Mailing-List: contact user-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hadoop.apache.org
MIME-Version: 1.0
Sender: shaposhnik@gmail.com
In-Reply-To: 
 <CAGHyZ6J-p_k3oY5F6tSs5kxjRhYuQ0PaY8cmD9qkpsXHbp2DHQ@mail.gmail.com>
References: 
 <CAN5siX6QRu0B+06Z=38N13K92zSs_pnw_NZykGgLQ4kV0f+8tw@mail.gmail.com>
	<D1C959ED.DD045%james@123dm.nl>
	<CAGHyZ6J-p_k3oY5F6tSs5kxjRhYuQ0PaY8cmD9qkpsXHbp2DHQ@mail.gmail.com>
Date: Mon, 13 Jul 2015 11:53:56 -0700
Message-ID: 
 <CA+ULb+uao8y6xf=pxWo6Bci4a4M9NNjB-C4pvMBzLv5wytOYOg@mail.gmail.com>
Subject: Re: Hadoop or RDBMS
From: Roman Shaposhnik <roman@shaposhnik.org>
To: user@hadoop.apache.org
Content-Type: text/plain; charset=UTF-8

On Mon, Jul 13, 2015 at 7:27 AM, Sean Busbey <busbey@cloudera.com> wrote:
> Given the relatively modest dataset size, this sounds like a straight
> forward use case for a traditional RDBMS.

To pile on top of Sean's reply I'd say that given the current estimates a
traditional RDBMS such as Postgres could fit the bill. If the potential
scalability needs to be taken into account you could look at MPP-type
of solutions to scale from Postgres. Ping me off-list if you're interested
in that route.

To get us back on topic for user@hadoop, I'd say that Hadoop ecosystem
strategy for your use case has to be warranted by more than a size
of data. In fact, I'd say that the Hadoop and/or Spark would only make
sense to you if you feel like taking advantage of various analytical
frameworks available for those.

Just my 2c.

Thanks,
Roman.