DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com;
	h=Message-ID:X-YMail-OSG:Received:X-Mailer:Date:From:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
	b=bZWoXCQYgTodZgMJz2ctmaZc2uviBFskVkLQgVrkcOF1a7xc1PGSwVSxwrFkGxfjKVCBLtUD1LAM/qkGfRVek3Fb+8Uve2vPClOjZhEiXVCMgz8cyN7BzTO2U8JUhiPhEVTyal6Uaj7THfCUIp4hz4cj8u5SsKRnKcWTloLx8NI=;
Message-ID: <535760.62572.qm@web111307.mail.gq1.yahoo.com>
Date: Mon, 28 Mar 2011 16:54:50 -0700 (PDT)
From: Andy <angelflow@yahoo.com>
Subject: Re: Intel SSDs that may not suck
To: "pgsql-performance@postgresql.org" <pgsql-performance@postgresql.org>,
	Greg Smith <greg@2ndQuadrant.com>
In-Reply-To: <4D90EDB6.2080609@2ndQuadrant.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

This might be a bit too little too late though. As you mentioned there real=
ly isn't any real performance improvement for the Intel SSD. Meanwhile, San=
dForce (the controller that OCZ Vertex is based on) is releasing its next g=
eneration controller at a reportedly huge performance increase.=0A=0AIs the=
re any benchmark measuring the performance of these SSD's (the new Intel vs=
. the new SandForce) running database workloads? The benchmarks I've seen s=
o far are for desktop applications.=0A=0AAndy=0A=0A--- On Mon, 3/28/11, Gre=
g Smith <greg@2ndQuadrant.com> wrote:=0A=0A> From: Greg Smith <greg@2ndQuad=
rant.com>=0A> Subject: [PERFORM] Intel SSDs that may not suck=0A> To: "pgsq=
l-performance@postgresql.org" <pgsql-performance@postgresql.org>=0A> Date: =
Monday, March 28, 2011, 4:21 PM=0A> Today is the launch of Intel's 3rd=0A> =
generation SSD line, the 320 series.=A0 And they've=0A> finally produced a =
cheap consumer product that may be useful=0A> for databases, too!=A0 They'v=
e put 6 small capacitors=0A> onto the board and added logic to flush the wr=
ite cache if=0A> the power drops.=A0 The cache on these was never very=0A> =
big, so they were able to avoid needing one of the big=0A> super-capacitors=
 instead.=A0 Having 6 little ones is=0A> probably a net reliability win ove=
r the single point of=0A> failure, too.=0A> =0A> Performance is only a litt=
le better than earlier generation=0A> designs, which means they're still be=
hind the OCZ Vertex=0A> controllers that have been recommended on this list=
.=A0 I=0A> haven't really been hearing good things about long-term=0A> reli=
ability of OCZ's designs anyway, so glad to have an=0A> alternative.=A0 *Im=
portant*:=A0 don't buy SSD for=0A> important data without also having a goo=
d redundancy/backup=0A> plan.=A0 As relatively new technology they do still=
 have=0A> a pretty high failure rate.=A0 Make sure you budget for=0A> two d=
rives and make multiple copies of your data.=0A> =0A> Anyway, the new Intel=
 drivers fast enough for most things,=0A> though, and are going to be very =
inexpensive.=A0 See http://www.storagereview.com/intel_ssd_320_review_300gb=
=0A> for some simulated database tests.=A0 There's more about=0A> the inter=
nals at http://www.anandtech.com/show/4244/intel-ssd-320-review=0A> and the=
 white paper about the capacitors is at http://newsroom.intel.com/servlet/J=
iveServlet/download/38-4324/Intel_SSD_320_Series_Enhance_Power_Loss_Technol=
ogy_Brief.pdf=0A> =0A> Some may still find these two cheap for enterprise u=
se,=0A> given the use of MLC limits how much activity these drives=0A> can =
handle.=A0 But it's great to have a new option for=0A> lower budget system =
that can tolerate some risk there.=0A> =0A> -- Greg Smith=A0=A0=A02ndQuadra=
nt US=A0 =A0=0A> greg@2ndQuadrant.com=A0=A0=A0Baltimore,=0A> MD=0A> Postgre=
SQL Training, Services, and 24x7 Support=A0=0A> www.2ndQuadrant.us=0A> "Pos=
tgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books=0A> =0A> =
=0A> -- Sent via pgsql-performance mailing list (pgsql-performance@postgres=
ql.org)=0A> To make changes to your subscription:=0A> http://www.postgresql=
.org/mailpref/pgsql-performance=0A> =0A=0A=0A