Researchers switch databases, just short of a petabyte
- By Joab Jackson
- Jun 30, 2004
The world will have to wait a while longer to see the first petabyte database.
The Stanford Linear Accelerator Center's Objectivity database, widely acknowledged as the world's largest, has stopped growing just short of 900T, a victim of database standardization. A petabyte is 1,000T.
Last fall, the Stanford University co-run International BaBar Collaboration switched its primary database for experimental data, said Jacek Becla, a database manager at the Menlo Park, Calif., center. The new database, nicknamed Kanga, runs the Root Object I/O System, a database format developed by CERN, the European Organization for Nuclear Research.
A smaller relational database management system serves the BaBar group for bookkeeping and indexing.
'Everyone in high-energy physics is using Root or planning to,' Becla said. 'We have students who want to reuse their knowledge on other experiments.'
Root is more familiar to outside researchers who access the BaBar data, and its tools are easier than the ones Stanford wrote for the Objectivity database, Becla said.
Last December, Winter Corp. of Waltham, Mass., recognized Stanford's Objectivity as the world's largest publicly known database. Each year, Winter publishes a list of the world's largest and fastest databases.
At last count the BaBar Objectivity database held 896T and, at its peak, was growing by about 1T a day. Most of the data was stored offline on magnetic tape. The database was named for its developer, Objectivity Inc. of Mountain View, Calif.Object-oriented
Objectivity/DB has a distributed, object-oriented format, whereas most of today's databases are relational.
'Objectivity seemed to have an architecture inherently scalable up to the amount of data we were going to handle,' said Richard Mount, head of computing at the Stanford center.
The center first considered Objectivity in 1986 when researchers from Stanford and Lawrence Berkeley National Laboratory were ramping up the International BaBar Collaboration. Their experiments involved crashing particle beams together inside a linear accelerator to understand the relationships between matter and antimatter.
The Lawrence Berkeley lab will keep the Objectivity data available and add small increments of new data. The database resides on about 100 older servers, mostly Sun Microsystems Inc. Sun Fire and Enterprise 4500 models.
Although Stanford publicized the fact that it ran the world's largest database, it had no plans to make an announcement when Objectivity reached the 1P benchmark.
'The main thing we do here is physics, not computer science,' Becla said. 'We were proud of Objectivity, and we were thinking it would hit a petabyte, but we were not storing the data for that.'
Kanga currently holds about 100T and is growing at about 1T a day.
Joab Jackson is the senior technology editor for Government Computer News.