Re: One process per session lack of sharing - Mailing list pgsql-hackers

From AMatveev@bitec.ru
Subject Re: One process per session lack of sharing
Date
Msg-id 06873259.20160718121754@bitec.ru
Whole thread Raw
In response to Re: One process per session lack of sharing  (Pavel Stehule <pavel.stehule@gmail.com>)
List pgsql-hackers
Hi


>  Such a host delegate process could be explicitly built with
> multithread support and not 'infect' the rest of the code with its requirements.
>  
>  Using granular RPC is nice for isolation but I am concerned that the latencies might be high.
I agree with you.
Moreover I think that some decision have not sense with this "thread model" in any way.
For example Embedded oracle XML DB applications:
http://docs.oracle.com/cd/B28359_01/appdev.111/b28369/xdb23jv1.htm#i1043708
"""
You can run a Java servlet.
Servlets work better as the top-level entry point into Oracle Database, and require using HTTP(S) as the protocol to
accessOracle Database.
 
"""

> What I know about Oracle, PL/SQL, Java - all is executed as
> outprocess calls. I am sure, so PL doesn't share process with SQL engine there

It's better to say that java is executed like outprocess calls.
It's done for wide libraries support.
But it is not separate  process.
"""
The JDBC server-side internal driver,
the Oracle JVM, the database, and the SQL engine all run within the same address space, and therefore, the issue of
networkround-trips is irrelevant
 
"""
http://docs.oracle.com/cd/E11882_01/java.112/e16548/overvw.htm#JJDBC28026




pgsql-hackers by date:

Previous
From: Antonin Houska
Date:
Subject: Typos in logical decoding
Next
From: AMatveev@bitec.ru
Date:
Subject: Re: One process per session lack of sharing