<?xml version="1.0" encoding="utf-8"?>

<?xml-stylesheet type="text/css" href="style/rss.css"?>

<!DOCTYPE rss PUBLIC "-//IDN Silmaril.IE/DTD RSS 2.0//EN"
  "http://WWW.Silmaril.IE/software/rss2.dtd">

<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>sabi notes</title>
    <atom:link rel="self" type="application/rss+xml"
      href="http://www.sabi.co.uk/changes.xml"/>
    <link>http://www.sabi.co.uk/</link>
    <description>Notes mostly about Linux and computer issues</description>
    <language>en-GB</language>

    <docs>http://blogs.law.Harvard.edu/tech/rss/</docs>
    <webMaster>notes {at} notes.for.sabi.co.UK</webMaster>
    <copyright>Copyright 2008 Peter G.</copyright>
    <ttl>14000</ttl>

    <category>AMD</category>
    <category>Intel</category>
    <category>Microsoft</category>
    <category>Nintendo</category>
    <category>Sony</category>

    <category>CPU</category>
    <category>OS</category>
    <category>VoIP</category>
    <category>file sytem</category>
    <category>free software</category>
    <category>hardware</category>
    <category>performance</category>
    <category>system administration</category>
    <category>videogames</category>
    <category>virtual memory</category>
    
    <category>ALSA</category>
    <category>Debian</category>
    <category>Fedora</category>
    <category>GNU</category>
    <category>JFS</category>
    <category>KDE</category>
    <category>Linux</category>
    <category>MS Windows 2000</category>
    <category>MS Windows XP</category>
    <category>MS Windows</category>
    <category>PS3</category>
    <category>RedHat</category>
    <category>SUSE</category>
    <category>UNIX</category>
    <category>XFS</category>
    <category>Xbox 360</category>
    <category>ext3</category>

    <lastBuildDate>Sun, 10 May 2009 18:55:00 +0100</lastBuildDate>

<!--
    Regexps to convert titles to items.

  <dt class="h"><a name="\(..\)\(..\)\(..\)\([a-z]?\)" href="#\1\2\3\4">\1\2\3\4 *\([A-Z][a-z][a-z]\)?</a>[	 
]*\([^<]+\)</dt>

<item>
      <title>\6</title>
      <link>http://WWW.sabi.co.UK/blog/\1\2%\2.html#\1\2\3\4</link>

      <description></description>

      <category></category>
      <category></category>
      <category></category>
      <category></category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/\1\2%\2.html#\1\2\3\4</guid>
      <pubDate>\5, \3 %\2 20\1 18:10:00 +0100</pubDate>
    </item>
-->

    <item>
      <title>Switching to Atom from RSS</title>
      <link>http://WWW.sabi.co.UK/blog/0904Apr.html?090418#090418</link>

      <description>I have decided to switch to Atom as the feed
        format for this site and blog, at http://WWW.sabi.co.UK/changed.atom
        and this RSS format feed will no longer be updated. This is the
        last entry that will appear in this feed.</description>

      <category>web</category>
      <category>blog</category>
      <category>XML</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0904Apr.html?090418#090418</guid>
      <pubDate>Sat, 18 Apr 2009 21:30:00 +0100</pubDate>
    </item>

    <item>
      <title>Impressive JFS and eSATA performance</title>
      <link>http://WWW.sabi.co.UK/blog/0901Jan.html?090131#090131</link>

      <description>JFS and eSATA perform fairly well when synchronizing
	a largish filesystem to an external drive. Not surprising, but
        still worth keeping in mind.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>file system</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0901Jan.html?090131#090131</guid>
      <pubDate>Sat, 31 Jan 2009 10:15:00 +0000</pubDate>
    </item>

    <item>
      <title>My laptop and cost of spares, Linux power management</title>
      <link>http://WWW.sabi.co.UK/blog/0812Dec.html?081227#081227</link>

      <description>Laptop prices have decreased so much that repairing them
        is not longer so useful, and Linux power management has
        greatly improved.</description>

      <category>Linux</category>
      <category>power</category>
      <category>hardware</category>
      <category>laptop</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0812Dec.html?081227#081227</guid>
      <pubDate>Sat, 27 Dec 2008 18:10:00 +0100</pubDate>
    </item>

    <item>
      <title>More laptop battery decay, new high capacity battery</title>
      <link>http://WWW.sabi.co.UK/blog/0812Dec.html?081224#081224</link>

      <description>Lithium-ion batteries age, mine has aged very rapidly
      especially when warm, and I'll be making minimal use of my new
      high capacity one.</description>

      <category>hardware</category>
      <category>Linux</category>
      <category>system administration</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0812Dec.html?081224#081224</guid>
      <pubDate>Wed, 24 Dec 2008 18:10:00 +0100</pubDate>
    </item> <!-- 19:10 -->

    <item>
      <title>Updated Linux power saving script</title>
      <link>http://WWW.sabi.co.UK/#sourcesSabiwatts?081202</link>

      <description>This script is designed to tweak the power
	related settings of the Linux kernel for normal operation
	when on permanent power and for low power operation when
	on battery power. This update fixes two mistakes in the
	X display enabling/disabling feature.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>laptop</category>
      <category>shell</category>
      <category>configuration</category>
      <category>power</category>

