT-Austin.

Paper PDF
Presentation Slides
What is VMware ESX Server?
Paper review
Zeyuan Hu
CS 736 @ UW-Madison Spring 2015 Panel Discussion
Ballooning adds a small overhead (1.4% to 4.4%, because guest OS uses more resources when given more memory initially) while obtaining similar performance compared with a VM w/o ballooning.
The evaluations are obtained by varying the tax rate on two VMs.
For large numbers of VMs, around 2/3 of all VM memory is reclaimed.