[...] two hundred thousand years - far longer than any radioactivity will last
Rubidium 87 has a half-life of 47 billion (10^9) years (our soloar system is not yet 5 billion years old). Uranium 238 has a half-life of 4.5 Billion (10^9) years, Plutonium 239 has a half-life of 25.000 years. Half-life means that after some billion years, you still have half of your nuclear waste happily emitting radioactivity, while the other half has decayed to other, possibly also radioactive elements. After 7 times the half-life (7*47*10^9 years = 329*10^9 years), you still have round about 1 % of the original radioactive waste (2^-7 = 1/128 ~ 1%) and a lot of other radioactive products.
There is no final solution of nuclear waste, and there probably will never be one. It is practically impossible to guarantee a safe place for at least 7 half-lifes of Rubidium 87 or Uranium 238. Even if you can dig a deep hole and fill it with nulear waste, there still is a possibility much larger than zero that someone digs it out again in some thousand years and does not know what that shiny stuff is. There have already been accidents in third world countries, where poor and uneducated people digged out radioactive materials from medical devices on waste dumps. And a mountain used to store nuclear waste may erode before the waste becomes harmless.
Lessons learned: PR people don't know anything about mathematics and radioactivity.
The seventies and eighties saw a home hifi market flooded with crap gear from japan (Manufacturers like Sansui and Sony and Kenwood and Pioneer) that boasted incredibly low THD... and provided its owners incredibly bad sound.
I could not agree more. But this "Hi Fi" junk did not only come from Japan. In Germany, the (now broken) company "Schneider" produced large ammounts of cheap stereo towers (with tuner, double tape, amplifier, and record player, later also CD player), featuring large cases containing much air, a small transformer and a type plate promising 1000W music output from 50W AC input. No, they did not invent a perpetuum mobile. But their unrealistic output ratings coined the term "Schneider-Watt" for ridiculous amplifier output ratings, especially above the rated input power. I won't talk much about the sound of this devices or about the speaker boxes made of cheap plastic and cardboard. You should be able to guess what you get from these pieces of crap.
Personally, I prefer old high quality transistor amplifiers from the time when Japan imports were just a footnote in the german HiFi business.
Why the heck is he using an FM transmitter to connect the iPod to his nice tube equipment.
Perhaps because he uses old tube equipment without line inputs. Old tube radios are often driven by rectified mains voltage (so you get some hundred volts inside the radio on nearly all components), without an insulating transformer (so this voltage is "available" against earth and can kill you). Adding a line input to such a (simple and cheap) design requires an insulating transformer either for the power supply or for the line input, which would have caused additional costs.
It is possible to retrofit a line input to most old tube radios, but not without dramatic changes to the device. You need at least an additional switch and a hole in the backside. Most people who love old tube radios would rather like several root canal treatments without anesthesia than that.
Some "newer" and expensive old tube radios have inputs for a record player and/or a tape, both could be used to connect modern audio devices like the iPod, but not necessarily without mechanical and electrical adapters.
So the most easiest way to "connect" an iPod to old tube radios is an FM transmitter. As a nice side effect, you can "connect" several radios to the same iPod, all without fiddling with cables.
And by the way, frequencies below 50 Hz and above 15 kHz can only be heard by very young people. The older you get, the narrower the bandwith of your ears becomes.
He also inspired the Scotty Rule:
When your commanding officer tells you to get something done,
Estimate the time to finish the job,
Double that,
Change up to the next highest units.
For example, if you think the job takes two weeks, say it's impossible to get it done in under four months. So when you spend three weeks drinking beer then knock it out in an afternoon, your boss will think you're some kind of miracle worker.
This rule works quite well, I use it whenever possible. From my own experiences, most people estimate very bad. The good ones can get to the point where they estimate 50% of the actual time needed (the other ones estimate much less time). So doubling that time gives you a very good estimation for your own. Never tell anyone that number, don't write it down! Rounding up to the next highest units gives your boss a nice round number and you get a little time for all those catastrophes you did not think about when estimating. Publish that number. And for the "miracle worker": Double the "boss estimation" and don't work too fast. Just make sure you look very busy.;-)
Speed: The USB 2.0 Hi-Speed FAQ tells us that the maximum speed of USB 2.0 is 480 Mbit/s. The maximum speed of parallel ATA is 133 MByte/s = 1064 MBit/s, plus it does not have the "ATA over USB" protocol overhead. Serial ATA does 150 MByte/s = 1200 MBit/s, IIRC. The ethernet interface of the device supports 100 MBit/s. Modern harddisks can not deliver 133 MByte/s = 1064 MByte/s, but they become faster every day. Flash memory can be that fast, at least for reading.
Power: Each USB port must be able to deliver 5V @ 0.5A.
Now do the maths: You can see that already a single USB device can deliver more data than the ethernet port could transport. The CPU (according to http://www.batbox.org/nslu2-linux.html) is an XScale CPU with 131.48 BogoMIPS, roughly comparable in Performance to a slow Pentium II. I'm sure it can't handle much more than 100 MBit/s Ethernet and two USB 2.0 ports.
Adding a second USB port is convenient to copy data directly between USB devices, e.g. for backup or upgrade purposes. But adding more USB ports costs 0.5A per port for the PSU, making it much more inefficient for each added port.
I would think most people who are able to hack their NSLU2 are also the type that have an extra computer around that they would use instead.
Sure you could use an old PC for that job. But that PC has at least a 150W PSU, often 200W, 250W or more, and almost every PC has at least one noisy fan. My tests on my ex-router (really old Compaq 486 without harddisk) show that a PC needs at least 40W AC power when IDLE, and much more with newer CPUs. According to the Datasheet, the device is specified for 5VDC @ 2A. USB ports must be able to deliver 0.5A each, so the "real" machine needs nothing more than 5V @ 1A. This means you never put more than 10W into the device, with a low power USB storage device, 5W should be possible IMHO. With a common wallbrick PSU (50% heat, 50% output), this translates to 20W AC power under FULL LOAD. With a modern switching PSU (20% heat, 80% output), and a low power USB storage device, you need about 7W AC power. That's what a modern ATX PC draws in standby mode (so-called "off").
Did I mention that the NSLU2 has no moving parts?
Tux2000, not related to Linksys except that I own a hacked WRT54G.
Some time ago, out IT department and an external IT consulting company (recommended by MS) tried to migrate our NT4 Domains (one per office plus some for special purposes) into a single W2k Active Directory. It took more than week full of night shifts and a second IT consulting company to limit the damage caused by scripts of the first IT consulting company. World readable "top secret" documents, completely locked transfer folders, and locked-out users were only the tip of the iceberg.