      <pubDate>Tue, 02 Dec 2008 19:00:00 +0000</pubDate>
    </item>

    <item>
      <title>Enabled video outputs can always draw power</title>
      <link>http://WWW.sabi.co.UK/blog/0812Dec.html?081201#081201</link>

      <description>Just discovered that an enabled but
        disconnected video output can draw a few watts of
        power on a laptop.</description>

      <category>system administration</category>
      <category>power</category>
      <category>laptop</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0812Dec.html?081201#081201</guid>
      <pubDate>Mon, 01 Dec 2008 19:45:00 +0000</pubDate>
    </item>

    <item>
      <title>Updated Linux power saving script</title>
      <link>http://WWW.sabi.co.UK/#sourcesSabiwatts?081201</link>

      <description>This script is designed to tweak the power
	related settings of the Linux kernel for normal operation
	when on permanent power and for low power operation when
	on battery power. This update adds disabling X11 video outputs
	when disconnected in low power mode and enabling connected ones
	when in normal power mode.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>laptop</category>
      <category>shell</category>
      <category>configuration</category>
      <category>power</category>

      <pubDate>Mon, 01 Dec 2008 19:35:00 +0000</pubDate>
    </item>

    <item>
      <title>Quick loss of battery capacity</title>
      <link>http://WWW.sabi.co.UK/blog/0811Nov.html?081124#081124</link>

      <description>After 9 months of use, my laptop battery now
        lasts 2 hours instead of 3.5 hours when new. It is in effect
        a consumable.</description>

      <category>hardware</category>
      <category>power</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0811Nov.html?081124#081124</guid>
      <pubDate>Mon, 24 Nov 2008 19:00:00 +0000</pubDate>
    </item> <!-- 19:00 -->

    <item>
      <title>Routing with OSPF and ECMP</title>
      <link>http://WWW.sabi.co.UK/blog/0810Oct.html?081014#081014</link>

      <description>Automatic routing discovery with OSPF and parallel
	routing with ECMP combine to give a fairly simple but very
	resilient network design</description>

      <category>system administration</category>
      <category>networking</category>
      <category>architecture</category>
      <category>design</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0810Oct.html?081014#081014</guid>
      <pubDate>Tue, 14 Oct 2008 19:50:00 +0000</pubDate>
    </item> <!-- 19:50 -->

    <item>
      <title>Syntax of Konqueror Advanced Web Shortcuts</title>
      <link>http://WWW.sabi.co.UK/Notes/linuxKDE.html#docsAWS?080912</link>
      <description>From a few sources I have compiled a hopefully
	complete summary of the Advanced Web Shortcuts that are
	quite convenient in Konqueror.</description>
      <category>KDE</category>
      <category>documentation</category>
      <category>Linux</category>

      <pubDate>Fri, 12 Sep 2008 19:10:00 +0100</pubDate>
    </item>

    <item>
      <title>Second guessing the UNIX security system</title>
      <link>http://WWW.sabi.co.UK/blog/0808Aug.html?080826#080826</link>

      <description>Programs that are note 'setuid' should not
        override the POSIX/UNIX security system and refuse to
        run unless the user has 'root' privileges: just checking
        for user id 0 pre-empts different security architectures,
        and it is pointless and obnoxious anyhow.</description>

      <category>UNIX</category>
      <category>Linux</category>
      <category>software engineering</category>
      <category>security</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0808Aug.html?080826#080826</guid>
      <pubDate>Tue, 26 Aug 2008 18:30:00 +0100</pubDate>
    </item> <!-- 18:30 -->

    <item>
      <title>A new log structured file system design</title>
      <link>http://WWW.sabi.co.UK/blog/0808Aug.html?080822#080822</link>

      <description>A new log structured file system has been
	developed in Japan, and it seems quite mature, to the
	point it can run extensive performance tests. A log
	structured filesystems aligns well with </description>

      <category>file systems</category>
      <category>Linux</category>
      <category>software engineering</category>
      <category>performance</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0808Aug.html?080822#080822</guid>
      <pubDate>Fri, 22 Aug 2008 19:10:00 +0100</pubDate>
    </item> <!-- 19:10 -->

    <item>
      <title>Another example of incomplete, misleading error message</title>
      <link>http://WWW.sabi.co.UK/blog/0808Aug.html?080812#080812</link>

      <description>Cryptic and misleading error messages seem to me
	fairly common with Kerberos and in this case Kerberos-related
        software. Much like MS-Windows or MS-Windows related software.
	This cultural compatibility may also be why Active Directory
        is based on Kerberos.</description>

      <category>software engineering</category>
      <category>Linux</category>
      <category>MS-Windows</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0808Aug.html?080812#080812</guid>
      <pubDate>Tue, 12 Aug 2008 21:30:01 +0100</pubDate>
    </item> <!-- 21:30 -->

    <item>
      <title>Java/JVM and C#/CIL, the patent angle</title>
      <link>http://WWW.sabi.co.UK/blog/0808Aug.html?080807#080807</link>

