What's so hard about using a trigger port/ports to open everything else up?
It's just another example of security through obscurity, that's what's wrong with it. IP-based authentication is worthless.
I've said it before and I'll say it again and again. Security through obscurity is a perfectly fine LAYER to add to a security regimen.
Sure you've got everything tuned up pretty good but there is nothing wrong and I strongly encourage the use of an obscurity layer in order to build up defences a little more. Relying on obscurity to protect you is one thing, and it's a very bad thing. But having the attacker have to guess what the hell he's looking at before he can apply his standard toolkits and procedures is always a good thing. Always. You don't have a little sticker on your house door which says "Dudley EX-145 model lock" now do you?
Now bringing this back to the topic at hand... If I have a trigger port that will open up all my services (or a selected service) when tickled just right is a fine way to keep the scanners at bay. Even better is if you put that trigger port on auth and your auth demon is tied in with your inetd server: If you get an auth request for port 12345 followed by a request for port 54321 within a 5 second window, it tells inetd to start listening on whichever ports you specify for the next 1 minute (or something).
Tell me, what's so insecure about that, if you've got all your other defences built up?
Its/possible/ for someone at the other end of the wire to send packets to you with a destination-IP of 127.0.0.1, and your box will happily accept them. Sure, this is a long-shot, and there's no way the hacker at the central-office will get a response, but there are a lot of attacks that dont need responses.
That's why my IPCHAINS input chain looks a little like this:
I've played a bit with deny policies on the input chain but they're really too restrictive and your chain starts to get a little long, especially if you're doing more than just telnet/ssh/ftp/web. or have more than just two interfaces (think incoming dialup, DSL, local network and VPN) The forward policy, however, is always deny. Always.
Given that a regular modem involves a CPU response to almost every single character (a DSL interface won't require that)
I believe that some cheap-ass NICs are almost as bad. 3COM's Parallel tasking chipset (in the 3C905B) is very good about not using your CPU to bring in data.
That's right, a Pentium 90. Not as bad as a 486, but no great shakes.
A pox on both your P90 and the other guy's 486!
I'm doing the masq/firewall mambo with a 12 (10?) year old 80386DX/40. This poor machine was purchased new by my father for a small fortune and a few years ago I claimed it.
It sits headless in a corner and does nothing but work. An old 400MB hard drive (the original 120 went screwy) and 8MB of memory and it's pretty much at its limit. But it doesn't complain at all.
but this isn't all good, it's a question of quantity vs. quality:P obviously. but i thought it needed to be said... some people don't realize, but if you listen... at least, i can't help but hear the difference. mp3s soudn awful to me.
Where I usually hear the difference between CD and MP3 is when there are a lot of cymballs, tambourine or hi hats and not much else. The encoders just can't seem to get the high frequency percussion right without using a lot of bits in the encode.
That being said, I was absolutely amazed when I heard an MP3 of Melissa Etheridge's "Like the Way I Do". I heard the MP3, threw in my CD and listened. I honestly cannot tell the difference in the opening bars at all!
For those who don't know the song, it starts out with a cymbal and hi hat laying down the rhythm and a tambourine "filling in". That's it. Nothing but highs and "noise" which would normally be completely mangled by most encoders.
I want to do some more testing when I have some free time, because I can't remember if it's an MP3 I ripped or if I had downloaded it. I usually use LAME and its VBR, although this MP3 is a plain jane 128kbit MP3 and it sounds absolutely incredible. If I didn't encode it, I want to know whose encoder it is because that encoder is worth a lot of money in my eyes (ears).
I listen to a lot of MP3s and I can usually tell when it's been encoded with a poor encoder. Most MP3 encoders do a fair job. It's good enough for me in the car or at a party. But if I want to keep the quality I usually set up a 128kbit mininmum with VBR and get acceptable results, most of the time. Whoever did the Etheridge track I mentioned has one KICK ASS encoder.
Thirteen? Is that how old I am? I must be doing something right to have a home, two cars, two kids and a wife and a career to pay for it all considering I'm only 13.
If you don't like what I have to say, ignore it. Everyone is entitled to my opinion.
There's an obligatory Richard Gere joke somewhere in this....
A relative of mine is a 17+ year Toronto General Hospital nurse. Many moons ago when she was in emergency she swears to god she pulled hamsters out of his arse. And this woman never ever lies. Ever.
She didn't say it was him for sure but she is willing to put money on it. She's got quite a number of interesting stories about her emergency room days, in fact.