So here is my advice: Have a verified backup of all working systems, run a lot of tests, and try the migration in a *good* lab environment first (a 1:1 copy of your production systems would be ideal). Repeat several times until everything works smoothly. Run the last tests with recent copies of the production system. DO NOT TRUST SCRIPTS! Verify the result of each script, and make all scripts abort if they find data they can not handle.
Windows / Samba is not a good benchmark. The protocol is so bloated and Windows does so many things behind your back that you can't get reliable results. Use FTP or raw sockets (netcat).
Don't benchmark networks with active Windows systems. Windows floods your net with broadcasts.
Don't copy from or to harddisks for benchmarks. Use ramdisks or packet generators. Recent S-ATA disks have a theoretical peak(!) transfer rate of 160 MByte/s, with a much lower average transfer rate. Compared to a ramdisk, this is just slightly more than a floppy.
Thing about bandwidths of all relevant busses. The PCI bus on a consumer mainboard can do no more that 133 MByte/s in a burst, shared for all available PCI slots. Again, the average transfer rate is much lower. One gigabit ethernet adapter is sufficient to fill the bus. That's why Intel includes the gigabit ethernet interface (except for the physical interface) into the north bridge of recent chipsets, where more bandwitdh is available. The backplane of any available gigabit ethernet switch has enough bandwidth to cope with several gigabit streams.
Don't let a multi-purpose OS on a multi-purpose hardware do the job of a specialised hardware + firmware. It may work using brute force, but special hardware and firmware can do the job with less energy and more performance. This is true for (ab)using linux as hub, switch or bridge. It is also true for WinModems and several other recent "using software instead of hardware" ideas. And as I must confess, Linux on x86 is a bad router platform. My old 486DX4-133 router needed 40W AC power and two noisy fans for the same job that is now done by a fanless 200 MHz MIPS CPU from less than 20W AC power, plus I got a four port switch for free.
A noname 5 port gigabit switch costs about 70 EUR (8 ports for 80 EUR, 1 EUR is about 1 US $). A gigabit ethernet adapter costs about 12 EUR. You currently need two extra cards in the Linux box, 24 EUR. To get a "Linux 5 port switch", you need 4 extra cards, 48 EUR. For a "Linux 8 port switch", you need 7 extra cards, 84 EUR, and perhaps another mainboard. If you look only at the hardware costs, the break even is at about 6 ports. But you trade money against performance, as you already know. Your ordinary PC can't handle the bandwidth of two gigabit connections, and it sure can not handle 8 connections. Looking at the AC power, your PC will draw between 50W (idle) and 300W (working), blowing most of it as heat through a lot of perhaps noisy fans. A real standalone switch is usually fanless, needs about 10W, and can handle 3 or more gigabit connections without trouble.
BTW: Why do you think that you would need a gigabit ethernet for just three home PCs? You have no device (except for/dev/null and ramdisks) that can source or sink a gigabit stream. Your internet connection is almost sure slower than good old ethernet (10 MBit/s). And I know from my own experiments that looking big videos on Windows (667 MHz) from a Linux fileserver (2x 233 MHz, ATA software RAID) is no problem at 100 MBit/s with only three PCs running.
Did you know that you can mix 10 MBit/s, 100 MBit/s and 1 GBit/s devices on a gigabit switch? You know that gigabit adapters can fall back to 100 MBit/s to work with 100 MBit/s devices (crossover cable, hub, 100 MBit/s switch,...) and even to 10 MBit/s for old devices? 100 MBit/s adapters can also fall back to 10 MBit/s.
... well known to my co-workers, for a web-based application:
Some day, all my co-workers at the main office seemed to have to work on my development machine (remote office), so I changed the master password. A phone call some time later: "Please tell me the password." Told him. Machine blocked again a few days later. Wash, rinse, repeat. Finally, I changed the password to "never". Phone call: "Please tell me the password." - "Never." (*klick* speaker on) - "Oh, come on. Tell me the password." - "Never." - "I really need the password." - "Never." And so on for ten funny minutes, with my local co-workers ROTFL.
The default master password was a stupid six-letter word, and often no one bothered to change it when installing the software at the client's site. All attempts to get some attention for the unchanged default master password failed. Now it is a long sentence about insecure default passwords, easy to remember but hard to type. Perhaps that will force them to change the f*ing default password.
Good for the CCD, bad for the accu battery and the LCD. I don't know any accu or primary cell that likes to be frozen. Lead acid batteries can handle it (see any car), but it shortens their live. (And no, I don't now a digital camera with a lead acid battery.) IIRC, alkaline, NiCD and NiMH should not be cooled below 0C. The same applies for LCDs.
Like others already posted, google for amateur astronomers, they know how to cool CCDs.
I doubt this will work. The PC "Keyboard BIOS" is really a small microcontroller (some 8XXX CPU, embedded into the chipset in "modern" machines), that communicates with another microcontroller inside the keyboard (often also an 8XXX or 6502, or a special chip in "modern" keyboards). Both microcontrollers use a crude semi-bidirectional serial protocol with a clock and a data line (For the ugly details, ask IBM or Google).
+5V is usually supplied to the keyboard trough a fuse (rated with up to 2A, often using a self-healing semiconductor fuse) directly from the power supply, ground is ground. Adding a resistor between +5V and ground will slightly increase the power usage of the system, and make your room a little bit warmer. But it won't tell the microcontroller inside the PC that a working keyboard is present.
I've seen these instructions before, but I doubt anyone had success with this. Maybe some other computer once monitored the power drained by the keyboard, but the IBM PC, XT, AT and PS/2 series never did this. They use the clock and data line to check the presence and type of a keyboard.
Right, this tool sometimes works. Especially if you own an ancient 386 or 486 Compaq that does not offer the user interface to set the network server mode flag. But it won't help you with modern Compaqs, they lack the matching BIOS. Quoting the embedded documentation:
SP0667.EXE - NO_F1.COM
Patch Instructions Introduction:
All Compaq 386 and 486 based machines have the ability to run without a keyboard attached. Some Compaq machines allow the user to select this mode of operation by selecting "Network Server" mode in their setup program. For those that do not offer this option NO_F1.COM is a DOS based program that is run one time which sets a bit in CMOS that instructs the BIOS not to wait for a keystroke after displaying the "Press [F1] to continue" message.
