Random Notes
  • Introduction
  • Reading list
  • Theory
    • Index
      • Impossibility of Distributed Consensus with One Faulty Process
      • Time, Clocks, and the Ordering of Events in a Distributed System
      • Using Reasoning About Knowledge to analyze Distributed Systems
      • CAP Twelve Years Later: How the “Rules” Have Changed
      • A Note on Distributed Computing
  • Operating System
    • Index
  • Storage
    • Index
      • Tachyon: Reliable, Memory Speed Storage for Cluster Computing Frameworks
      • Exploiting Commutativity For Practical Fast Replication
      • Don’t Settle for Eventual: Scalable Causal Consistency for Wide-Area Storage with COPS
      • Building Consistent Transactions with Inconsistent Replication
      • Managing Update Conflicts in Bayou, a Weakly Connected Replicated Storage System
      • Spanner: Google's Globally-Distributed Database
      • Bigtable: A Distributed Storage System for Structured Data
      • The Google File System
      • Dynamo: Amazon’s Highly Available Key-value Store
      • Chord: A Scalable Peer-to-peer Lookup Service for Internet Applications
      • Replicated Data Consistency Explained Through Baseball
      • Session Guarantees for Weakly Consistent Replicated Data
      • Flat Datacenter Storage
      • Small Cache, Big Effect: Provable Load Balancing forRandomly Partitioned Cluster Services
      • DistCache: provable load balancing for large-scale storage systems with distributed caching
      • Short Summaries
  • Coordination
    • Index
      • Logical Physical Clocks and Consistent Snapshots in Globally Distributed Databases
      • Paxos made simple
      • ZooKeeper: Wait-free coordination for Internet-scale systems
      • Just Say NO to Paxos Overhead: Replacing Consensus with Network Ordering
      • Keeping CALM: When Distributed Consistency is Easy
      • In Search of an Understandable Consensus Algorithm
      • A comprehensive study of Convergent and Commutative Replicated Data Types
  • Fault Tolerance
    • Index
      • The Mystery Machine: End-to-end Performance Analysis of Large-scale Internet Services
      • Gray Failure: The Achilles’ Heel of Cloud-Scale Systems
      • Capturing and Enhancing In Situ System Observability for Failure Detection
      • Check before You Change: Preventing Correlated Failures in Service Updates
      • Efficient Scalable Thread-Safety-Violation Detection
      • REPT: Reverse Debugging of Failures in Deployed Software
      • Redundancy Does Not Imply Fault Tolerance
      • Fixed It For You:Protocol Repair Using Lineage Graphs
      • The Good, the Bad, and the Differences: Better Network Diagnostics with Differential Provenance
      • Lineage-driven Fault Injection
      • Short Summaries
  • Cloud Computing
    • Index
      • Improving MapReduce Performance in Heterogeneous Environments
      • CLARINET: WAN-Aware Optimization for Analytics Queries
      • MapReduce: Simplified Data Processing on Large Clusters
      • Dryad: Distributed Data-Parallel Programs from Sequential Building Blocks
      • Resource Management
      • Apache Hadoop YARN: Yet Another Resource Negotiator
      • Mesos: A Platform for Fine-Grained Resource Sharing in the Data Center
      • Dominant Resource Fairness: Fair Allocation of Multiple Resource Types
      • Large-scale cluster management at Google with Borg
      • MapReduce Online
      • Delay Scheduling: A Simple Technique for Achieving Locality and Fairness in Cluster Scheduling
      • Reining in the Outliers in Map-Reduce Clusters using Mantri
      • Effective Straggler Mitigation: Attack of the Clones
      • Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster Computing
      • Discretized Streams: Fault-Tolerant Streaming Computation at Scale
      • Sparrow: Distributed, Low Latency Scheduling
      • Making Sense of Performance in Data Analytics Framework
      • Monotasks: Architecting for Performance Clarity in Data Analytics Frameworks
      • Drizzle: Fast and Adaptable Stream Processing at Scale
      • Naiad: A Timely Dataflow System
      • The Dataflow Model:A Practical Approach to Balancing Correctness, Latency, and Cost in Massive-Scale
      • Interruptible Tasks:Treating Memory Pressure AsInterrupts for Highly Scalable Data-Parallel Program
      • PACMan: Coordinated Memory Caching for Parallel Jobs
      • Multi-Resource Packing for Cluster Schedulers
      • Other interesting papers
  • Systems for ML
    • Index
      • A Berkeley View of Systems Challenges for AI
      • Tiresias: A GPU Cluster Managerfor Distributed Deep Learning
      • Gandiva: Introspective Cluster Scheduling for Deep Learning
      • Workshop papers
      • Hidden Technical Debt in Machine Learning Systems
      • Inference Systems
      • Parameter Servers and AllReduce
      • Federated Learning at Scale - Part I
      • Federated Learning at Scale - Part II
      • Learning From Non-IID data
      • Ray: A Distributed Framework for Emerging AI Applications
      • PipeDream: Generalized Pipeline Parallelism for DNN Training
      • DeepXplore: Automated Whitebox Testingof Deep Learning Systems
      • Distributed Machine Learning Misc.
  • ML for Systems
    • Index
      • Short Summaries
  • Machine Learning
    • Index
      • Deep Learning with Differential Privacy
      • Accelerating Deep Learning via Importance Sampling
      • A Few Useful Things to Know About Machine Learning
  • Video Analytics
    • Index
      • Scaling Video Analytics on Constrained Edge Nodes
      • Focus: Querying Large Video Datasets with Low Latency and Low Cost
      • NoScope: Optimizing Neural Network Queriesover Video at Scale
      • Live Video Analytics at Scale with Approximation and Delay-Tolerance
      • Chameleon: Scalable Adaptation of Video Analytics
      • End-to-end Learning of Action Detection from Frame Glimpses in Videos
      • Short Summaries
  • Networking
    • Index
      • Salsify: Low-Latency Network Video through Tighter Integration between a Video Codec and a Transport
      • Learning in situ: a randomized experiment in video streaming
      • Short Summaries
  • Serverless
    • Index
      • Serverless Computing: One Step Forward, Two Steps Back
      • Encoding, Fast and Slow: Low-Latency Video Processing Using Thousands of Tiny Threads
      • SAND: Towards High-Performance Serverless Computing
      • Pocket: Elastic Ephemeral Storage for Serverless Analytics
      • Fault-tolerant and Transactional Stateful Serverless Workflows
  • Resource Disaggregation
    • Index
  • Edge Computing
    • Index
  • Security/Privacy
    • Index
      • Differential Privacy
      • Honeycrisp: Large-Scale Differentially Private Aggregation Without a Trusted Core
      • Short Summaries
  • Misc.
    • Index
      • Rate Limiting
      • Load Balancing
      • Consistency Models in Distributed System
      • Managing Complexity
      • System Design
      • Deep Dive into the Spark Scheduler
      • The Actor Model
      • Python Global Interpreter Lock
      • About Research and PhD
Powered by GitBook
On this page
  • Summary
  • Max-Min Fairness
  • Motivation
  • Dominant Resource Fairness(DRF)
  • An Example