      <description>C#/CIL are somewhat more polished and nicer
	versions of Java/JVM, but there is a large risks that
        Microsoft owns enough patents to exclude if it wishes
        independent implementations of C#/CIL.</description>

      <category>software engineering</category>
      <category>patents</category>
      <category>Microsoft</category>
      <category>Java</category>
      <category>C#</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0808Aug.html?080807#080807</guid>
      <pubDate>Thu, 07 Aug 2008 06:45:01 +0100</pubDate>
    </item> <!-- 06:45 -->

    <item>
      <title>New large Samsung flash drives</title>
      <link>http://WWW.sabi.co.UK/blog/0808Aug.html?080805#080805</link>

      <description>Some new flash hard drives have been launched
	and they have remarkable speed, low weight and low power
	consumption. Surely good for laptops, but also for high
	density mass storage arrays.</description>

      <category>storage</category>
      <category>hardware</category>
      <category>performance</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0808Aug.html?080805#080805</guid>
      <pubDate>Tue, 05 Aug 2008 19:00:01 +0100</pubDate>
    </item> <!-- 19:00 -->

    <item>
      <title>Storage systems, indices and sequential access</title>
      <link>http://WWW.sabi.co.UK/blog/0807Jul.html?080718#080718</link>

      <description>Indices in data storage systems are used to avoid
	sequentially a large dataset when only a subset is sought,
	but for both RAM and disks bulk sequential access is becoming
	ever faster than random access, making the usefulness of indices
	lower with time. But flash devices may change that for mass
	storage at least.</description>

      <category>performance</category>
      <category>hardware</category>
      <category>DBMS</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0807Jul.html?080718#080718</guid>
      <pubDate>Fri, 18 Jul 2008 08:05:00 +0100</pubDate>
    </item> <!-- 08:05 -->

    <item>
      <title>When large application services are unavailable</title>
      <link>http://WWW.sabi.co.UK/blog/0807Jul.html?080711#080711</link>

      <description>An article talks about online service unavailability
	but does not make a vital distinction: when a web shop is
	not available that's inconvenient because one cannot buy
	or read somebody else's data, when a web service is
        not available that's bad because one cannot sell or read one's
	own data.</description>

      <category>web</category>
      <category>business</category>
      <category>reliability</category>
      <category>architecture</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0807Jul.html?080711#080711</guid>
      <pubDate>Fri, 11 Jul 2008 06:55:00 +0100</pubDate>
    </item> <!-- 06:55 -->

    <item>
      <title>Updated Linux power saving script</title>
      <link>http://WWW.sabi.co.UK/#sourcesSabiwatts?080709</link>

      <description>This script is designed to tweak the power
	related settings of the Linux kernel for normal operation
	when on permanent power and for low power operation when
	on battery power. Thus update adds setting the hard disk
        APM level which is useful to reduced load cycles when
        on permanent power.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>laptop</category>
      <category>shell</category>
      <category>configuration</category>
      <category>power</category>

      <pubDate>Wed, 09 Jul 2008 19:25:00 +0100</pubDate>
    </item>

    <item>
      <title>Good Linux power management</title>
      <link>http://WWW.sabi.co.UK/blog/0807Jul.html?080709#080709</link>

      <description>Thanks to suitable investigation in Linux
        power management, I am getting several hours of work
        on my 6 cells battery laptop. The good quality of the
        LCD screen and backlight look quite important.</description>

      <category>system administration</category>
      <category>hardware</category>
      <category>performance</category>
      <category>power consumption</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0807Jul.html?080709#080709</guid>
      <pubDate>Wed, 09 Jul 2008 08:50:00 +0100</pubDate>
    </item> <!-- 08:50 -->

    <item>
      <title>UML and future of American Programmers</title>
      <link>http://WWW.sabi.co.UK/blog/0807Jul.html?080708#080708</link>

      <description>Someone thinks that analyasts and users will
	do their own programming thanks to advanced tools, but
	while this is an old unrealized dream, the bigger issue
	is that the global price for programmers is much lower
	than the cost of analyst or user time.</description>

      <category>software engineering</category>
      <category>outsourcing</category>
      <category>offshoring</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0807Jul.html?080708#080708</guid>
      <pubDate>Tue, 08 Jul 2008 06:35:00 +0100</pubDate>
    </item> <!-- 06:35 -->

   <item>
      <title>Penny foolish is not pounds wise for applications</title>
      <link>http://WWW.sabi.co.UK/blog/0807Jul.html?080704#080704</link>

      <description>Current individual computers are often powerful
	enough to run even poorly written applications tolerably,
        but I would rather use that power to run several applications
        well instead of one tolerably.</description>

      <category>performance</category>
      <category>system administration</category>
      <category>programming</category>
      <category>optimization</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0807Jul.html?080704#080704</guid>
      <pubDate>Fri, 04 Jul 2008 07:20:00 +0100</pubDate>
    </item> <!-- 07:20 -->

   <item>
      <title>Internet "computing" vs. "publishing", the book</title>
      <link>http://WWW.sabi.co.UK/blog/0807Jul.html?080702#080702</link>