This Softpaq (SP0667.EXE) replaces SP0378.EXE which is an older version of NO_F1.COM which does not properly support Prolinea MT and Presario 800 models.
System Requirements:
Any Compaq 386 or 486 based system that does not support "Network Server" mode in it's setup utility program.
Installation instructions:
From a DOS prompt run the program NO_F1.COM
If you subsequently run your setup/configuration program, it may be necessary to run NO_F1.COM in order to run without a keyboard.
Get a working receiver for a cordless keyboard. You do not need the actual keyboard, or a mouse. It also does not matter wheater it works with infrared or radio. Plug it into the computer. Finished. (If you are paranoid, you wrap a radio receiver into grounded tin foil to build a faraday cage, thus eleminating all incoming transmissions from wireless keyboards around. For infrared, some layers of duct tape across the infrared receiver should be sufficient, but tin foil will also do the job.)
Way 2:
Get a really, really, really cheep, but working keyboard (at least one key should work as expected). Open the case. Use brute force if needed. You should find a small printed circuit board (PCB) with one chip and three (or more) LEDs, it should be connected to a keyboard matrix made mostly of transparent foil. Rip off the keyboard matrix, throw away everything except the PCB and the keyboard cable. Place the PCB in a small case, or wrap it into duct tape. Connect the keyboard cable to the computer. Finished.
(Note: If you are really out of luck, the entire keyboard IS a PCB, or even worse, the entire keyboard is made of plastic foil, including the part carrying the chip and LEDs. In this case, getting another keyboard is the easiest way.)
The IT at "my" company seems to love single points of failure. Their motto seems to be "if there is a way to build a SPoF, do it". Recent examples:
The "services office" (where IT, language service, human resources and so on work) is connected through a single line to the "main office" 10 km away. One day, an excavator cut that line. Result: No one could work for hours, because each and every device including all computers and all printers use DHCP to get an IP address. And the DHCP server (and the DNS server) is located in the main office. There was a dedicated print server, but it was not allowed to work as DHCP and DNS server.
All servers in a remote office run on a single UPS. One day, yet another evil excavator cut the power line. All rooms went dark, the UPS switched to battery, all servers were running smoothly. The PBX had and still has no UPS, so only mobile phones still worked. The hotline of the local power authoritiy told us it would take some hours to get the line fixed. So we needed to shut down the servers before the UPS battery was drained. But except for one or two servers, our IT supporter had no privileges to shut down the servers, so it had to be done from the main office. But neither the ethernet switches nor the router to the main office were connected to the UPS. We finally decided that the servers had had enough time to write their caches to the disks and simply disconnected them. And no, the UPS signal output was not connected to the servers. Now, it could signal a power outage and a low battery via ethernet -- if the switches were connected to a UPS.
Did I mention that all servers in that remote office are connected to a single switch (out of three), using up to three ethernet lines?
Did I mention various air conditions that can not cope with the heat of the servers on a hot summer day?
Did I mention that all remote office data lines (yes, one line per office) end in a single point in the main office?
Did I mention that we have a single mail server (MX for the domain) at our provider for all incoming external mail which is regularily blacklisted and that our internal MX consults that black lists to fight spam?
(Hmm, I should really stop here or I won't finish until tomorrow.)
First, you don't need the OS encrypted. The most you'll need is/var,/tmp,/home and swap (and/data or wherever you put your bulk data files).
Right. Encrypting the OS is encrypting known data. <paranoia>This may lead to a key recovery by just comparing the encrypted OS and the unencrypted OS.</paranoia> Regarding encrypted swap: I think encrypting swap slows down too much. Just a dd if=/dev/random of=swappartition in the shutdown script should do the job as well. OK, if someone knocks you down, rips out the HDD of your PC before it has a chance to shut down properly, and runs strings swappartition | grep somepattern, he might find something useful. Decide yourself how important your data security is. You could also add some more RAM and do not use a swap partition at all.
I think "Windows in VMware" is the way to do work with encrypted data with more than one OS. Always boot Linux, configure its firewall as paranoid as possible, and "wire" Windows in it's VMware box to the inside of the firewall. Mount your data directory via Samba in Windows. Encrypt and decrypt it in Linux. Make sure there is no way to connect to the VMware virtual LAN except inside your PC; this also means to set up firewall (iptables) rules to prohibit a connection to Samba from the real LAN or Wireless LAN.
I use rechargeables for everything and in every size except 9V.
Rechargeables are not the best idea for low(est) power devices. A common TV remote control operates on two, sometimes three Alkaline cells for two or three years, sometimes even longer. Most, if not all, remote control chips are optimized for this operation mode. True, if you press a button on a remote, it draws a lot of power for the infrared LEDs. One ampere is not unusual. But it sends very short pulses (or else the LEDs would literally burn), and the real power does not come from the battery, but from an electrolyte capacitor contacted in parallel, via a very small resistor, to the battery. It is rapidly discharged by the LEDs, and then charged again by the battery. The battery just has to provide a very small current to recharge the capacitor after each pulse, and an even smaller, permanent current for the controller chip. Alkaline cells have a very low self-discharge factor, so they can work like this for years. The remote control continues to work even if the alkaline cell is nearly dead (1.3 V), but with a reduced range (because of the lower available energy due to the lower voltage) and with a noticeable recovery time. When the voltage is so low that the chip does not work properly, it starts to send nonsense to the TV (and may under certain circumstances crash the microprocessor inside the TV - nobody is perfect).
If you replace the alkaine cells with NiCD or NiMH, this lowers the available voltage by 0.3 volts per cell (NiCD=1.2 V, alkaline=1.5 V), thus reducing the available engergy for the LEDs, thus reducing the range of the remote control. Those rechargable cells discharge themselves, so you need to recharge them often. But at that time, you want to use your remote control as well, so you need two pairs of rechargeable cells, and you need a charger. Probably a good one, so it does not kill your cells while loading them. And after some time, your rechargeables are dead because they had nothing "to do". So you need two new pairs of cells, probably after two years. Just about after the same time that you would have bought the first new pair of alkalines (the first one is usually included).
Just calculate for yourself: Using rechargeables in remote control is nonsense. And if you are concerned about the environmental impact: At least here in Germany, there is a well-working return system for alkaline (and zinc carbon) cells (in fact, you have to return your used cells).
The same is also true for other low(est) power devices, like clocks and calculators with liquid crystal displays (LCDs), and most analog quartz clocks. Radio controlled analog clocks need a little bit more power and a "high" voltage for the receiver, so they can not drain alkaline cells as much as clocks without receiver and remote controls.
