Re: Vertica targeting PostgreSQL users

On Mon, Nov 20, 2017 at 4:43 PM, Jonathan S. Katz <jkatz(at)postgresql(dot)org>
wrote:

> Hi,
>
> Vertica current marketing is heavily targeting Postgres users:
>
> https://www.vertica.com/postgresql/
>
> and features this directly from their homepage. I also know of at least
> one campaign that has gone out around this.
>
> The good news for our community is that because of all the advances in
> Postgres and the overall growth to our user base, there are database
> companies looking to win users over to their platforms. This is also a
> good learning experience for us too: because we have so many users, we also
> have to look at user retention to ensure that we have enough features and
> available knowledge to manage Postgres. I have not read the case study
> that they proposed, but knowing the group of people who work on Postgres
> and our software ecosystem, we may have been able to retain those databases
> on Postgres.
>

To my mind I am not that concerned about who moves away to PostgreSQL.
People migrate for various reasons. Some of those are technical. Some of
them are political. In the end whether a given user migrates is not,
really, in our control. I don't think focusing on "why did we lose these"
is helpful. What is helpful instead is understanding how to do the same
sorts of things on PostgreSQL as Vertica.

>
> Going forward, we must continue to understand our users’ needs while
> ensuring that we can provide them as many resources as possible to help
> them manage Postgres and show them that here is great help available when
> it is required.
>

There are other important topics here in analytics workloads. Up until
this year most of my experience, for example, was in workflows which were
mixed buy oriented primarily towards OLTP engines. So things like ERP,
scientific data processing and the like. Now I have stepped into analytics
and am working in a very large environment. The big differences in terms
of how you handle sparse matrices and other sorts of things crop up.
Systems like Cassandra and Vertica solve some analytics problems but they
don't solve all problems extremely well. They have their own caveats and
learning curves. At least as far as our experience at Adjust we have
typically solved these problems ourselves.

The arguments about the development direction of PostgreSQL in my view is
similarly problematic. Companies contribute what their users need. And
that's good. But almost all the literature published about PostgreSQL
focuses on OLTP problems and solutions. What we are actually missing is a
series of good solid papers, discussions, and so forth regarding how to do
OLAP on PostgreSQL effectively. This is an area where PostgreSQL is
rapidly growing but without guidance, people aren't going to know how to
think about, model, and approach a lot of these problems.

So I am going to suggest a different approach, namely that we see what we
can do to create space and exposure for stuff discussing OLAP etc. on
PostgreSQL. Maybe a special category of case studies and maybe contributed
white papers or the like. But this is an area where a lot of knowledge
sharing could be very helpful.

>
> Best,
>
> Jonathan
>
>

--
Best Wishes,
Chris Travers

Efficito: Hosted Accounting and ERP. Robust and Flexible. No vendor
lock-in.
http://www.efficito.com/learn_more