      <description>The distiction between the "publishing", "content"
	culture, and the "transaction", "technology" one is vividly
        argued in "Burn rate" by Michael Wolff.</description>

      <category>business</category>
      <category>networking</category>
      <category>dotcom</category>
      <category>web</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0807Jul.html?080702#080702</guid>
      <pubDate>Wed, 02 Jul 2008 21:00:00 +0100</pubDate>
    </item> <!-- 21:00 -->

    <!-- 22:35 -->
    <item>
      <title>Simple script for flexible SSH tunneling</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080630#080630</link>

      <description>With a simple wrapper script for OpenSSH commands
	it is possible to tunnel over SSH via a bastion host fairly
        conveniently.</description>

      <category>system administration</category>
      <category>network administration</category>
      <category>SSH</category>
      <category>shell</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080630#080630</guid>
      <pubDate>Mon, 30 Jun 2008 22:35:00 +0100</pubDate>
    </item>

    <item>
      <title>Toshiba U300-13U review update</title>
      <link>http://WWW.sabi.co.UK/Notes/hwNotesReviews.html#toshibaU300u1?080628</link>
      <description>Some small updates to my review of my Toshiba U300
        laptop.</description>
      <category>hardware</category>
      <category>laptop</category>
      <category>review</category>

      <pubDate>Sat, 28 Jun 2008 22:50:00 +0100</pubDate>
    </item>

    <item>
      <title>Updated Linux network traffic shaping script</title>
      <link>http://WWW.sabi.co.UK/#sourcesSabishape?080627</link>

      <description>The overall logic has not changed much, but
	the options are slightly different, in particular there is
	now no default command, and slightly different defaults
	for bandwidth.</description>

      <category>Linux</category>
      <category>networking</category>
      <category>Netfilter</category>
      <category>iptables</category>
      <category>shell</category>
      <category>shaping</category>
      <category>IP</category>
      <category>configuration</category>
      <category>traffic control</category>

      <pubDate>Fri, 27 Jun 2008 21:00:00 +0000</pubDate>
    </item>

    <item>
      <title>Updated Linux IPv4/IPv6 leaf firewall script</title>
      <link>http://WWW.sabi.co.UK/#sourcesSabifire?080627</link>

      <description>Much improved version of this leaf firewall
	script with some small improprieties corrected, and rather
	better internal organisation. The command interface has
	changed slightly, and now the default on stopping the
	firewall is to accept all packets. To reject all packets
	use 'close' instead of 'stop'. There have been quite a few
	unreleased versions between the previous release and
	this one.</description>

      <category>Linux</category>
      <category>networking</category>
      <category>Netfilter</category>
      <category>iptables</category>
      <category>shell</category>
      <category>shaping</category>
      <category>IP</category>
      <category>configuration</category>
      <category>firewall</category>

      <pubDate>Fri, 27 Jun 2008 21:00:00 +0000</pubDate>
    </item>

    <item>
      <title>New Linux power saving script</title>
      <link>http://WWW.sabi.co.UK/#sourcesSabiwatts?080627</link>

      <description>This new script is designed to tweak the power
	related settings of the Linux kernel for normal operation
	when on permanent power and for low power operation when
	on battery power. This is a polished and documented version
	of a script that I have been using for a while. It is not
	yet configurable, but the defaults embedded in it should be
	reasonable for most users.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>laptop</category>
      <category>shell</category>
      <category>configuration</category>
      <category>power</category>

      <pubDate>Fri, 27 Jun 2008 21:00:00 +0000</pubDate>
    </item>

    <item>
      <title>Distribution of services to servers, some examples</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080626#080626</link>

      <description>Some examples to illustrate different ways to
        split redundancy and services and users given a fixed
        number of servers.</description>

      <category>system administration</category>
      <category>network administration</category>
      <category>performance</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080626#080626</guid>
      <pubDate>Thu, 26 Jun 2008 07:25:00 +0100</pubDate>
    </item>

    <item>
      <title>Toshiba does Cell based laptop graphics</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080625#080625</link>

      <description>Notably, Toshiba have launched a laptop with a Cell
        based graphics chip. Which supports the idea that the Sony
        PS3 was originally intended to have two Cell chips, one for
        graphics.</description>

      <category>performance</category>
      <category>graphics</category>
      <category>Sony</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080625#080625</guid>
      <pubDate>Wed, 25 Jun 2008 07:15:00 +0100</pubDate>
    </item> <!-- 07:15 -->

    <item>
      <title>DBMSes, automatic and simple tuning</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080621#080621</link>

      <description>Indices are very important to performance in DBMses,
        yet often DBAs forget about creating them or compacting them.
        The conclusion to me is that performance tuning should be as
        automatic as possible, as while not being as good as good
        manual tuning, it is more likely to happen.</description>

      <category>DBMS</category>
      <category>performance</category>
      <category>system administration</category>
      <category>software engineering</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080621#080621</guid>
      <pubDate>Sat, 21 Jun 2008 18:05:00 +0100</pubDate>
    </item> <!-- 18:05 -->

    <item>
      <title>The irony of Python and Lisp</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080620#080620</link>