My "BRAUN" electical toothbrush uses a single NiCd cell which is permanently charged when it sits in its "docking station". After about one year of permanent "keep alive" charging, the cell is now nearly dead. It still can hold enough energy to work for five minutes, but when it is disconnected from the wall outlet for more than a few hours, it simply does not work and needs two days to recharge. The cell is sealed into the toothbrush, having a predetermined breaking point in the case to take out the cell for disposal, so I can not replace the cell without permanently damaging the toothbrush. A trick I've seen too late on TV: Do not permanently charge the toothbrush, discharge it while using, and connect it only to mains when it is empty. It will need round about a day to charge. This will avoid the memory effect.
About batteries and cells:
A cell consists of two electrodes (the positive anode and the negative cathode), the electrolyte, and a case. Depending on the materials used, a single cell has 1.2V (NiCd and NiMH), 1.4V (zinc air button cells), 1.5V (alkaline and zinc carbon), 1.55V (Silver / Quicksilver button cells), 2V (Pb=lead acid), 3V (Lithium), o
Like many others said: Get a cheap "internet router" that does NAT (Network Address Translation). If the attackers can't get to the fresh XP machine, they can't kill it. Easy, isn't it? Just turn OFF UPNP support and all DMZ / port forwarding stuff on the router.
If you still have a spare PC (minimum 486SX-25, 8 MB RAM, Floppy, two ethernet cards), give fli4l (or any other small Linux router software) a try. Download size is a few MBytes (ask your friends / neighboors), complete boot floppy is created within a few minutes on any Windows system. No linux knowledge required.
Keep the NAT router between the XP machine and your internet connection even after you have completed the XP setup. Though the router may not help against using IE and Outlook, it will help against all TCP and UDP based attacks. All viri and worms that spread by connecting to any TCP or UDP port on your machine will fail to infect your machine thanks to the NAT router.
Buy a "real" modem with a UART, preferably an external serial modem (RS232). Yes, a lot of people already said this already. But it's the only way to avoid trouble. There is no need to buy an expensive brand, just any external modem with a 9-pin or 25-pin connector will do the job. USB modems are often WinModems, so are most PCI modems. ISA is dead. ISA modems are often "real" modems with a build-in COM port (i.e. UART), but there are some ISA WinModems.
Even if someone would try to build a serial port WinModem, he would fail terribly: the serial port is fast enough for the well-known Hayes commands even at 56.000 baud, but it is way to slow for a WinModem sampling the phone line and doing the modem part in Software. So an external serial modem can't ever be a WinModem.
And by the way: Yes, I have a success story. My WinModem in my old Toshiba Tecra 8200 "accidentally" works. I just had to try two or three different drivers that all claimed not to work with my WinModem. Thanks to http://linmodems.technion.ac.il/! (But I just don't want to know what happens when I update my kernel.)
I do not use a smartphone and I think I will never use one - for a simple reason: My "ancient" Nokia 5130 (modified 5110 for 1800 MHz) works, works, works.
There are no crashes because of a bug in some "smart" component I do not need. It does not support ringtones, logos or other kiddie nonsense. It is a telephone I can simply carry around and use it whenever I need it. It shows the caller's number (and name if it is in the phone book). Nice to have, but not really needed. It has a simple calculator. Nice to have, but not really needed. It can send and receive SMS. Nice to have. It has two or three stupid games. Definitely not needed.
Why should I want to use a smartphone? Taking notes? I prefer pen and paper, or I call my own answering machine if I can't get a piece of paper. Photos? That's why I own a good digital camera and a cheap analog camera. Writing letters and documentations? I own several PCs, thank you. Navigation? Installed in my car (PDA-based).
I really can't see any use for a smart phone. ("Nanana I have a smartphone and you don't!" and "My smartphone is much smarter than your smartphone." are no reasons for buying a smartphone.)
APM is only defined for single CPU machines. APM + SMP (Hyperthreading counts as a secondary CPU!) simply does not work with Linux (except for power-off via APM) last time I checked (2.4.20+n).
I wish my old Dell dual CPU machine had ACPI so I could shut it down using the power button. But alas, APM does not "hook" the power button like ACPI does, so pressing the power button is just like unplugging the power cord - a bad idea. For now, ca::ctrlaltdel:/sbin/shutdown -t5 -h now in/etc/inittab must do the job.
Maybe the best choice would be to have translatable keywords for a language
Have you ever seen Visual Basic for Applications in a localized version of MS Office? It really hurts the eye. If you have ever coded an advanced hello world programm in nearly any language, you know what a FOR loop looks like. If you look at VBA with translated keywords, you can't see anything but bla bla because of the translated keywords.
English is very helpful for keywords because you can understand english sentences no matter what order the words are in ("to station go now I"). In other languages (like german and french), order of words is more important, so the pseudo-english grammar of many computer languages does not match the translated keywords. It is simply much harder to read german with a pseudo-english grammar than reading english with a pseudo-english grammar.
I use to code completely in english (including comments and docs) for about 5 years now. It is just easier to read than mixing german comments and message strings with the english grammar of the language. And it has the nice side-effect that others can read my code without the babelfish.
BTW: See also Scarblac's posting "Translated Visual Basic" for a nice comparison to music and its italian "keywords".
Don't re-invent the wheel. Get a customizable product and an expert that can customize it.
I suggest Livelink. Well, it's not free. It costs money. It may cost lots of money if you want all those nice features. It's not open source. But I have enough Karma to burn.;-)
The consulting company I work for is based on knowledge. Fast, reliable and secure (permisson based) access to archived knowledge is mission critical. So there never was a problem buying the software we need for business, no matter what it costs.
My job is not Livelink. But I work in the same room as our Livelink expert. So I collect a little bit of knowlegde about Livelink. I'm the one he asks for Unix and network tricks.
Livelink has a document management (that's the main part), team rooms, workstreams, and a lot of other nice features. For details, have a look at the web page. Livelink is a core server, extended by a lot of scripts (in a custom language named Oscript), and a tiny CGI that passes requests from the webserver to the core server. If you own a development kit, you can customize nearly every aspect of Livelink, and you can see lots of code written by Opentext. So if you have the money, you can at least see most of the sources.
We use three dual-CPU W2K machines with Apache 1.3.x as Web and application servers, a fourth dual-CPU W2K machine for the indexer and search engine, a Sun 420 running Solaris 9 for the database (Oracle), and Linux Virtual Server (LVS) as load balancer for the webservers. We have about 1500 users all around the world.
