Memory upgrades for Lenovo S10e, Aspire M5201, eMachines T6212

Last month my Gateway GT4014j suddenly died and I replaced it with an Acer Aspire M5201, which came with 2 GB of memory installed. That was less than the 3 GB I had in the Gateway and it affected some very memory hungry apps I use. So three weeks later I did a memory upgrade – a series of upgrades to be more precise. I bought two sets of memory upgrades and upgraded three machines.

In my experience, most computers don’t become obsolete because of CPU clock speeds that are too slow but because either there is too little main memory (RAM) for newer, more demanding (i.e. more bloated) applications or because the machine runs out of disk space. The latter used to be more of a problem than it is now, because of the availability of external USB drives, but lack of RAM is still a problem.

Giving your computer enough memory usually is the best way to give it a performance boost because insufficient memory will result in disk swapping. Without enough memory your applications will spend more time waiting for data to get swapped to and from disk, which slows even the fastest CPU down to a crawl such that even a medium or low end CPU with sufficient RAM could run circles around it.

When Vista came out, many manufacturers still sold 512 MB configurations with Vista Home Basic, while advertising mainly the processor speed. Buyers would have been far better off picking a somewhat slower CPU but equipped with a whole GB of RAM. The same is basically still true today for 1 GB machines (usually with Vista Home Premium) vs. 2 GB machines: More main memory almost always beats more Gigahertz!

Back to my Aspire M5201: It has four DIMM sockets (2 banks) for up to 8 GB of memory (2GBx4), but unless you run a 64-bit operating system such as a 64-bit Linux, Vista 64-bit or (from 22 Oct 2009) Windows 7 64-bit, you can’t actually make use of more than 4 GB. Therefore I opted to go as far as 4 GB only for two 1 GB DDR2 PC2-6400 DIMMs from Crucial Technology (CT12864AA800, from Tsukumo.co.jp). They were 1,980 yen each (about US$22) including sales tax.

I also ordered one 2 GB DDR2 PC2-5300 SODIMM (CT25664AC667) for 3,800 yen (about US$42) for a Lenovo IdeaPad S10e. The S10e is a typical Atom N270 netbook with 1 GB of RAM. One 512 MB SODIMM base memory is soldered in while the other can be replaced with either a 1 GB or 2 GB SODIMM. You end up with 1.5 GB or with 2 GB (not 2.5 GB because the 945 GSE chip set is not capable of addressing more than 2 GB of total memory). The Intel Atom itself supports up to 4 GB, which is why nettops with the NVIDIA Ion chip set such as the Acer Revo 3600 can handle up to 4 GB of RAM. The memory upgrade went very smoothly, especially the netbook. The Acer Aspire M5201 (AMD 780G chip set) now shows 3.75 GB of RAM in Windows while the Lenovo shows 1.99 GB.

Encouraged by that I removed the 1.5 GB of DDR DIMMs from my 2005-vintage eMachines T6212 and transplanted the 3 GB of DDR DIMMs (1 GB x 2 and 512 MB x 2) from the dead Gateway GT4014j to it. DDR memory was superseded by DDR2 3 years ago, but now that old machine that originally came with 512 MB shows 2.75 GB of memory in Windows.

If you’re not sure what memory is right for your machine or how much memory you can fit, Crucial have a convenient Memory Finder application that will figure it out for you.

IE7, IE8 and .exe files on network drive: “The publisher could not be verified…”

I recently upgraded a Windows XP machine from Internet Explorer 6 to IE 8. Since then I’ve been getting warnings whenever I wanted to run any .EXE file that doesn’t use code signing off a network drive.

My main browser on that machine is FireFox, but that doesn’t matter. The warning that comes up is the following:

The publisher could not be verified. Are you sure you want to run this software?

Apparently the same behaviour was already introduced when upgrading to IE 7. It may also happen with Service Pack 3 (SP3) for XP. Unlike for the similar query that happens for executable files downloaded off the Internet, there is no option to turn off the warning for a specific .EXE file.