      <description>Python is mostly an alternate syntax for Lisp,
        which had a reputation for being slow motivating Lisp
        implementors to create very efficient implementations.
        Ironically Python implementation are much slower and
        much more popular.</description>

      <category>programming</category>
      <category>Python</category>
      <category>Lisp</category>
      <category>performance</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080620#080620</guid>
      <pubDate>Fri, 20 Jun 2008 17:30:00 +0100</pubDate>
    </item> <!-- 17:30 -->

    <item>
      <title>Europe says to upgrade to IPv6</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080618#080618</link>

      <description>IPv6 is being pushed by Eropean authorities,
      and that seems a good idea, as Europe is not as vested in
      the existing IPv4 based Internet as much as the USA.</description>

      <category>networking</category>
      <category>IPv6</category>
      <category>business</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080618#080618</guid>
      <pubDate>Wed, 18 Jun 2008 06:45:00 +0100</pubDate>
    </item> <!-- 06:45 -->

    <item>
      <title>High quality realtime sw ray tracing</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080615#080615</link>

      <description>Intel demonstrates an IA32 SMP systems
      capable of playing a contemporary game with realtime sw
      raytracing.</description>

      <category>performance</category>
      <category>architecture</category>
      <category>CPU</category>
      <category>games</category>
      <category>Intel</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080615#080615</guid>
      <pubDate>Sun, 15 Jun 2008 11:15:00 +0100</pubDate>
    </item> <!-- 11:15 -->

   <item>
      <title>Google and Amazon "cloud" suitability</title>
      <link>http://WWW.sabi.co.UK/blog/0806jun.html?080616#080616</link>

      <description>Google's platform is mostly about large scale
      read-only best effort publishing, Amazon's mostly about global
      state transactional e-commerce, and cloud computing could be more
      about the latter than the former. But it is more likley perhaps
      to be about small local transactions and small publishing, so
      neither may have the advantage.</description>

      <category>business</category>
      <category>search</category>
      <category>e-commerce</category>
      <category>DBMS</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0806jun.html?080616#080616</guid>
      <pubDate>Mon, 16 Jun 2008 21:15:00 +0100</pubDate>
    </item> <!-- 21:15 -->

    <item>
      <title>Google and Amazon "cloud" pricing and consequences</title>
      <link>http://WWW.sabi.co.UK/blog/0805may.html?080531#080531</link>

      <description>The availability of cloud computing retail from both
      Google and Amazon means that low initial outlay e-commerce web
      sites will be even easier to setup. This will favour sites that
      sell third world labor intensive services to first world
      customers, as it makes available first world class (and priced)
      infrastructure that would be difficult otherwise for third world
      companies to arrange.</description>

      <category>business</category>
      <category>clusters</category>
      <category>e-commerce</category>
      <category>DBMS</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0805may.html?080531#080531</guid>
      <pubDate>Sat, 31 May 2008 22:00:00 +0100</pubDate>
    </item>

    <item>
      <title>Large storage pools are rarely necessary</title>
      <link>http://WWW.sabi.co.UK/blog/0805may.html?080516#080516</link>

      <description>Except in some really scalable applications
      large storage pools are rarely necessary, and collections of
      localized storage pools seem to be easier and cheaper to setup
      and maintain.</description>

      <category>storage</category>
      <category>networking</category>
      <category>file systems</category>
      <category>system administration</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0805may.html?080516#080516</guid>
      <pubDate>Fri, 16 May 2008 19:05:00 +0100</pubDate>
    </item> <!-- 19:05 -->

   <item>
      <title>Parallel assignment and functional programming</title>
      <link>http://WWW.sabi.co.UK/blog/0805may.html?080512#080512</link>

      <description>Parallel assignment probably is what makes
	parallel programming interesting, as ordering of side
        effects is the hard issue, not side effects in general.</description>

      <category>programming</category>
      <category>software engineering</category>
      <category>computer languages</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0805may.html?080512#080512</guid>
      <pubDate>Mon, 12 May 2008 11:21:00 +0100</pubDate>
    </item>

   <item>
      <title>The case for parallel assignment</title>
      <link>http://WWW.sabi.co.UK/blog/0805may.html?080507#080507</link>

      <description>Parallel assingment can be compiled efficiently
	and easily and simplifies the basic blocks of programs.</description>

      <category>programming</category>
      <category>software engineering</category>
      <category>performance</category>
      <category>programming languages</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0805may.html?080507#080507</guid>
      <pubDate>Wed, 07 May 2008 07:06:00 +0100</pubDate>
    </item>

   <item>
      <title>The tragic depth of the Microsoft cultural hegemony</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080430#080430</link>

      <description>An upgraded version of the Linux logging daemon
	uses MS-DOS shell style syntax.</description>

      <category>software engineering</category>
      <category>Linux</category>
      <category>UNIX</category>
      <category>Microsoft</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080430#080430</guid>
      <pubDate>Wed, 30 Apr 2008 17:20:00 +0100</pubDate>
    </item>

    <item>
      <title>Much improved file system checking for XFS</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080406#080406</link>

