用Java开发VoltDB程序

jieforest

Q: Is there any recommended way to keep the db schema evolution in order (I.E. as a set of "patches")? Thank you.

A: VoltDB supports catalog updates where you can add and remove tables. You can also add, remove and update stored procedures. Over the coming weeks we’ll be adding additional support for online schema changes such as adding/dropping indexes and columns.

jieforest

Q: How well VoltDB gels with hibernate/mybatis or other java based persistence framework?

A: We do not currently support Hibernate or MyBatis directly.

jieforest

Q: You mentioned about the limitation of 50MB for the result set of a query, does it mean the developer should predetermine the result set size of the query and not use such?

A: Yes, you need to be sure that your result set is less than 50MB or the query or stored procedure will fail.

jieforest

Q: VoltDB is described as an in-memory database, but what makes it different from loading a complete Oracle database in memory?

A:  We’re dramatically faster than Oracle in memory.  Research at MIT, Brown and Yale have identified over 90% of processing in traditional databases can be classified as “overhead” leaving a scant 7% processing to the actual “work”.  This overhead includes such things as buffer management, locking, latching and logging.  VoltDB was architected from the ground up to eliminate this overhead by being in-memory, thus eliminating buffering, and to run each transaction in parallel and in isolation, thus eliminating the need to latch or lock.  When traditional databases are running in memory, they still carry the traditional overhead, or baggage.

jieforest

Thanks to everyone who tuned in and especially to those with who sent along some great questions. Don’t forget to check out the forums section on our community site for more information as well. Please join us next week when I will be presenting our next webinar “Writing VoltDB Apps in NodeJS” on October 11th.

over.

coldljy

有用于生产不