Why so many servers? Most of the time, one web server is completely idle. Opentext would recommend a single server setup, and that would be sufficient. But we have demanding consultants, our problems are response time and availability. We have some queries that block a server for a while. So we need at least two servers. The third server is for load peaks and for downtimes of one of the other servers. Index and search also need a lot of power that would block a single machine, so it's placed on the fourth server.
Why W2K? The most recent version of Livelink requires it.
Why Sun? Oracle on Windows simply sucks, the raw CPU power of the previous multi-CPU x86 database machine was larger than the one of the Sun machine, but Oracle runs much faster on the Sun. (Now all corporate databases are switched to a Oracle/Sun cluster, but that's a different story.)
Why LVS? Simple: It works. We tried a load-balancing software called Resonate, a really fitting name for a piece of software that should implement a control loop. We kicked it because it was hard to maintain and did not work reliably on our machines. We tried LVS on a really old desktop and it worked great, even if we tried really hard to confuse it. Now it has its own x86 server running Slackware, and we did not have a single second of trouble with it.
Why Apache? We used Netscape Enterprise Server / iPlanet. It had a pretty web-based config tool and much bloat, and it costs money. Apache does the same job for free, and its configuration is a simple text file that can be copied to the various servers. MS IIS has bugs. Lots of bugs. Its mouse controlled. We did not even think about a test system with the IIS.
Slashdot Mirror
Rubidium 87 has a half-life of 47 billion (10^9) years (our soloar system is not yet 5 billion years old). Uranium 238 has a half-life of 4.5 Billion (10^9) years, Plutonium 239 has a half-life of 25.000 years. Half-life means that after some billion years, you still have half of your nuclear waste happily emitting radioactivity, while the other half has decayed to other, possibly also radioactive elements. After 7 times the half-life (7*47*10^9 years = 329*10^9 years), you still have round about 1 % of the original radioactive waste (2^-7 = 1/128 ~ 1%) and a lot of other radioactive products.
There is no final solution of nuclear waste, and there probably will never be one. It is practically impossible to guarantee a safe place for at least 7 half-lifes of Rubidium 87 or Uranium 238. Even if you can dig a deep hole and fill it with nulear waste, there still is a possibility much larger than zero that someone digs it out again in some thousand years and does not know what that shiny stuff is. There have already been accidents in third world countries, where poor and uneducated people digged out radioactive materials from medical devices on waste dumps. And a mountain used to store nuclear waste may erode before the waste becomes harmless.
Lessons learned: PR people don't know anything about mathematics and radioactivity.
Tux2000
Someone may want to read http://www.hifiaktiv.at/diverses/realistische_betr achtungen.htm (in german).
Tux2000
I could not agree more. But this "Hi Fi" junk did not only come from Japan. In Germany, the (now broken) company "Schneider" produced large ammounts of cheap stereo towers (with tuner, double tape, amplifier, and record player, later also CD player), featuring large cases containing much air, a small transformer and a type plate promising 1000W music output from 50W AC input. No, they did not invent a perpetuum mobile. But their unrealistic output ratings coined the term "Schneider-Watt" for ridiculous amplifier output ratings, especially above the rated input power. I won't talk much about the sound of this devices or about the speaker boxes made of cheap plastic and cardboard. You should be able to guess what you get from these pieces of crap.
Personally, I prefer old high quality transistor amplifiers from the time when Japan imports were just a footnote in the german HiFi business.
Tux2000
Perhaps because he uses old tube equipment without line inputs. Old tube radios are often driven by rectified mains voltage (so you get some hundred volts inside the radio on nearly all components), without an insulating transformer (so this voltage is "available" against earth and can kill you). Adding a line input to such a (simple and cheap) design requires an insulating transformer either for the power supply or for the line input, which would have caused additional costs.
It is possible to retrofit a line input to most old tube radios, but not without dramatic changes to the device. You need at least an additional switch and a hole in the backside. Most people who love old tube radios would rather like several root canal treatments without anesthesia than that.
Some "newer" and expensive old tube radios have inputs for a record player and/or a tape, both could be used to connect modern audio devices like the iPod, but not necessarily without mechanical and electrical adapters.
So the most easiest way to "connect" an iPod to old tube radios is an FM transmitter. As a nice side effect, you can "connect" several radios to the same iPod, all without fiddling with cables.
And by the way, frequencies below 50 Hz and above 15 kHz can only be heard by very young people. The older you get, the narrower the bandwith of your ears becomes.
Tux2000
This rule works quite well, I use it whenever possible. From my own experiences, most people estimate very bad. The good ones can get to the point where they estimate 50% of the actual time needed (the other ones estimate much less time). So doubling that time gives you a very good estimation for your own. Never tell anyone that number, don't write it down! Rounding up to the next highest units gives your boss a nice round number and you get a little time for all those catastrophes you did not think about when estimating. Publish that number. And for the "miracle worker": Double the "boss estimation" and don't work too fast. Just make sure you look very busy. ;-)
Tux2000, yet another Scotty fan.
Speed: The USB 2.0 Hi-Speed FAQ tells us that the maximum speed of USB 2.0 is 480 Mbit/s. The maximum speed of parallel ATA is 133 MByte/s = 1064 MBit/s, plus it does not have the "ATA over USB" protocol overhead. Serial ATA does 150 MByte/s = 1200 MBit/s, IIRC. The ethernet interface of the device supports 100 MBit/s. Modern harddisks can not deliver 133 MByte/s = 1064 MByte/s, but they become faster every day. Flash memory can be that fast, at least for reading.
Power: Each USB port must be able to deliver 5V @ 0.5A.
Now do the maths: You can see that already a single USB device can deliver more data than the ethernet port could transport. The CPU (according to http://www.batbox.org/nslu2-linux.html) is an XScale CPU with 131.48 BogoMIPS, roughly comparable in Performance to a slow Pentium II. I'm sure it can't handle much more than 100 MBit/s Ethernet and two USB 2.0 ports.
Adding a second USB port is convenient to copy data directly between USB devices, e.g. for backup or upgrade purposes. But adding more USB ports costs 0.5A per port for the PSU, making it much more inefficient for each added port.
Tux2000
I would think most people who are able to hack their NSLU2 are also the type that have an extra computer around that they would use instead.