      <description>The checking utilities for XFS have been improved and
	now take a lot less memory; some approximate rules are given as
        to how much.</description>

      <category>system administration</category>
      <category>performance</category>
      <category>storage</category>
      <category>XFS</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080406#080406</guid>
      <pubDate>Sun, 06 Apr 2008 18:10:00 +0100</pubDate>
    </item>

    <item>
      <title>Large decreases in flash disk prices expected</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080428#080428</link>

      <description>The capacity/size/price ratio of flash disks is
	expected to increase to 50%/year for at least a few years,
	which is higher than the current rate of increase for hard
	disks so they might reach parity in a few years. Good news,
	even if currently flash disks are bit too slow and way more
        expensive than that.</description>

      <category>hardware</category>
      <category>storage</category>
      <category>performance</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080428#080428</guid>
      <pubDate>Mon, 28 Apr 2008 07:50:00 +0100</pubDate>
    </item>

    <item>
      <title>Multiple CPUs, developers, power saving</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080420#080420</link>

      <description>Perhaps apart from servers the best applications for
      multiple CPU chips are development and power saving</description>

      <category>performance</category>
      <category>software engineering</category>
      <category>CPU</category>
      <category>architecture</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080420#080420</guid>
      <pubDate>Sun, 20 Apr 2008 21:20:00 +0100</pubDate>
    </item>

    <item>
      <title>Large storage pools and Lustre</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080417#080417b</link>

      <description>Building large storage pools is not easy, for
	single pools up to some hundred TB I think Lustre seems
        currently quite attractive.</description>

      <category>file systems</category>
      <category>storage</category>
      <category>Linux</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080417#080417b</guid>
      <pubDate>Thu, 17 Apr 2008 21:30:00 +0100</pubDate>
    </item>

    <item>
      <title>Number of patches in Red Hat kernels over time</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080417#080417</link>

      <description>The large number of patches in RHEL kernels
	is fruit both of policy and "scratch my itch" practice.</description>

      <category>RedHat</category>
      <category>software engineering</category>
      <category>Linux</category>
      <category>operating systems</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080417#080417</guid>
      <pubDate>Thu, 17 Apr 2008 21:15:00 +0100</pubDate>
    </item>

    <item>
      <title>Two rather different types of clusters</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080416#080416</link>

      <description>Parallelism in clusters can be used to achieve
	greater speed or reliability or both. Initially it was
        about reliability, currently it is mostly about speed.</description>

      <category>operating system</category>
      <category>performance</category>
      <category>storage</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080416#080416</guid>
      <pubDate>Wed, 16 Apr 2008 18:10:00 +0100</pubDate>
    </item>

    <item>
      <title>Dimensions of filesystem performance</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080415#080415</link>

      <description>File system performance envelopes have many
	dimensions, and several are important, because the performance
	envelopes of common file system designs are rather differently
	shaped.</description>

      <category>file systems</category>
      <category>performance</category>
      <category>software engineering</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080415#080415</guid>
      <pubDate>Tue, 15 Apr 2008 19:35:00 +0100</pubDate>
    </item>

    <item>
      <title>Performance tests of 2-CPU and 4-CPU chips</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080414#080414</link>

      <description>Looking at some recent 2-CPU and 4-CPU chip tests
        the 2-CPU ones seem still the most convenient, even if some
        important applications can take advantage of more CPUs.</description>

      <category>hardware</category>
      <category>performance</category>
      <category>parallelism</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080414#080414</guid>
      <pubDate>Mon, 14 Apr 2008 21:05:00 +0100</pubDate>
    </item>

    <item>
      <title>A cheap large reliable storage pool system</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080407#080407</link>

      <description>How to design a cheap reliable storage pool
      system It is not simple, but a compromise can be found in
      Lustre and network redundancy.</description>

      <category>storage</category>
      <category>networking</category>
      <category>system administration</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080407#080407</guid>
      <pubDate>Mon, 07 Apr 2008 21:55:00 +0100</pubDate>
    </item>

    <item>
      <title>Some other recent CPU developments</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080405#080405c</link>

      <description>Rather interesting new about recent CPUs,
        for example CPU ICs with 410m transistors, and simple
        and small x86 CPUs</description>

      <category>architecture</category>
      <category>hardware</category>
      <category>CPU</category>
      <category>Intel</category>
      <category>AMD</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080405#080405c</guid>
      <pubDate>Sat, 05 Apr 2008 21:20:00 +0000</pubDate>
    </item>

    <item>
      <title>Recent AMD and Intel quad-CPU chip test</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080405#080405b</link>

      <description>Comparison of recetn AMD and Intel CPU chips,
        including 45nm low power ones.</description>

      <category>architecture</category>
      <category>hardware</category>
      <category>CPU</category>
      <category>Intel</category>
      <category>AMD</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080405#080405b</guid>
      <pubDate>Sat, 05 Apr 2008 19:15:00 +0000</pubDate>
    </item>

    <item>
      <title>Another example of inappropriate error messages</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080405#080405</link>

      <description>Bad error reporting is common, and I found a
        particularly galling and classic example.</description>

