Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (athena.apache.org: domain of greenemj@gmail.com designates
 209.85.222.199 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:from:date:message-id:subject:to
         :content-type;
        b=MpzhRo25Wu8Z/+V/L+Cj4Tm/gCbs9hOI6RCjLqhYEMvau91lvmp5dPY/xibaZFWR2u
         SBcgi2T2olQZh9gXUKP5DYh8QnrMtTw9LoCXPyH2j2fPjdNXfzy8oZ6fySPbzo/CZOMh
         ykjfx9yKXbS4y0A29jklvNjNHZsRbNcl60+u4=
MIME-Version: 1.0
In-Reply-To: <AANLkTikRY7acZSZ_LMeUeCuhX4Mcx2RaPCmCuZseSRND@mail.gmail.com>
References: <AANLkTilzlNZ8clRJjx2hEyZP2gJsUHZpvO2MAiCBdhtB@mail.gmail.com>
	<AANLkTikRY7acZSZ_LMeUeCuhX4Mcx2RaPCmCuZseSRND@mail.gmail.com>
From: Mark Greene <greenemj@gmail.com>
Date: Tue, 25 May 2010 07:04:08 -0400
Message-ID: <AANLkTilSJjvYsfMhoKVGi1oNpipAe114QZDrYUKLphkK@mail.gmail.com>
Subject: Re: Why are writes faster than reads?
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=000e0cd13beefd9e300487691e68

--000e0cd13beefd9e300487691e68
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

I'm fairly certain the write path hits the commit log first, then the
memtable.

2010/5/25 Peter Sch=FCller <scode@spotify.com>

> > I have seen several off-hand mentions that writes are inherently faster
> than
> > reads. Why is this so?
>
> I believe the primary factor people are referring to is that writes
> are faster than reads in terms of disk I/O because writes are
> inherently sequential. Writes initially only happen in-memory plus in
> a (sequentially written) commit log; when flushed out to an sstable
> that is likewise sequential writing.
>
> Reads on the other hand, to the extent that they go down to disk, will
> suffer the usual overhead associated with disk seeks.
>
> See http://wiki.apache.org/cassandra/ArchitectureInternals for details.
>
> --
> / Peter Schuller aka scode
>

--000e0cd13beefd9e300487691e68
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

I&#39;m fairly certain the write path hits the commit log first, then the m=
emtable.<br><br><div class=3D"gmail_quote">2010/5/25 Peter Sch=FCller <span=
 dir=3D"ltr">&lt;<a href=3D"mailto:scode@spotify.com">scode@spotify.com</a>=
&gt;</span><br>

<blockquote class=3D"gmail_quote" style=3D"margin: 0pt 0pt 0pt 0.8ex; borde=
r-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;"><div class=3D"im"=
>&gt; I have seen several off-hand mentions that writes are inherently fast=
er than<br>


&gt; reads. Why is this so?<br>
<br>
</div>I believe the primary factor people are referring to is that writes<b=
r>
are faster than reads in terms of disk I/O because writes are<br>
inherently sequential. Writes initially only happen in-memory plus in<br>
a (sequentially written) commit log; when flushed out to an sstable<br>
that is likewise sequential writing.<br>
<br>
Reads on the other hand, to the extent that they go down to disk, will<br>
suffer the usual overhead associated with disk seeks.<br>
<br>
See <a href=3D"http://wiki.apache.org/cassandra/ArchitectureInternals" targ=
et=3D"_blank">http://wiki.apache.org/cassandra/ArchitectureInternals</a> fo=
r details.<br>
<font color=3D"#888888"><br>
--<br>
/ Peter Schuller aka scode<br>
</font></blockquote></div><br>

--000e0cd13beefd9e300487691e68--