Clash of the Titans: MapReduce vs. Spark for Large Scale Data Analytics
Summary: Profiling-driven, architecture-level comparison of MapReduce and Spark across shuffle, execution model, and caching, with tools linking task plans to resource use. Spark is ~2.5x–5x faster on WordCount, k-means, and PageRank; gains from hash-based aggregation and RDD caching, while MapReduce wins Sort via shuffle efficiency. (summarized by gpt-5-nano on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Juwei Shi
- 2. Yunjie Qiu
- 3. Umar Farooq Minhas
- 4. Limei Jiao
- 5. Chen Wang
- 6. Berthold Reinwald
- 7. Fatma Ozcan
Incoming Citations (Sorted by Pagerank)
Showing 7 of 7 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 3,343 | Comparative Evaluation of Big-Data Systems on Scientific Image Analytics Workloads | 2017 | VLDB | 7.1967343e-05 |
| 5,793 | Lifetime-Based Memory Management for Distributed Data Processing Systems | 2016 | VLDB | 5.3258796e-05 |
| 7,476 | Lachesis: Automatic Partitioning for UDF-Centric Analytics | 2021 | VLDB | 4.7188928e-05 |
| 9,332 | PlinyCompute: A Platform for High-Performance, Distributed, Data-Intensive Tool Development | 2018 | SIGMOD | 4.3556432e-05 |
| 11,427 | Accelerating Complex Analytics using Speculation | 2021 | CIDR | 4.1945683e-05 |
| 11,694 | An Experimental Evaluation of Garbage Collectors on Big Data Applications | 2019 | VLDB | 4.1945683e-05 |
| 11,802 | Query-able Kafka: An agile data analytics pipeline for mobile wireless networks | 2017 | VLDB | 4.1945683e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 3 of 3 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 4 | Pregel: A System for Large-Scale Graph Processing | 2010 | SIGMOD | 0.0019005923 |
| 868 | Profiling, What-if Analysis, and Cost-based Optimization of MapReduce Programs | 2011 | VLDB | 0.00015789681 |
| 7,304 | MRTuner: A Toolkit to Enable Holistic Optimization for MapReduce Jobs | 2014 | VLDB | 4.7684491e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| Overall Rank | Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 6,560 | Efficient In-memory Data Management: An Analysis | 2014 | VLDB | 5.010074e-05 |
| 868 | Profiling, What-if Analysis, and Cost-based Optimization of MapReduce Programs | 2011 | VLDB | 0.00015789681 |
| 8,300 | sPCA: Scalable Principal Component Analysis for Big Data on Distributed Platforms | 2015 | SIGMOD | 4.5435639e-05 |
| 5,888 | Magnet: Push-based Shuffle Service for Large-scale Data Processing | 2020 | VLDB | 5.2873617e-05 |
| 3,535 | Scaling Spark in the Real World: Performance and Usability | 2015 | VLDB | 6.9992495e-05 |
| 3,129 | Scalable Big Graph Processing in MapReduce | 2014 | SIGMOD | 7.5008242e-05 |
| 2,476 | A Platform for Scalable One-Pass Analytics using MapReduce | 2011 | SIGMOD | 8.6960139e-05 |
| 7,019 | Bridging the Gap Between HPC and Big Data Frameworks | 2017 | VLDB | 4.860057e-05 |
| 42 | A Comparison of Approaches to Large-Scale Data Analysis | 2009 | SIGMOD | 0.00073498298 |
| 1,615 | The Performance of MapReduce: An In-depth Study | 2010 | VLDB | 0.00011132319 |