      <category>software engineering</category>
      <category>programming</category>
      <category>Linux</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080405#080405</guid>
      <pubDate>Sat, 05 Apr 2008 15:40:00 +0000</pubDate>
    </item>

    <item>
      <title>More data on the cost of the Linux page cache</title>
      <link>http://WWW.sabi.co.UK/blog/0804apr.html?080402#080402</link>

      <description>The Linux page cache shows again how incredibly
        expensive it is in terms of CPU time.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>IO</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0804apr.html?080402#080402</guid>
      <pubDate>Wed, 02 Apr 2008 21:05:00 +0000</pubDate>
    </item>

    <item>
      <title>Outsourcing computer service</title>
      <link>http://WWW.sabi.co.UK/blog/0803mar.html?080315#080315</link>

      <description>Outsourcing computing services is not quite like
        buying computing power supply from a utility; it is like
        using the services of a bank, because it usually involves
        trusting business critical data to the supplier of those
        services.</description>

      <category>system administration</category>
      <category>project management</category>
      <category>business</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0803mar.html?080315#080315</guid>
      <pubDate>Sat, 15 Mar 2008 20:50:00 +0000</pubDate>
    </item>

    <item>
      <title>Linux block device performance issues</title>
      <link>http://WWW.sabi.co.UK/blog/0803mar.html?080303#080303</link>

      <description>The Linux block cache layer has some performance
	issues probably due to misdesigns and lack of care.</description>

      <category>Linux</category>
      <category>system administration</category>
      <category>performance</category>
      <category>storage</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0803mar.html?080303#080303</guid>
      <pubDate>Mon, 03 Mar 2008 21:20:00 +0000</pubDate>
    </item>

    <item>
      <title>Free software and the value of proprietary platforms</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080226#080226</link>

      <description>Runtime compatibility with competing platforms makes
	users prefer the product of competitors, and development
	compatibility with competiting platforms makes developers prefer
	the competiting platform.</description>

      <category>software engineering</category>
      <category>development</category>
      <category>Linux</category>
      <category>MS Windows</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080226#080226</guid>
      <pubDate>Tue, 26 Feb 2008 21:20:00 +0000</pubDate>
    </item>

    <item>
      <title>A 48 drive setup, it had to happen</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080220#080220</link>

      <description>Someone describes a setup for a 48 drive storage
	system that creates a single storage pool</description>

      <category>system administration</category>
      <category>RAID</category>
      <category>Linux</category>
      <category>storage</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080220#080220</guid>
      <pubDate>Wed, 20 Feb 2008 08:20:00 +0000</pubDate>
    </item>

    <item>
      <title>Switching my main work PC to a laptop</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080219#080219</link>

      <description>I have bought a new laptop that is more powerful
	than my desktop, and has enough storage space that I can use
	it as my main system. Thanks to recent high capacity 2.5"
	disk drives.</description>

      <category>storage</category>
      <category>hardware</category>
      <category>laptop</category>
      <category>performance</category>
      <category>Linux</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080219#080219</guid>
      <pubDate>Tue, 19 Feb 2008 22:00:00 +0000</pubDate>
    </item>

    <item>
      <title>Another RAID and volume management perversity</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080217#080217</link>

      <description>Someone wants to start with a 2+1 RAID5
	and thn expand it to a 13+1 RAID5 or a 12+2 RAID6
	by adding drives one at a time. That does not seem wise,
	and better alternatives can be described.</description>

      <category>RAID</category>
      <category>Linux</category>
      <category>system administration</category>
      <category>storage</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080217#080217</guid>
      <pubDate>Sun, 17 Feb 2008 11:30:00 +0000</pubDate>
    </item>

    <item>
      <title>A RAID and filesystem perversity</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080216#080216</link>

      <description>A large filesystem with a 100 million small files
	does not look like being a good idea, especially compared to a
	simple indexed file. Some numbers from a smaller case
	illustrate why.</description>

      <category>storage</category>
      <category>file systems</category>
      <category>DBMS</category>
      <category>Linux</category>
      <category>XFS</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080216#080216</guid>
      <pubDate>Sat, 16 Feb 2008 12:30:00 +0000</pubDate>
    </item>

    <item>
      <title>Toshiba U300-13U review</title>
      <link>http://WWW.sabi.co.UK/Notes/hwNotesReviews.html#toshibaU300?080219</link>
      <description>Just got a Toshiba U300-13U and overall it is quite
	good, well suited to GNU/Linux, with a few annoying physical
	issues.</description>
      <category>hardware</category>
      <category>laptop</category>
      <category>review</category>

      <pubDate>Tue, 19 Feb 2008 08:45:00 +0000</pubDate>
    </item>

    <item>
      <title>Some more data on filesystem checking speed</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080210#080210</link>

      <description>Some more reports of filesystem checking speed
	or slowness.</description>

      <category>performance</category>
      <category>file systems</category>
      <category>system administration</category>
      <category>Linux</category>
      <category>ext3</category>
      <category>XFS</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080210#080210</guid>
      <pubDate>Sun, 10 Feb 2008 21:40:00 +0000</pubDate>
    </item>

    <item>
      <title>Coupling and active redundancy</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080202#080202</link>

