Aster Data has built a business out of
map/reduce, and the release today of a thousand new SQL query building
blocks—what it calls "functions"—is designed to give business users
access to map/reduce computed analytics.
Whereas open-source big data solution Hadoop, for instance, is based entirely
on its own implementation of map/reduce, as well as a set of homegrown query
structuring frameworks like Pig and Hive, Aster Data wants to use SQL right
from the start. Sharmila Shahani-Mulligan, executive vice president of
marketing at Aster Data, said that this is a significant advantage over Hadoop.
“Hadoop lends itself more to batch-type processing. Most of our customers are
running analytics on a daily basis with the expectation of results returned
every few minutes," she said. "It's not real-time, but it's near
real-time.
"The second advantage is SQL map/reduce. We are literally targeting the
business analyst with SQL using full map/reduce underneath.”
Map/reduce is the framework for processing huge amounts of data, and it is the
basis of the Apache Hadoop project, as well as of Big Table, which runs
Google's search engine. Using map/reduce, huge stores of data can be processed,
and the results can be combined into a cohesive set of information.
Stephanie McReynolds, director of product marketing at Aster Data, said the new
sets of query-building tools aren't limited to business users. “We introduced
many new business analyst-ready functions," she said. "[These]
functions address particular business issues, like path analysis for website
traffic.
"We also have a series of packages for power users. These are for people
building their own SQL map/reduce applications. They want to use Java or C
functions to get ahead. These are smaller building blocks."
Shahani-Mulligan said that Aster Data's analytics can be tweaked and queried by
business users, a major advantage over Hadoop. She said that many business
users already know SQL, which cannot be said of Hive or Pig. She said that with
Hadoop, developers likely need to be called in to implement any analytics
batches that need to be run, but with Aster Data, the business users can do
that themselves.
“With almost any of our [customers] you talk to, one of the big appeals has
been that their existing business analysts can work with functions and don't
have to use a new language," said Shahani-Mulligan. "This is why we
came out with SQL map/reduce. Some of them also have Hadoop, but it requires
you to do constant programming in map/reduce versus having a simple-to-use
interface."
By Alex Handy
Be the first to rate this post
- Currently 0/5 Stars.
- 1
- 2
- 3
- 4
- 5