Re:A database built like an operating system: the architectu...
by ablock on Friday March 31, 2017 @06:47PM
(#14009066)
So Daytona compiles and runs a new binary for each query, not just for each connection, and does all its coordination through files + pipes + shared memory. Query binaries even fork themselves for parallelism(!). My understanding is that the cost of switching between kernel and user space is pretty overwhelming. Maybe STM would help that someday but now you’ve implemented an actual database in the kernel...
PostgreSQL’s process per user model makes way more sense but as you say still has upper limits on concurrency.
Re:A database built like an operating system: the architectu...
by eternalban on Friday March 31, 2017 @06:48PM
(#14009077)
Not sure how to read that. Isn't simply 1 process per connection?
Re:A database built like an operating system: the architectu...
by sargun on Saturday April 01, 2017 @02:06AM
(#14010767)
Yes. But you can't run more than one query per connection.
Re:A database built like an operating system: the architectu...
by eternalban on Saturday April 01, 2017 @08:58AM
(#14011883)
One executing query per connection, meaning a connection maintains a serialized query processing model.
Although it doesn't handle unbounded concurrency, there are techniques (like manual locking, or pgbouncer) to deal with this.