MapReduce: Simplified Data Processing on Large Clusters
OSDI, 2004 · paper · Computer science
Presents MapReduce, a programming model and implementation for processing large data sets on clusters of commodity machines. Users write map and reduce functions, and the system handles parallelization and failures; it is reported to be scalable and widely used at Google.
Checked: PDF from static.googleusercontent.com, fetched 4 Oct 2026. Page numbers are those of that PDF.
1 finding
Reference or label Reference [17]: "L. G. Valiant. A bridging model for parallel computation. Communications of the ACM, 33(8):103–111, 1997." The year does not match volume 33(8).
- p. 13 References, entry [17]
Communications of the ACM, 33(8):103–111, 1997.
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: CACM volume 33, issue 8 is August 1990, when Valiant's paper appeared. The p.13 image confirms that the printed year is 1997.
Second check: another model agreed
Report a problem with this finding
4 minor (typos and the like; most checked by the judge only)
Wording The introduction roadmap says "Section 7 discusses related and future work", but Section 7 is headed "7 Related Work" and the paper has no discussion of future work.
- p. 2 Introduction roadmap, top of the left column
Sec-tion 7 discusses related and future work.
- p. 11 Section 7 heading
7 Related Work
Judge's note: The p.2 crop shows the roadmap sentence. Searching the text layer finds 'future' only in that roadmap line (main.txt line 86). Section 7 is titled 'Related Work' and covers only prior systems. The roadmap describes content that is not in the paper.
Second check: another model agreed
Report a problem with this finding
Wording The Figure 3(c) caption "200 tasks killed" describes the experiment differently from Section 5.5, which "intentionally killed 200 out of 1746 worker processes".
- p. 9 Figure 3, panel (c) caption
(c) 200 tasks killed
- p. 10 Section 5.5
killed 200 out of 1746 worker processes
Judge's note: Checked the p.9 caption and the p.10 text. The paper separates tasks from worker processes, so the caption's unit is loose shorthand for the same experiment. This is a minor labelling inconsistency.
Second check: another model agreed
Report a problem with this finding
Typo M and R are defined as the numbers of map and reduce tasks ("There are M map tasks and R reduce tasks to assign"), but Section 3.3 reuses them for individual tasks: "Consider map task M and reduce tasks R1 and R2."
- p. 5 Section 3.3, Semantics in the Presence of Failures, right column
Consider map task M and reduce tasks R1 and R2.
- p. 4 Section 3.1, step 2
There are M map tasks and R reduce tasks to assign.
Judge's note: Checked the p.5 image. The symbol is overloaded, but context makes the meaning clear. This is a cosmetic notational inconsistency only.
Second check: none (judge only)
Report a problem with this finding
Typo Section 5.3: "This functions passes the intermediate key/value pair unchanged" should read "This function passes".
- p. 9 Section 5.3, first paragraph of the left column
This functions passes the intermediate key/value pair un-
- p. 9 Section 5.3, left column, first paragraph
This functions passes the intermediate key/value pair un- changed
Judge's note: The p.9 image shows 'This functions passes'. It refers to the single Identity function, so this is a number-agreement typo.
Second check: none (judge only)