Sure you could use an old PC for that job. But that PC has at least a 150W PSU, often 200W, 250W or more, and almost every PC has at least one noisy fan. My tests on my ex-router (really old Compaq 486 without harddisk) show that a PC needs at least 40W AC power when IDLE, and much more with newer CPUs. According to the Datasheet, the device is specified for 5VDC @ 2A. USB ports must be able to deliver 0.5A each, so the "real" machine needs nothing more than 5V @ 1A. This means you never put more than 10W into the device, with a low power USB storage device, 5W should be possible IMHO. With a common wallbrick PSU (50% heat, 50% output), this translates to 20W AC power under FULL LOAD. With a modern switching PSU (20% heat, 80% output), and a low power USB storage device, you need about 7W AC power. That's what a modern ATX PC draws in standby mode (so-called "off").
Did I mention that the NSLU2 has no moving parts?
Tux2000, not related to Linksys except that I own a hacked WRT54G.
Some time ago, out IT department and an external IT consulting company (recommended by MS) tried to migrate our NT4 Domains (one per office plus some for special purposes) into a single W2k Active Directory. It took more than week full of night shifts and a second IT consulting company to limit the damage caused by scripts of the first IT consulting company. World readable "top secret" documents, completely locked transfer folders, and locked-out users were only the tip of the iceberg.
So here is my advice: Have a verified backup of all working systems, run a lot of tests, and try the migration in a *good* lab environment first (a 1:1 copy of your production systems would be ideal). Repeat several times until everything works smoothly. Run the last tests with recent copies of the production system. DO NOT TRUST SCRIPTS! Verify the result of each script, and make all scripts abort if they find data they can not handle.
Tux2000
A noname 5 port gigabit switch costs about 70 EUR (8 ports for 80 EUR, 1 EUR is about 1 US $). A gigabit ethernet adapter costs about 12 EUR. You currently need two extra cards in the Linux box, 24 EUR. To get a "Linux 5 port switch", you need 4 extra cards, 48 EUR. For a "Linux 8 port switch", you need 7 extra cards, 84 EUR, and perhaps another mainboard. If you look only at the hardware costs, the break even is at about 6 ports. But you trade money against performance, as you already know. Your ordinary PC can't handle the bandwidth of two gigabit connections, and it sure can not handle 8 connections. Looking at the AC power, your PC will draw between 50W (idle) and 300W (working), blowing most of it as heat through a lot of perhaps noisy fans. A real standalone switch is usually fanless, needs about 10W, and can handle 3 or more gigabit connections without trouble.
BTW: Why do you think that you would need a gigabit ethernet for just three home PCs? You have no device (except for /dev/null and ramdisks) that can source or sink a gigabit stream. Your internet connection is almost sure slower than good old ethernet (10 MBit/s). And I know from my own experiments that looking big videos on Windows (667 MHz) from a Linux fileserver (2x 233 MHz, ATA software RAID) is no problem at 100 MBit/s with only three PCs running.
Did you know that you can mix 10 MBit/s, 100 MBit/s and 1 GBit/s devices on a gigabit switch? You know that gigabit adapters can fall back to 100 MBit/s to work with 100 MBit/s devices (crossover cable, hub, 100 MBit/s switch, ...) and even to 10 MBit/s for old devices? 100 MBit/s adapters can also fall back to 10 MBit/s.
Tux2000
... well known to my co-workers, for a web-based application:
Some day, all my co-workers at the main office seemed to have to work on my development machine (remote office), so I changed the master password. A phone call some time later: "Please tell me the password." Told him. Machine blocked again a few days later. Wash, rinse, repeat. Finally, I changed the password to "never". Phone call: "Please tell me the password." - "Never." (*klick* speaker on) - "Oh, come on. Tell me the password." - "Never." - "I really need the password." - "Never." And so on for ten funny minutes, with my local co-workers ROTFL.
The default master password was a stupid six-letter word, and often no one bothered to change it when installing the software at the client's site. All attempts to get some attention for the unchanged default master password failed. Now it is a long sentence about insecure default passwords, easy to remember but hard to type. Perhaps that will force them to change the f*ing default password.
Tux2000
Good for the CCD, bad for the accu battery and the LCD. I don't know any accu or primary cell that likes to be frozen. Lead acid batteries can handle it (see any car), but it shortens their live. (And no, I don't now a digital camera with a lead acid battery.) IIRC, alkaline, NiCD and NiMH should not be cooled below 0C. The same applies for LCDs.
Like others already posted, google for amateur astronomers, they know how to cool CCDs.
Tux2000
I doubt this will work. The PC "Keyboard BIOS" is really a small microcontroller (some 8XXX CPU, embedded into the chipset in "modern" machines), that communicates with another microcontroller inside the keyboard (often also an 8XXX or 6502, or a special chip in "modern" keyboards). Both microcontrollers use a crude semi-bidirectional serial protocol with a clock and a data line (For the ugly details, ask IBM or Google).
+5V is usually supplied to the keyboard trough a fuse (rated with up to 2A, often using a self-healing semiconductor fuse) directly from the power supply, ground is ground. Adding a resistor between +5V and ground will slightly increase the power usage of the system, and make your room a little bit warmer. But it won't tell the microcontroller inside the PC that a working keyboard is present.
I've seen these instructions before, but I doubt anyone had success with this. Maybe some other computer once monitored the power drained by the keyboard, but the IBM PC, XT, AT and PS/2 series never did this. They use the clock and data line to check the presence and type of a keyboard.
Tux2000
Right, this tool sometimes works. Especially if you own an ancient 386 or 486 Compaq that does not offer the user interface to set the network server mode flag. But it won't help you with modern Compaqs, they lack the matching BIOS. Quoting the embedded documentation:
Tux2000
Way 1:
Get a working receiver for a cordless keyboard. You do not need the actual keyboard, or a mouse. It also does not matter wheater it works with infrared or radio. Plug it into the computer. Finished. (If you are paranoid, you wrap a radio receiver into grounded tin foil to build a faraday cage, thus eleminating all incoming transmissions from wireless keyboards around. For infrared, some layers of duct tape across the infrared receiver should be sufficient, but tin foil will also do the job.)
Way 2:
Get a really, really, really cheep, but working keyboard (at least one key should work as expected). Open the case. Use brute force if needed. You should find a small printed circuit board (PCB) with one chip and three (or more) LEDs, it should be connected to a keyboard matrix made mostly of transparent foil. Rip off the keyboard matrix, throw away everything except the PCB and the keyboard cable. Place the PCB in a small case, or wrap it into duct tape. Connect the keyboard cable to the computer. Finished.
(Note: If you are really out of luck, the entire keyboard IS a PCB, or even worse, the entire keyboard is made of plastic foil, including the part carrying the chip and LEDs. In this case, getting another keyboard is the easiest way.)
