Re: proposal GSoC 2015 task: Allow access to the database via HTTP - Mailing list pgsql-hackers

2015-03-23 18:10 GMT+01:00 Вадим Горбачев <bmsdave@gmail.com>:

>> What is a benefit of this implementation for Postgres community? 

It will be possible to clean this task from https://wiki.postgresql.org/wiki/Todo

>> It can be interesting as well integrated project to Postgres - implemented in C as background worker (if it possible)

I.e. as I understand http_api has to work in separate process at the server of the DBMS.

And why not to start the free-standing application?

In this case there will be great opportunities for scaling:

1. through one http_api appendix it will be possible to be connected to several DB

2. to use pgpool

3. to share burden between servers and so forth.

4. to carry out other functions of a frontend.

And what pluses writing of background worker will give?

integration to Postgres - simply (usual) deployment. Fast access to PostgreSQL shared memory, fast access to database.

Now anybody can using pgpool, pgbouncer, light httpd,  simple fast CGI script without problems.

If background worker is necessary, I am ready to realize it and it will be interesting to me.

But whether it will be more necessary for postgresql, than the free-standing application?

2015-03-23 19:24 GMT+03:00 Pavel Stehule <pavel.stehule@gmail.com>:

Hi

2015-03-22 23:28 GMT+01:00 Вадим Горбачев <bmsdave@gmail.com>:

Hi Team.

I would like to solve a problem of "Allow access to the database via HTTP".

But before drawing up the demand in GSOC I wanted to consult here. Therefore I will be grateful to comments from attendees here!

1. I think, will better use access to DB through the stand-alone program which not necessarily has to be on the same server. At least because it will give certain freedom in cluster systems.

2. Whether it is obligatory to use a programming language C for this purpose? After all as the stand-alone program ( frontend ) it has to be not necessarily written in the same programming language as the server ( backend ). I would prefer to use the python language for writing as I consider that this language is more clear to system administrators + to bring much more simply editings in a code.

What is a benefit of this implementation for Postgres community?

Proposed implementation is few lines in Python - and it is not big benefit for us. More, similar project exists.

It can be interesting as well integrated project to Postgres - implemented in C as background worker (if it possible)

Regards

Pavel

 

3. What you will advise what to pass a selection stage in GSOC 2015 from postgresql?)

PS: my English is poor. I ask you to forgive me for it.

Best Regards,
Vadim Gorbachov