There is however a way to allow all unsigned EXEs to run off a given network drive or network server without the warning. Here is what you do:

Open the Control Panel, select Internet Options, click the Security tab, select Local intranet, click Sites and the Advanced button. Add any server (e.g. \\myserver) by entering the name and clicking Add. The resources will be added as file://myserver. When you’re done click Close and OK.

This will allow all unsigned applications on the listed servers to execute without security prompt just as if they were on your local machine.

IPv6 with DD-WRT router and Hurricane Electric

Last weekend I got IPv6 working on my US$60 router, allowing all my machines here to talk IPv6 to the outside world. That includes an Ubuntu Linux server, 4 PCs and one Mac.

The biggest incentive for upgrading to IPv6 is the fact that at the current pace we’ll run out of (IPv4) IP addresses in about two years. These are the unique host addresses (usually written in dotted decimal format like 209.85.171.100) that identify client and server computers on the Internet. The newer IPv6 standard that replaces 32-bit addresses with 128-bit addresses will forever take care of this shortage. It will also do away with the primary need for Network Address Translation (NAT) which has been a big headache for voice over IP (VoIP) and other peer to peer applications.

However, over a decade after the introduction of the newer standard (in 1997), uptake is still slow. Many ISPs still don’t support IPv6 and neither does a lot of the equipment used at homes and offices. This is gradually starting to change. IPv6 is an integral part of modern operating systems such as Linux, Mac OS X, Windows 7, Windows 2008 Server, Windows Vista, Windows 2003 Server and Windows XP (where it’s optional).

If your ISP does not support IPv6, you can still use it by employing the services of a tunnel broker, which gives you IPv6 connectivity over an IPv4 tunnel. This lets you test your software with the new APIs, though you won’t gain native IPv6 performance. If you have a static IP you can use tunnelbroker.net by Hurricane Electric, Inc. Their service is professionally run and free. Another option is SixXS, but I have not tried them.

My router is a Buffalo WHR-HP-G54, which is compatible with the Linux based open source DD-WRT firmware. Recent versions of DD-WRT have IPv6 support. My first attempt with the v24 sp1 std build which is supposed to include IPv6 was unsuccessful, but I had more luck after trying the v24 10070 crushedhat version (dd-wrt.v24-10070_crushedhat_4MB.bin). Here’s what you do:

  • The following instructions assume that your WHR-HP-G54 router is running open source DD-WRT firmware. If your router is still running the default firmware, install DD-WRT v24 sp1 mini generic (SVN build 10020, 27-July-2008) on it. See my blog post on the WHR-HP-G54 with DD-WRT for detailed instructions. The WHR-HP-G54DD is a version of this router that comes with DD-WRT preinstalled.
  • Go to Security / Firewall on your DD-WRT and remove the check mark on Block anonymous WAN requests (ping) so that Hurricane Electric can verify your router exists by pinging it.
  • Go to www.tunnelbroker.net and sign up for an account. Then log in and go to Create a Regular tunnel. You’ll need to enter your static IP, which will be conveniently displayed. You have a choice of tunnel endpoints. Pick one that has a short ping time from where you are. Make a note of all the details of the tunnel that is created. You will need to enter some of these details on your router, in particular these:
    • Server IPv4 address
    • Server IPv6 address
    • Routed /64
  • Read crushedhat’s description of how to configure the router with his firmware, which should work with most Broadcom-based DD-WRT-compatible routers.
  • I’m assuming you have updated the firmware of your router before and know the usual caveats about “bricking” your router if anything goes wrong. I won’t be responsible for that. 😉 I went from the factory Buffalo firmware to v24 sp1 mini to v24 sp1 std to v24 sp1 mini to v24 crushedhat 10070, with no problems, but your mileage may vary. I downgraded from v24 sp1 std (4 MB) to v24 sp1 mini (2 MB) “just in case” before flashing crushedhat’s std (4 MB) build. I did not opt to reset the NVRAM to factory defaults.
  • Download a copy of the v24 crushedhat 10070 build and save it on your hard disk. Use a computer with a wired connection to the router, not WLAN for the firmware upgrade. Go to Administration / Firmware Upgrade and select the dd-wrt.v24-10070_crushedhat_4MB.bin file. Click the upgrade button. Don’t touch anything until after the router has reset and is running the new firmware.
  • Go to Administration / Management and check Enable for IPv6 and Radvd enabled. Then paste the following into the Radvd config box:

