MapReduce A Step Backwards: Is Comparison to Relational Databases Fair?
...As a data processing paradigm, MapReduce represents a giant step backwards. The database community has learned the following three lessons from the 40 years that have unfolded since IBM first released IMS in 1968....Given the experimental evaluations to date, we have serious doubts about how well MapReduce applications can scale. Moreover, the MapReduce implementers would do well to study the last 25 years of parallel DBMS research literature.
The article goes on to list criteria such as:
- MapReduce is a poor implementation (in comparison to B-trees)
- MapReduce is not novel
- MapReduce is missing features (such as loading and indexing)
- MapReduce is incompatible with the DBMS tools
The blogsphere has quickly called foul on the comparison and its reasoning. Greg Jorgensen provides a detailed rebuttal. Among the items he notes are that MapReduce is not a database but an algorithmic technique for distributed processing and should not be compared to one. Jorgensen proposes that a better comparison would have been to SimpleDB:
...What the authors really want to gripe about is distributed “cloud” data management systems like Amazon’s SimpleDB; in fact if you change “MapReduce” to “SimpleDB” the original article almost makes sense...
Rich Skrenta comments on the angle of disruption:
...The thing that disrupts you is always uglier and worse in some way. Less features, less developed. But if there's a 10X price win in there somewhere, the cheap rickety thing wins in the end. Think Linux vs. AT&T Unix, or mysql vs. Oracle...
All in all, it is sad to read such a misguided piece…
GridGain – Grid Computing Made Simple
more research needed.
Swapping it with SimpleDB or BigTable is a more logical perspective.
Also - if they were referring to BigTable, then in fact, it does support indexes and doesn't do brute force searches.
Prior to our development of MapReduce, the authors and many others
at Google implemented hundreds of special-purpose computations that
process large amounts of raw data, such as crawled documents, Web
request logs, etc., to compute various kinds of derived data, such as
inverted indices, various representations of the graph structure of Web
documents, summaries of the number of pages crawled per host, and
the set of most frequent queries in a given day. Most such computa-
tions are conceptually straightforward. However, the input data is usu-
ally large and the computations have to be distributed across hundreds
or thousands of machines in order to finish in a reasonable amount of
time. The issues of how to parallelize the computation, distribute the
data, and handle failures conspire to obscure the original simple com-
putation with large amounts of complex code to deal with these issues.
MapReduce is for doing computation on raw data. In Google's case this data is usually crawled from the web. Google likely stores some of the data they glean from raw data they process using MapReduce in a ... DBMS. *sigh*
InfoQ Sep 01, 2015