Mailing-List: contact cassandra-user-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cassandra-user@incubator.apache.org
Received-SPF: pass (nike.apache.org: domain of driftx@gmail.com designates
 209.85.221.193 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type;
        b=CcX+qGvyZpUIGKSGgElYWSzBB06IYjiBgMDvh+ix9rZkcnlNRgbtjlrdXcSe+DpOpL
         0PnelRJXoUV2oXVoVd+f2EYZN08F8LY1lbag5zVz0ZH3QzxBV14sg5BFWn2X7Co+LXRw
         mn/Xz4LlWG5pFagBzBKHzVFEcDk70HV8oQksU=
MIME-Version: 1.0
In-Reply-To: <ac68ab851003091114t665e9c95i64bea99517a7a72@mail.gmail.com>
References: <ac68ab851003090515y4b053009uda0fe63cd5f9db9e@mail.gmail.com>
	 <e06563881003090552u4c0602aftb2367bae79df6f18@mail.gmail.com>
	 <ac68ab851003090631x2da162dr99e803efd38a8266@mail.gmail.com>
	 <e06563881003090634o13653c31vbf4f80e2c066faf6@mail.gmail.com>
	 <ac68ab851003091114t665e9c95i64bea99517a7a72@mail.gmail.com>
Date: Tue, 9 Mar 2010 14:13:50 -0600
Message-ID: <cdc5ad201003091213i21b703fese824be06293f300b@mail.gmail.com>
Subject: Re: Bad read performances: 'few rows of many columns' vs 'many rows
	of few columns'
From: Brandon Williams <driftx@gmail.com>
To: cassandra-user@incubator.apache.org
Content-Type: multipart/alternative; boundary=0016361e7da8f074d4048163d1e6

--0016361e7da8f074d4048163d1e6
Content-Type: text/plain; charset=ISO-8859-1

On Tue, Mar 9, 2010 at 1:14 PM, Sylvain Lebresne <sylvain@yakaz.com> wrote:

> I've inserted 1000 row of 100 column each (python stress.py -t 2 -n
> 1000 -c 100 -i 5)
> If I read, I get the roughly the same number of row whether I read the
> whole row
> (python stress.py -t 10 -n 1000 -o read -r -c 100) or only the first column
> (python stress.py -t 10 -n 1000 -o read -r -c 1). And that's less that
> 10 rows by
> seconds.
>
> So sure, when I read the whole row, that almost 1000 columns by
> seconds, which is
> roughly 50M/s troughput, which is quite good. But when I read only the
> first column,
> I get 10 columns by seconds, that 500K/s, which is less good. Now,
> from what I've
> understood so far, cassandra doesn't deserialize whole row to read a
> single column
> (I'm not using supercolumn here), so I don't understand those numbers.
>

A row causes a disk seek while columns are contiguous.  So if the row isn't
in the cache, you're being impaired by the seeks.  In general, fatter rows
should be more performant than skinny ones.

-Brandon

--0016361e7da8f074d4048163d1e6
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div class=3D"gmail_quote">On Tue, Mar 9, 2010 at 1:14 PM, Sylvain Lebresne=
 <span dir=3D"ltr">&lt;<a href=3D"mailto:sylvain@yakaz.com" target=3D"_blan=
k">sylvain@yakaz.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_qu=
ote" style=3D"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex=
">

I&#39;ve inserted 1000 row of 100 column each (python stress.py -t 2 -n<br>
1000 -c 100 -i 5)<br>
If I read, I get the roughly the same number of row whether I read the whol=
e row<br>
(python stress.py -t 10 -n 1000 -o read -r -c 100) or only the first column=
<br>
(python stress.py -t 10 -n 1000 -o read -r -c 1). And that&#39;s less that<=
br>
10 rows by<br>
seconds.<br>
<br>
So sure, when I read the whole row, that almost 1000 columns by<br>
seconds, which is<br>
roughly 50M/s troughput, which is quite good. But when I read only the<br>
first column,<br>
I get 10 columns by seconds, that 500K/s, which is less good. Now,<br>
from what I&#39;ve<br>
understood so far, cassandra doesn&#39;t deserialize whole row to read a<br=
>
single column<br>
(I&#39;m not using supercolumn here), so I don&#39;t understand those numbe=
rs.<br></blockquote><div><br></div><div>A row causes a disk seek while colu=
mns are contiguous. =A0So if the row isn&#39;t in the cache, you&#39;re bei=
ng impaired by the seeks. =A0In general, fatter rows should be more perform=
ant than skinny ones.</div>
<div><br></div><div>-Brandon</div>
</div>

--0016361e7da8f074d4048163d1e6--