    interface br0
    {
    AdvSendAdvert on;
    prefix 2001:470:YYYY:YY::/64
    {
    AdvOnLink on;
    AdvAutonomous on;
    };
    };

    where 2001:470:YYYY:YY::/64 matches the value of “Routed /64” in the created tunnel given to you by Tunnelbroker.net:

    Server IPv4 address: 216.218.226.238
    Server IPv6 address: 2001:470:XXXX:XX::1/64
    Client IPv4 address: 219.110.159.121
    Client IPv6 address: 2001:470:YYYY:YY::2/64
    Routed /48: 2001:470:ZZZZ::/48
    Routed /64: 2001:470:YYYY:YY::/64

  • Go to Administration / Commands and enter these commands, then click Save Startup:

    ip tunnel add he-ipv6 mode sit remote 216.218.226.238 ttl 64
    ip link set he-ipv6 up
    ip addr add 2001:470:XXXX:XX::2/64 dev he-ipv6
    ip route add ::/0 dev he-ipv6
    ip addr add 2001:470:YYYY:YY:200:00ff:fe00:0000/64 dev br0

    Replace 216.218.226.238 with Server IPv4 address from your tunnel settings, 2001:470:XXXX:XX:: with the Server IPv6 address value and 2001:470:YYYY:YY:: with the Routed /64 value.

  • Go to Administration / Commands and enter these commands, then click Save Firewall:

    insmod ip6t_REJECT
    ip6tables -F
    ip6tables -A FORWARD -p tcp -i he-ipv6 –syn -m multiport –dports ftp-data,ftp,ssh,smtp,http,https,ntp,domain -j ACCEPT
    ip6tables -A FORWARD -p tcp -i he-ipv6 –syn -j REJECT –reject-with adm-prohibited
    ip6tables -A FORWARD -p udp -i he-ipv6 -m multiport –dports ntp,domain -j ACCEPT
    ip6tables -A FORWARD -p udp -i he-ipv6 -j REJECT –reject-with adm-prohibited

  • Now it’s time to check if everything works. It may take a few minutes or one reboot for your client to obtain an IPv6 address. Here is what things should look like after that:

    C:\>ipconfig

    Windows IP Configuration

    Ethernet adapter Motherboard Network Connection:

    Connection-specific DNS Suffix . :
    IP Address. . . . . . . . . . . . : 192.168.100.2
    Subnet Mask . . . . . . . . . . . : 255.255.255.0
    IP Address. . . . . . . . . . . . : 2001:470:YYYY:YY:290:feff:fe66:e237
    IP Address. . . . . . . . . . . . : fe80::290:feff:fe66:e237%6
    Default Gateway . . . . . . . . . : 192.168.100.1
    fe80::21d:73ff:fe3a:3b8c%6

    Tunnel adapter Teredo Tunneling Pseudo-Interface:

    Connection-specific DNS Suffix . :
    IP Address. . . . . . . . . . . . : fe80::ffff:ffff:fffd%5
    Default Gateway . . . . . . . . . :

    Tunnel adapter Automatic Tunneling Pseudo-Interface:

    Connection-specific DNS Suffix . :
    IP Address. . . . . . . . . . . . : fe80::5efe:192.168.42.2%2
    Default Gateway . . . . . . . . . :

    You can ping Google’s IPv6 servers:

