Blaze: Holistic Caching for Iterative Data Processing
Open Access
- 22 April 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 370-386
Abstract
Modern data processing workloads, such as machine learning and graph processing, involve iterative computations to converge generated models into higher accuracy. An effective caching mechanism is vital to expedite iterative computations since the intermediate data that needs to be stored in memory grows larger over iterations, often exceeding the memory capacity. However, existing systems handle intermediate data through separate operational layers (e.g., caching, eviction, and recovery), with each layer working independently in a greedy or cost-agnostic manner. These layers typically rely on user annotations and past access patterns, failing to make globally optimal decisions for the workload. To overcome these limitations, Blaze introduces a unified caching mechanism that integrates the separate operational layers. Blaze dynamically captures the workload structure and metrics using profiling and inductive regression, and automatically estimates the potential data caching efficiency associated with different operational decisions based on the profiled information. To achieve this goal, Blaze incorporates potential data recovery costs across stages into a single cost optimization function, which informs the optimal partition state and location. This approach reduces the significant disk I/O overheads caused by oversized partitions and the recomputation overheads for partitions with long lineages, while efficiently utilizing the constrained memory space. Our evaluations demonstrate that Blaze can accelerate end-to-end application completion time by 2.02 - 2.52× compared to recomputation-based MEM_ONLY Spark, and by 1.08 - 2.86× compared to checkpoint-based MEM+DISK Spark, while reducing the cache data stored on disk by 95% on average.Keywords
Funding Information
- Institute for Information and Communication Technology Promotion (2015-0-00221)
This publication has 30 references indexed in Scilit:
- Echo: Compiler-based GPU Memory Footprint Reduction for LSTM RNN Training2020
- DAG-Aware Joint Task Scheduling and Cache Management in Spark Clusters2020
- LCRC: A Dependency-Aware Cache Management Policy for Spark2018
- LERC: Coordinated Cache Management for Data-Parallel Systems2017
- LRC: Dependency-aware cache management for data analytics clusters2017
- LCS: An Efficient Data Eviction Strategy for SparkInternational Journal of Parallel Programming, 2016
- MEMTUNE: Dynamic Memory Management for In-Memory Data Analytic Platforms2016
- The HiBench benchmark suite: Characterization of the MapReduce-based data analysis2010
- Utility-Based Cache Partitioning: A Low-Overhead, High-Performance, Runtime Mechanism to Partition Shared Caches40th Annual IEEE/ACM International Symposium on Microarchitecture (MICRO 2007), 2006
- Connected Components in Random Graphs with Given Expected Degree SequencesAnnals of Combinatorics, 2002