Given the factor of an 80-fold difference in this record-parsing benchmark, we suspect the absolute numbers for the Hadoop benchmarks in the comparison paper are inflated...
Whole article could be argued to be fighting words, considering they basically point out flawed assertions in the comparison paper" over and over again. Though catching an 80-fold difference is pretty ridiculous.
Looks like the author's of the "comparison paper" have a new paper out (the byline says Jan 1 2010) titled "MapReduce and Parallel DBMSs: Friends or Foes?" which seems to take a different approach at characterizing Map/Reduce, but unfortunately the contents are only available for premium members.
Parallel DBMSs excel at efficient querying of large data sets; MapReduce-style systems excel at complex analytics and ETL tasks. Neither is good at what the other does well. Hence, the two technologies are complementary.
Whoa! Those are some serious fighting words!