    C:\>ping ipv6.google.com

    Pinging ipv6.l.google.com [2001:4860:c004::68] from 2001:470:YYYY:YY:290:feff:fe66:e237 with 32 bytes of data:

    Reply from 2001:4860:c004::68: time=307ms
    Reply from 2001:4860:c004::68: time=307ms
    Reply from 2001:4860:c004::68: time=331ms
    Reply from 2001:4860:c004::68: time=318ms

    Ping statistics for 2001:4860:c004::68:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
    Approximate round trip times in milli-seconds:
    Minimum = 307ms, Maximum = 331ms, Average = 315ms

    Fire up FireFox 3 or the browser of your choice and go to http://www.kame.net/ – if the image of the turtle is dancing then you have IPv6 working. Go to http://whatismyv6.com/ to see your IPv6 address.

Good luck! 🙂

Windows 7 versus Linux on netbooks

“Does Linux stand a chance now that Windows 7 will run on netbooks?”, Shane O’Neill asks in an article in ComputerWorld on 15 January 2009 that overall sounds fairly optimistic on Microsoft’s prospects. However it largely avoids one crucial subject that matters for Microsoft in the struggle over market share in the booming nettop market: Money.

In 1985 Jack Tramiel, head of Atari Corporation came to visit Digital Research Inc. (DRI) to license its GEM graphical desktop environment for the new Atari 520ST. It was going to be a low-cost machine based on the same Motoroloa 68K CPU as Apple’s much more expensive Macintosh, which itself was a low-cost derivative of the Apple Lisa (that was long before Microsoft Windows became a viable product). Tramiel had a reputation as a fierce negotiator, so his counterpart at DRI, then the main competitor of Microsoft and Apple, was only half joking when he said to Tramiel: “Jack, I know you’ll probably start off by offering us a dollar per copy.” – “No,” replied Tramiel dryly. “50 cents.”

Tramiel knew that by coming out with a fully-functional product at rock-bottom prices he could grow the PC market. In the segment he envisaged there simply was no margin for a $50 operating system license. What was true when an Atari machine cost around $1000 is even more true today with $250-$450 netbooks, and future netbooks will be even cheaper than that. Soon we will also see netbooks based on the same low-power, low-cost ARM processors that power virtually all mobile phones.

Commentators cited by Computerworld on Windows 7 don’t really talk about money:

Analyst Rob Enderle, president of technology research firm The Enderle Group, agrees that Microsoft doesn’t see Linux as much of a threat and that refocusing on the netbook market is more about “Microsoft addressing the problem of having to keep shipping Windows XP long after its expiration date.”

Enderle says that getting XP on netbooks was clearly a response to Linux gaining traction, but that Microsoft is not afraid of consumers or OEMs having a preference for Linux.

“The problem was that Linux could run on a netbook and Vista couldn’t, not any consumer or OEM love for Linux,” he adds.

But Microsoft’s real problem wasn’t just that Vista was too big to fit on a 4 GB flash drive and too slow and bulky to run on an Intel Atom with 512 MB of RAM. It was also too expensive. So Microsoft could save face by charging next to nothing for its 5 year old Windows XP, but it didn’t make any real money on it. So what’s going to happen when Windows Vista 1.1 aka Windows 7 hits the streets in volume maybe a year from now?

Does it really matter to Microsoft shareholders and employees if the 21 million or so netbooks expected to be sold this year (and the even bigger numbers in 2010) will be running some version of Windows or a version of Linux (which is free), if previously those buyers would have picked up a more powerful machine that netted Microsoft $40-$100 per license?

Whether Windows 7 will run with decent performance on low-cost machines is really only half the question. The other is, how much Asus, Acer and the other netbook OEMs will offer to pay Steve Ballmer of Microsoft. Is it going to be $1 or 50c per copy? That is no way to sustain a business with a market capitalization of $150 billion and almost 90,000 employees worldwide (Jan 2009 numbers), as Microsoft is realizing to its horror.

