|
Mammoth Database Wins Grand Prize
By Mason Inman
SLAC not only boasts the world’s longest building, but also the world’s
largest recognized database. The
BABAR
database—now containing about 900 terabytes, or 900,000
gigabytes—recently earned a Grand Prize in a prestigious annual survey
of very large databases, the Winter Corporation’s Top Ten.
 |
|
Award winners from SCS (l to r): Igor
Gaponenko,Yemi Adesanya, Artem Trunov, Daniel Wang, Wilko Kroeger,
Tofigh Azemoon and Jacek Becla. (Photo by
Diana Rogers) |
The information in the
BABAR
database, if printed out, would fill about 25 billion
books, nearly 1000 times the number of books in the world’s largest
library, the Library of Congress.
In 2002, the
BABAR
database group applied for the Winter Corporation award,
but since the database was a special type that didn’t fit into the
survey’s existing categories, it didn’t win an award. The group knew
that their database dwarfed the others, so they applied again in 2003,
not knowing what to expect.
“When I first submitted the survey, we didn’t have much hope,” said
Jacek Becla (SCS),
BABAR
database manager. As the date for the announcement of the
award got closer, however, the group had inklings that this year would
be different. “I sort of knew it was coming,” Becla said, “Too many
people started complaining that we were ignored last year.”
On March 3 the Winter Corporation announced the awards, with the
BABAR
database in first place in a newly created category: Most Data in a
Hybrid System. The database is called a hybrid because it uses both hard
disk drives and tape drives to store the data. The second place
database, run by the UK Meteorological Office, weighed in at 184
terabytes, about a fifth the size of the
BABAR
database.
While this award brings well-deserved recognition, the
BABAR
database is already well known within the database
community as a pioneering project, in part because of presentations by
the
BABAR
database team at database conferences.
“Database gurus are still very interested in things we’re doing,” said
Yemi Adesanya (SCS), a database developer on the project. Becla added,
“They want to hear the story of what we did and why we did it this way.”
When the database was being planned in 1996, the database group talked
to professors and professionals to find the best system for handling the
reams of data that would come pouring out of
BABAR.
No one was sure what system would be best, Becla said. The group took a
chance on a new object-oriented database technology created by
Objectivity, Inc. in Mountain View. Their leap of faith has paid off
famously.
Numerous DOE science projects—from combustion and proteomics studies to
climate models and nuclear physics—collect huge data sets, and many look
to the
BABAR
database for guidance, Becla said. “Everybody’s talking about petabytes
in the next couple of years. Everybody has similar issues to ours, so
there’s a lot of common ground.”
Many other databases are quickly catching up with
BABAR’s,
especially since all new
BABAR
data is being stored in another data storage system,
designed at CERN, rather than in the
BABAR
database. But for the near future, the group thinks they’ll hold their
title. “It’s very likely that we will be the largest next year,” Becla
said, “even though we’ve stopped growing.” |
