On Wed, Jan 16, 2013 at 8:19 PM, Robert Haas <robertmhaas@gmail.com> wrote:
> On Tue, Jan 15, 2013 at 4:50 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>> I find the argument that this supports compression-over-the-wire to be
>> quite weak, because COPY is only one form of bulk data transfer, and
>> one that a lot of applications don't ever use. If we think we need to
>> support transmission compression for ourselves, it ought to be
>> integrated at the wire protocol level, not in COPY.
>>
>> Just to not look like I'm rejecting stuff without proposing
>> alternatives, here is an idea about a backwards-compatible design for
>> doing that: we could add an option that can be set in the connection
>> request packet. Say, "transmission_compression = gzip".
>
> But presumably this would transparently compress at one end and
> decompress at the other end, which is again a somewhat different use
> case. To get compressed output on the client side, you have to
> decompress and recompress. Maybe that's OK, but it's not quite the
> same thing.
Well, libpq could give some access to raw compressed streams, but,
really, even with double compression on the client, it solves the
bandwidth issue, not only for pg_dump, pg_restore, and copy, but also
for all other transfer-intensive applications. I do think it's the
best option.