Linux Eee Box now available, MS giving away XP?

Two months ago I commented on the non-availability of either of the two Linux versions of the Asus Eee Box 202 that had been announced months ago along with its XP-equipped sibling, which has been shipping for two months already.

It appears you can now also get Linux versions of this tiny energy saving desktop machine, but the pricing is most curious. Originally Asus announced an 80 GB Linux version for $269, $30 less than the similarly equipped XP version. This would presumably pay for the Windows XP license fee that Microsoft collects from every Windows OEM. Linux, which is free software licensed under GPL, comes without any such fees.

However, the actual pricing is now quite different. For example, Amazon.com sells the 1 GB RAM / 80 GB hard disk Windows XP Home version (in black) for $298 while the 1 GB RAM / 160 GB hard disk Linux version (in black) is available for $321 (prices may vary by the time you read this). The picture is similar at Newegg.com, which offers the 1 GB RAM / 80 GB hard Windows XP Home version for $299 and the 1 GB RAM / 160 GB hard disk Linux version for $319. That’s $20-$23 more for a bigger hard disk, but a cheaper operating system. Retail prices of the 80 and 160 GB Seagate Momentus 5400 RPM SATA drives used in the two versions differ by only $10 (about $50 vs. $60, at Egghead.com).

Adjusting for the hard drive cost, the Windows version is now $10 cheaper instead of $30 more expensive than the Linux version, money which we can assume came out of Microsoft’s revised OEM pricing. Asus is making at least $10 more per machine that ships with Linux than with Windows, even if Microsoft was giving away XP for free to Asus, which is probably not far from the truth.

Let’s put it this way: At this stage, Microsoft must have been pretty desperate to keep Linux off as many high volume, low end machines as possible. It had made a blunder by heavily betting on the success of Vista, which requires fairly high end hardware, but the market decided otherwise. To its surprize it found that customers wanted to stick with less resource-hungry XP. Then the boom in low end machines triggered by the One Laptop per Child initiative and the Asus Eee PC took Microsoft completely by surprize.

You see, Microsoft makes money on Windows by licensing it to OEMs like HP and Dell. In a mature market where almost everyone already owns a PC these OEMs were supposed to be able to sell more machines (and bundled Windows licenses) only if a new Windows version made the old machines obsolete. The actual needs of users played almost no role in this.

Microsoft had bet on selling ever more powerful machines, thinking that if old machines were not powerful enough to run Vista, owners of XP machines would dump their hardware and buy new machines from Microsoft’s OEM customers. Instead both retail and corporate customers revolted, as Vista offered them few real benefits but a number of potential problems. Instead of triggering a rush to Vista, Microsoft motivated customers to look at other alternatives such as Mac OS X, Ubuntu and other Unix or Linux flavours or simply to sit it out and stick with XP as long as possible.

This of course is a mortal threat to Microsoft’s business. Once millions of customers get used to no longer owning the latest and greatest Windows software or even start surfing the net, doing email, watching YouTube and running OpenOffice on non-Windows machines, where is that going to leave Microsoft in 2010, when Windows 7 and perhaps Office 2010 come out?

Right now Microsoft’s business model is still generating large cash flows from selling operating systems and application software to OEMs and retail customers. This is the classic model the microcomputer software industry has used since the 1970s, even before “microcomputers” where renamed to personal computers (yes, I was around then). It’s not how the mainframe business operated before and it’s not how new Internet businesses work now.

Witness successful companies such as Google and eBay, which are making money by acting as go-betweens between buyers and sellers (ads in the case of Google, merchandise in the case of eBay). Canonical and other Linux companies get their revenue from support contracts while giving away the software. Many new business models rely on advertising or shared sales revenue. Google makes billions of dollars and has hundreds of millions of daily users without in most cases selling directly to the people who use its services. The world as Microsoft knew it is changing and it will have to change with it. That will be extremely painful for them.