Was this helpful?

  1. Cloud Computing
  2. Index

Dominant Resource Fairness: Fair Allocation of Multiple Resource Types

https://cs.stanford.edu/~matei/papers/2011/nsdi_drf.pdf

PreviousMesos: A Platform for Fine-Grained Resource Sharing in the Data CenterNextLarge-scale cluster management at Google with Borg

Last updated 5 years ago

Was this helpful?

Summary

To address the problem of fair resource allocation in a system containing different resource types, the paper proposes Dominant Resource Fairness(DRF), a generalization of max-min fairness to multiple resource types. DRF satisfies four properties: sharing incentive, strategy-proofness, Pareto efficiency and envy-freeness.

One of the most popular resource allocation strategy is max-min fairness, which maximize the minimum allocation received by a user in the system. More specifically, we say a sharing of a network resource(e.g. bandwidth) is max-min fair when:

  1. The lowest demand on the network resource is maximized

  2. After the lowest demand on the network resource has been satisfied, only then the second lowest demand on the network resource will be maximized;

  3. After the second lowest demand on the network resource has been satisfied, only then the third lowest demand on the network resource will be maximized;

  4. and so on...

The Max-Min Fairness Assignment Algorithm

  • Repeat until converge

    • Compute the fair share of each unsatisfied request

    • Assign the fair share to each unsatisfied request

    • Some flows may have received more than its request

    • If there are over-assignments:

      • Take back the over-assigned shares and compute the sum of the over-assigned shares, called the residual amount.

      • Assign the residual amount to the remaining unsatisfied requests

For example: Consider the problem of the max-min fair allocation for a set of four sources with demands 2, 2.6, 4, 5 when the resource has a capacity of 10. If we run the above algorithm, the fair allocation is: source 1 gets 2, source 2 gets 2.6, sources 3 and 4 get 2.7 each.

Prior to this work, despite the vast amount of work on fair allocation, the focus has so far been primarily on single resource type.

Motivation

Existing fair schedulers for clusters, such as Quincy and the Hadoop Fair Scheduler, ignore the heterogeneity of user demands, and allocate resources at the granularity of slots, where a slot is a fixed fraction of a node. This leads to inefficient allocation as a slot is more often than not a poor match for the task demands. The paper provides an analysis of a real-world trace showing that most of the tasks either underutilize or overutilize some of their slot resources.

Dominant Resource Fairness(DRF)

For every user, DRF computes the share of each resource allocated to that user. The maximum among all shares of a user is called that user’s dominant share, and the resource corresponding to the dominant share is called the dominant resource. Then, DRF simply applies max-min fairness across users' dominant shares.

The algorithm tracks the total resources allocated to each user as well as the user's dominant share. At each step, DRF picks the user with the lowest dominant share among those with task ready to run.

An Example

Consider a system consisting of 9 CPUs, 18 GB RAM and two users, where user A runs tasks with demand <1 CPU, 4GB>, and user B runs tasks with demand <3 CPUs, 1 GB> each. In this case, user A's dominant resource is memory, and user B's dominant resource is CPU.

At the end of the above allocation, user A gets <3 CPU, 12GB>, while user B gets <6 CPU, 2 GB>, i.e. each user gets 2/3 of its dominant resource.

Max-Min Fairness