      <description>Active redundancy implies some degree of coupling,
	and this by itself introduces common modes of failure.</description>

      <category>system administration</category>
      <category>network administration</category>
      <category>architecture</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080202#080202</guid>
      <pubDate>Sat, 02 Feb 2008 19:35:00 +0000</pubDate>
    </item>

    <item>
      <title>Much better read latency with AMD's HyperTransport</title>
      <link>http://WWW.sabi.co.UK/blog/0712dec.html?071210#071210</link>

      <description>AMD CPUs have much better read latency than Intel
	ones thanks to HyperTransport. But memory sticks are designed to
	have high latency as Intel CPUs have lots of onchip cache
	anyhow.</description>

      <category>architecture</category>
      <category>hardware</category>
      <category>AMD</category>
      <category>Intel</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0712dec.html?071210#071210</guid>
      <pubDate>Mon, 10 Dec 2007 21:30:00 +0000</pubDate>
    </item>

    <item>
      <title>Slow remote X display with Exceed and antialiases fonts</title>
      <link>http://WWW.sabi.co.UK/blog/0706jun.html?070609#070609</link>

      <description>Peculiar case where a remote X display using
	anti-aliased fonts is abnormally slow</description>

      <category>system administration</category>
      <category>X11</category>
      <category>fonts</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0706jun.html?070609#070609</guid>
      <pubDate>Sat, 09 Jun 2007 22:00:00 +0000</pubDate>
    </item>

    <item>
      <title>Software versus hardware RAID: bandwidth, latency, convenience</title>
      <link>http://WWW.sabi.co.UK/blog/0706jun.html?070602#070602</link>

      <description>Software RAID has a very different profile from
	hardware RAID, both as to performance and administration</description>

      <category>RAID</category>
      <category>system administration</category>
      <category>storage</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0706jun.html?070602#070602</guid>
      <pubDate>Sat, 02 Jun 2007 22:00:00 +0000</pubDate>
    </item>

    <item>
      <title>Slow transfer rate over SSH and improvements</title>
      <link>http://WWW.sabi.co.UK/blog/0802feb.html?080205#080205</link>

      <description>Why WinSCP by default is slow, and general
	performance issues of bulk data transfer over SSH and
        solutions or palliatives</description>

      <category>performance</category>
      <category>networking</category>
      <category>Linux</category>
      <category>MS Windows</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0802feb.html?080205#080205</guid>
      <pubDate>Tue, 05 Feb 2008 21:40:00 +0000</pubDate>
    </item>

    <item>
      <title>Number of NFS server instances</title>
      <link>http://WWW.sabi.co.UK/blog/0801jan.html?080122#080122</link>

      <description>The number of NFS instances can be a surprising
	bottleneck, and in the past it exposed a disjointed hardware
	vs. software design on Sun servers.</description>

      <category>Linux</category>
      <category>performance</category>
      <category>networking</category>
      <category>system administration</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0801jan.html?080122#080122</guid>
      <pubDate>Tue, 22 Jan 2008 21:05:00 +0000</pubDate>
    </item>

    <item>
      <title>BCM5752 jumbo frames work only in output</title>
      <link>http://WWW.sabi.co.UK/blog/0712dec.html?071223#071223</link>

      <description>Discovered that a popular 1gb/s chipset can
	transmit but not receive jumbo frames, and how to take
	advantage of jumbo frames at least for transmitting.</description>

      <category>hardware</category>
      <category>networking</category>
      <category>performance</category>
      <category>Linux</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0712dec.html?071223#071223</guid>
      <pubDate>Sun, 23 Dec 2007 14:50:00 +0000</pubDate>
    </item>

    <item>
      <title>Network accelerators run Linux</title>
      <link>http://WWW.sabi.co.UK/blog/0712dec.html?071215#071215b</link>

      <description>Two network accelerators recently announced,
	both offloading network processing to onboard Linux.</description>

      <category>hardware</category>
      <category>networking</category>
      <category>Linux</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0712dec.html?071215#071215b</guid>
      <pubDate>Sat, 15 Dec 2007 15:10:00 +0000</pubDate>
    </item>

    <item>
      <title>2.5" hard disks differences</title>
      <link>http://WWW.sabi.co.UK/blog/0712dec.html?071215#071215</link>

      <description>A group test of 2.5" hard discs reported significant
	performance differences among supposedly very similar commodity
	products.</description>

      <category>hardware</category>
      <category>storage</category>
      <category>performance</category>

      <guid isPermaLink="true"
        >http://www.sabi.co.uk/blog/0712dec.html?071215#071215</guid>
      <pubDate>Sat, 15 Dec 2007 16:25:00 +0000</pubDate>
    </item>
  </channel>

  <script xmlns="http://www.w3.org/1999/xhtml" type="text/javascript"
    src="style/rss.js"></script>
<!--
  ;;; Local Variables:			***
  ;;;   tab-width:		8	***
  ;;;   indent-tabs-mode:	t	***
  ;;;   fill-column:		72	***
  ;;; End:				***
  vim:set noet ts=8 tw=72 sw=2:
-->
</rss>