Microsoft is well aware of the need to move to a new net-centric business model. For evidence look no further than its recent desperate attempt to acquire Yahoo’s user base by buying out the whole company, after having spent years and pouring billions of dollars into largely unsuccessful efforts to match Yahoo with its own MSN/Hotmail/OfficeLive offerings. The problem is, Yahoo is only up for sale because it has not managed to compete terribly well against Google either. If Microsoft and Yahoo merge, it will be a case of the blind leading the blind.

Without a working alternative, Microsoft is finding it very difficult to move away from its classic business model, which depends on its ability to get customers to buy new copies of Windows and Office every couple of years at fairly hefty prices, whether users are keen on these new versions or not.

Microsoft is going to find this ever more challenging as new low-cost mobile devices come along, especially in the sub-$500 bracket. These new gadgets will bridge the gap between mobile phones, consumer devices and PCs, offering adequate performance and convenience at low prices. Think about it: How many people are still going to pay $499.95 for a copy of Microsoft Office Professional 2007 when the mini notebook or nettop they will be running it on only cost $300-$400? That pricing model only worked while Microsoft had a de-facto monopoly on desktop operating systems and office suite applications and especially when hardware was still expensive.

Microsoft’s pricing is really a relic of the 1980s, when a 6 MHz PC AT with half a MB of RAM and a 20 MB hard drive cost as much as $6,700. Then it still made sense to spend an extra 10-30% of that amount on software to make the machine useful, but today’s numbers look very different.

The recent announcement by Canonical to port Ubuntu to ARM-based low-power devices further illustrate the danger for Microsoft from the market moving to a low-cost, low-power, net-centric model. Most mobile phones use ARM-based chipsets. So do many consumer broadband routers. It’s the de-facto standard in those markets. Both the Apple iPhone and the Google Android-based G1 run variants of Unix or Linux on ARM CPUs. So do many $50-$150 broadband routers. Meanwhile Intel is working on integrating its Atom CPU core with the graphics chip in its upcoming “Pineview” chips, going head to head with ARM for low-cost “PC on a chip” designs. Competition in this segment will result in a plethora of interesting products available to consumers.

While there will always be a demand for high end PCs for gaming (where Microsoft competes with Sony’s PlayStation and other dedicated games machines), the low end mobile market will be the real growth market, especially in emerging markets such as Asia, Africa and Latin America.

What Microsoft is effectively doing in the netbook / nettop market today is giving away Windows XP just to keep a foot in the door for a remote chance of profits a few years from now. That’s only about as smart as selling mortgages to people who can’t really afford to buy a house. Sure, you get customers that way to keep the market growing, but what is going to happen a year or two later? As the sub-prime mortgage crisis has taught us, you can’t fudge revenue numbers forever. In the short term, giving away Windows may prop up market share numbers, but the result is no more than window dressing. The Vista debacle seriously damaged Microsoft’s value as a brand amongst consumers. Don’t count on their brand loyalty too much. Microsoft has lost its ability to steer the market that it used to control, just as big old IBM did before it.

Microsoft’s Windows and Office sales are the equivalent of a subprime crisis waiting to happen. Microsoft’s revenue model is being hollowed out. The software giant is now too big to quickly move to a model that will still fit the market a few years down the road.

When the big crunch comes, expect things to get nasty. Microsoft employs 4 1/2 times more people than Google which already makes more money than Microsoft does. This ratio is not going to stay that way.

UPDATE (2008-11-20):

I mentioned that the XP version went from initially being $30 more expensive to $10 cheaper than the Linux version and credited this to Microsoft deciding to essentially give XP away in order to hang on to market share.

A review of the Dell Inspiron mini 9 sub-notebook mentions that Dell charges $40 extra for XP compared to the Linux version. That’s exactly the margin by which the price of the Eee Box XP version dropped relative to its Linux sibling between the announcement and the start of shipments, so it would back up my reasoning.