Prev: How to specify the main() class in "jar -cfe ..." ?
Next: bundling efficiency, too good to be true?
From: peter on 31 Mar 2010 08:57 using java to create a key-value database for high-performance computing, is it a stupid idea? thanks from Peter (cmk128(a)hotmail.com)
From: markspace on 31 Mar 2010 10:30 peter wrote: > using java to create a key-value database for high-performance > computing, is it a stupid idea? > thanks > from Peter (cmk128(a)hotmail.com) Yes, a little. There are lots of databases already available, use one of them.
From: Lew on 31 Mar 2010 12:38 peter wrote: >> using java [sic] to create a key-value database for high-performance >> computing, is it a stupid idea? > markspace <nos...(a)nowhere.com> wrote: > Yes, a little. There are lots of databases already available, use one > of them. Depending on what one means by "database" and its projected size, one might need no more than a standard 'Map'. If that's not quite enough, use Derby (a.k.a. "Java DB", included with the JDK) or PostgreSQL. -- Lew
From: jaap on 31 Mar 2010 15:17 Op 31-03-10 14:57, schreef peter: > using java to create a key-value database for high-performance > computing, is it a stupid idea? > thanks > from Peter (cmk128(a)hotmail.com) Some large internetsites use this idea. Maybe http://hadoop.apache.org/hbase/ is what your looking for.
From: Arne Vajhøj on 31 Mar 2010 19:29 On 31-03-2010 08:57, peter wrote: > using java to create a key-value database for high-performance > computing, is it a stupid idea? No. That is quite common. But if you have to ask here, then I don't think a multi billion dollar company has hired you to design their new PB size databse. For ordinary computing a relational database is ready to be used. If you like the key-value API, then use a table with a key field and a BLOB value field. Arne
|
Next
|
Last
Pages: 1 2 Prev: How to specify the main() class in "jar -cfe ..." ? Next: bundling efficiency, too good to be true? |