Tux2000
The IT at "my" company seems to love single points of failure. Their motto seems to be "if there is a way to build a SPoF, do it". Recent examples:
The "services office" (where IT, language service, human resources and so on work) is connected through a single line to the "main office" 10 km away. One day, an excavator cut that line. Result: No one could work for hours, because each and every device including all computers and all printers use DHCP to get an IP address. And the DHCP server (and the DNS server) is located in the main office. There was a dedicated print server, but it was not allowed to work as DHCP and DNS server.
All servers in a remote office run on a single UPS. One day, yet another evil excavator cut the power line. All rooms went dark, the UPS switched to battery, all servers were running smoothly. The PBX had and still has no UPS, so only mobile phones still worked. The hotline of the local power authoritiy told us it would take some hours to get the line fixed. So we needed to shut down the servers before the UPS battery was drained. But except for one or two servers, our IT supporter had no privileges to shut down the servers, so it had to be done from the main office. But neither the ethernet switches nor the router to the main office were connected to the UPS. We finally decided that the servers had had enough time to write their caches to the disks and simply disconnected them. And no, the UPS signal output was not connected to the servers. Now, it could signal a power outage and a low battery via ethernet -- if the switches were connected to a UPS.
Did I mention that all servers in that remote office are connected to a single switch (out of three), using up to three ethernet lines?
Did I mention various air conditions that can not cope with the heat of the servers on a hot summer day?
Did I mention that all remote office data lines (yes, one line per office) end in a single point in the main office?
Did I mention that we have a single mail server (MX for the domain) at our provider for all incoming external mail which is regularily blacklisted and that our internal MX consults that black lists to fight spam?
(Hmm, I should really stop here or I won't finish until tomorrow.)
Tux2000
First, you don't need the OS encrypted. The most you'll need is /var, /tmp, /home and swap (and /data or wherever you put your bulk data files).
Right. Encrypting the OS is encrypting known data. <paranoia>This may lead to a key recovery by just comparing the encrypted OS and the unencrypted OS.</paranoia> Regarding encrypted swap: I think encrypting swap slows down too much. Just a dd if=/dev/random of=swappartition in the shutdown script should do the job as well. OK, if someone knocks you down, rips out the HDD of your PC before it has a chance to shut down properly, and runs strings swappartition | grep somepattern, he might find something useful. Decide yourself how important your data security is. You could also add some more RAM and do not use a swap partition at all.
I think "Windows in VMware" is the way to do work with encrypted data with more than one OS. Always boot Linux, configure its firewall as paranoid as possible, and "wire" Windows in it's VMware box to the inside of the firewall. Mount your data directory via Samba in Windows. Encrypt and decrypt it in Linux. Make sure there is no way to connect to the VMware virtual LAN except inside your PC; this also means to set up firewall (iptables) rules to prohibit a connection to Samba from the real LAN or Wireless LAN.
Tux2000
Rechargeables are not the best idea for low(est) power devices. A common TV remote control operates on two, sometimes three Alkaline cells for two or three years, sometimes even longer. Most, if not all, remote control chips are optimized for this operation mode. True, if you press a button on a remote, it draws a lot of power for the infrared LEDs. One ampere is not unusual. But it sends very short pulses (or else the LEDs would literally burn), and the real power does not come from the battery, but from an electrolyte capacitor contacted in parallel, via a very small resistor, to the battery. It is rapidly discharged by the LEDs, and then charged again by the battery. The battery just has to provide a very small current to recharge the capacitor after each pulse, and an even smaller, permanent current for the controller chip. Alkaline cells have a very low self-discharge factor, so they can work like this for years. The remote control continues to work even if the alkaline cell is nearly dead (1.3 V), but with a reduced range (because of the lower available energy due to the lower voltage) and with a noticeable recovery time. When the voltage is so low that the chip does not work properly, it starts to send nonsense to the TV (and may under certain circumstances crash the microprocessor inside the TV - nobody is perfect).
If you replace the alkaine cells with NiCD or NiMH, this lowers the available voltage by 0.3 volts per cell (NiCD=1.2 V, alkaline=1.5 V), thus reducing the available engergy for the LEDs, thus reducing the range of the remote control. Those rechargable cells discharge themselves, so you need to recharge them often. But at that time, you want to use your remote control as well, so you need two pairs of rechargeable cells, and you need a charger. Probably a good one, so it does not kill your cells while loading them. And after some time, your rechargeables are dead because they had nothing "to do". So you need two new pairs of cells, probably after two years. Just about after the same time that you would have bought the first new pair of alkalines (the first one is usually included).
Just calculate for yourself: Using rechargeables in remote control is nonsense. And if you are concerned about the environmental impact: At least here in Germany, there is a well-working return system for alkaline (and zinc carbon) cells (in fact, you have to return your used cells).
The same is also true for other low(est) power devices, like clocks and calculators with liquid crystal displays (LCDs), and most analog quartz clocks. Radio controlled analog clocks need a little bit more power and a "high" voltage for the receiver, so they can not drain alkaline cells as much as clocks without receiver and remote controls.
My "BRAUN" electical toothbrush uses a single NiCd cell which is permanently charged when it sits in its "docking station". After about one year of permanent "keep alive" charging, the cell is now nearly dead. It still can hold enough energy to work for five minutes, but when it is disconnected from the wall outlet for more than a few hours, it simply does not work and needs two days to recharge. The cell is sealed into the toothbrush, having a predetermined breaking point in the case to take out the cell for disposal, so I can not replace the cell without permanently damaging the toothbrush. A trick I've seen too late on TV: Do not permanently charge the toothbrush, discharge it while using, and connect it only to mains when it is empty. It will need round about a day to charge. This will avoid the memory effect.
About batteries and cells:
A cell consists of two electrodes (the positive anode and the negative cathode), the electrolyte, and a case. Depending on the materials used, a single cell has 1.2V (NiCd and NiMH), 1.4V (zinc air button cells), 1.5V (alkaline and zinc carbon), 1.55V (Silver / Quicksilver button cells), 2V (Pb=lead acid), 3V (Lithium), o
Like many others said: Get a cheap "internet router" that does NAT (Network Address Translation). If the attackers can't get to the fresh XP machine, they can't kill it. Easy, isn't it? Just turn OFF UPNP support and all DMZ / port forwarding stuff on the router.
If you still have a spare PC (minimum 486SX-25, 8 MB RAM, Floppy, two ethernet cards), give fli4l (or any other small Linux router software) a try. Download size is a few MBytes (ask your friends / neighboors), complete boot floppy is created within a few minutes on any Windows system. No linux knowledge required.
Keep the NAT router between the XP machine and your internet connection even after you have completed the XP setup. Though the router may not help against using IE and Outlook, it will help against all TCP and UDP based attacks. All viri and worms that spread by connecting to any TCP or UDP port on your machine will fail to infect your machine thanks to the NAT router.
Tux2000
Buy a "real" modem with a UART, preferably an external serial modem (RS232). Yes, a lot of people already said this already. But it's the only way to avoid trouble. There is no need to buy an expensive brand, just any external modem with a 9-pin or 25-pin connector will do the job. USB modems are often WinModems, so are most PCI modems. ISA is dead. ISA modems are often "real" modems with a build-in COM port (i.e. UART), but there are some ISA WinModems.
Even if someone would try to build a serial port WinModem, he would fail terribly: the serial port is fast enough for the well-known Hayes commands even at 56.000 baud, but it is way to slow for a WinModem sampling the phone line and doing the modem part in Software. So an external serial modem can't ever be a WinModem.
And by the way: Yes, I have a success story. My WinModem in my old Toshiba Tecra 8200 "accidentally" works. I just had to try two or three different drivers that all claimed not to work with my WinModem. Thanks to http://linmodems.technion.ac.il/! (But I just don't want to know what happens when I update my kernel.)
Tux2000
Simple:
Advanced:
I do not use a smartphone and I think I will never use one - for a simple reason: My "ancient" Nokia 5130 (modified 5110 for 1800 MHz) works, works, works.
There are no crashes because of a bug in some "smart" component I do not need. It does not support ringtones, logos or other kiddie nonsense. It is a telephone I can simply carry around and use it whenever I need it. It shows the caller's number (and name if it is in the phone book). Nice to have, but not really needed. It has a simple calculator. Nice to have, but not really needed. It can send and receive SMS. Nice to have. It has two or three stupid games. Definitely not needed.
Why should I want to use a smartphone? Taking notes? I prefer pen and paper, or I call my own answering machine if I can't get a piece of paper. Photos? That's why I own a good digital camera and a cheap analog camera. Writing letters and documentations? I own several PCs, thank you. Navigation? Installed in my car (PDA-based).
I really can't see any use for a smart phone. ("Nanana I have a smartphone and you don't!" and "My smartphone is much smarter than your smartphone." are no reasons for buying a smartphone.)
Tux2000
APM is only defined for single CPU machines. APM + SMP (Hyperthreading counts as a secondary CPU!) simply does not work with Linux (except for power-off via APM) last time I checked (2.4.20+n).
I wish my old Dell dual CPU machine had ACPI so I could shut it down using the power button. But alas, APM does not "hook" the power button like ACPI does, so pressing the power button is just like unplugging the power cord - a bad idea. For now, ca::ctrlaltdel:/sbin/shutdown -t5 -h now in /etc/inittab must do the job.
Tux2000
Have you ever seen Visual Basic for Applications in a localized version of MS Office? It really hurts the eye. If you have ever coded an advanced hello world programm in nearly any language, you know what a FOR loop looks like. If you look at VBA with translated keywords, you can't see anything but bla bla because of the translated keywords.
English is very helpful for keywords because you can understand english sentences no matter what order the words are in ("to station go now I"). In other languages (like german and french), order of words is more important, so the pseudo-english grammar of many computer languages does not match the translated keywords. It is simply much harder to read german with a pseudo-english grammar than reading english with a pseudo-english grammar.
I use to code completely in english (including comments and docs) for about 5 years now. It is just easier to read than mixing german comments and message strings with the english grammar of the language. And it has the nice side-effect that others can read my code without the babelfish.
BTW: See also Scarblac's posting "Translated Visual Basic" for a nice comparison to music and its italian "keywords".
Tux2000
... completely without letters if you do it right! ;-)
Don't re-invent the wheel. Get a customizable product and an expert that can customize it.
I suggest Livelink. Well, it's not free. It costs money. It may cost lots of money if you want all those nice features. It's not open source. But I have enough Karma to burn. ;-)
Web page: http://www.opentext.com/
The consulting company I work for is based on knowledge. Fast, reliable and secure (permisson based) access to archived knowledge is mission critical. So there never was a problem buying the software we need for business, no matter what it costs.
My job is not Livelink. But I work in the same room as our Livelink expert. So I collect a little bit of knowlegde about Livelink. I'm the one he asks for Unix and network tricks.
Livelink has a document management (that's the main part), team rooms, workstreams, and a lot of other nice features. For details, have a look at the web page. Livelink is a core server, extended by a lot of scripts (in a custom language named Oscript), and a tiny CGI that passes requests from the webserver to the core server. If you own a development kit, you can customize nearly every aspect of Livelink, and you can see lots of code written by Opentext. So if you have the money, you can at least see most of the sources.
We use three dual-CPU W2K machines with Apache 1.3.x as Web and application servers, a fourth dual-CPU W2K machine for the indexer and search engine, a Sun 420 running Solaris 9 for the database (Oracle), and Linux Virtual Server (LVS) as load balancer for the webservers. We have about 1500 users all around the world.
Why so many servers? Most of the time, one web server is completely idle. Opentext would recommend a single server setup, and that would be sufficient. But we have demanding consultants, our problems are response time and availability. We have some queries that block a server for a while. So we need at least two servers. The third server is for load peaks and for downtimes of one of the other servers. Index and search also need a lot of power that would block a single machine, so it's placed on the fourth server.
Why W2K? The most recent version of Livelink requires it.
Why Sun? Oracle on Windows simply sucks, the raw CPU power of the previous multi-CPU x86 database machine was larger than the one of the Sun machine, but Oracle runs much faster on the Sun. (Now all corporate databases are switched to a Oracle/Sun cluster, but that's a different story.)
Why LVS? Simple: It works. We tried a load-balancing software called Resonate, a really fitting name for a piece of software that should implement a control loop. We kicked it because it was hard to maintain and did not work reliably on our machines. We tried LVS on a really old desktop and it worked great, even if we tried really hard to confuse it. Now it has its own x86 server running Slackware, and we did not have a single second of trouble with it.
Why Apache? We used Netscape Enterprise Server / iPlanet. It had a pretty web-based config tool and much bloat, and it costs money. Apache does the same job for free, and its configuration is a simple text file that can be copied to the various servers. MS IIS has bugs. Lots of bugs. Its mouse controlled. We did not even think about a test system with the IIS.
Tux2000