Of course, one coudl always revert to a standard news reader, get onto usenet, and forgo the middleman altogether, couldn't they???
The thing I use dejanews most for is its ability to do a search of all their carried newsgroups for terms I want and give me only the results I want. I dont' want to have to have the bandwidth required to do that on my own machine.
Are there any services (or newsreader software) which allow me to ask a server to search and give me the newsgroups and articles which contain the terms I asked for?
Resistor color codes - you know Victory Garden Walls - are just unfathomable to me.
Victory Garden Walls??! Sounds like you had a sissy shop teacher.:-)
The correct colour code memory aid goes thusly:
Bad Boys Rape Our Young Girls But Violet Goes Willingly
My first year electronics teacher (I took grade 11 electronics in grade 9) taught us that way, and I've heard it from almost every old fart electronics guy (teacher or no). The new ones teach you something like "Better Be Right Or Your Whole Big Venture Goes Wrong" or some such namby-pamby thing.:-)
Those are the one working on the module. Don't want it, don't compile it in. There are very few interdependency bugs, and that's why mozilla kicks all the other browsers asses.
How the hell can it kick "the other browser's asses" when it doesn't work?!?!!
Java/JScript often crashes or just doesn't work
The scrollbars always work, but they don't always scroll the screen
AutoComplete doesn't
The browser tends to crash if I look at it funny
The widgets are often slow to respond
The throbber doesn't work right
That's just a short list of problems under Linux. You've got to be smoking some serious crack if you think Moz kicks any other browser's ass right now. About the only part of Mozilla that does rock is the renderer. Too bad it's buried in 18 megs of shit and the access points are on the other side of broken bridges.
C is an unsafe language (i.e. bugs in one part of a program can cause unforseen errors in totally unrelated parts of the same or different programs).
If you scribble memory, yes you'll have problems. Is C bad because it allows you to scribble memory? Not in my eyes (with power comes danger) but the jury is still out.
Maybe I'm a C zealot then, or maybe I've become cynical enough to realize that computers do exactly what you tell them.:-)
Zero is a number. That was the hardest lesson I've ever learned with computers. Zero Is A Number; All Things Start With Zero.
I assume you know why that code doesn't work, but it all comes down to that premise. You told malloc() to give you enough mem to hold two ints. Then you only use 1/2 your allocated memory and write to memory beyond what you allocated because you forgot that All Things Start At Zero.
That error will bite you in the ass in Assembly and perl too, will it not? You don't need to C to have logical errors. C may not be "safe" but I'd rather have a language which allows you to shoot your head off than one which won't let you try.
Mozilla DOES give you the option to install or not install everything except the browser.
That's not the point; The browser is still buggy as hell because they haven't finished it because they've been pissing away all their time with the other components.
When you write code, do you build each module and test it, or do you build all the modules, throw them together and try to debug the whole thing? One way is much easier, and inter-module problems can be solved much faster if you know how each module responds on its own.
Wow, you have absolutly no clue, did you ever try NSbeta1??
Netscape 1.0 or Netscape 6 beta 1?
I can't download just a browser from Netscape. The last browser-only release was 4.01 IIRC.
Netscape 6.0 Beta 1 I have not tried. However I have tried several milestone relases of Mozilla. You can't get away from the kitchen sink syndrome with Moz.
For example, I use ICQ, telnet, IRC, FTP, and email. Why would anyone expect all of their standard daily tools be built into one, huge, bloated, takes-forever-to-load-because-of-all-this-crap "internet" tool?
Mirabilis does. Have you taken a look at their ICQ2000a client? Looks like they're making room to give you ads too. Too bad KICQ doesn't appear to be actively developed anymore, it was one of the best.
And, as a side note, why do so many people think that mail/news are usuless. When I'm on a new system, or borrowing someone's laptop, mail becomes extremely useful, even with the HTML-based access to my e-mail that my ISP, Earthlink/Mindspring provides. Again, I have to ask the question, "who are you people to decide what's useful for a browser client?" You are not the only bloody users, are you? There are a large number of people who do use the mail/news client from the 4.x series.
Right back at you
Who are YOU to decide that everyone NEEDS the email/news client? Give me a plain browser and the option to download the mail/news client (and IRC client, and XUL templating, and the MozillaOS extensions, and that kitchen sink module while you're at it) and I'll (as would be 99% of the people bitching about Mozilla) be happier than you can imagine.
But no... Mozilla developers figure that since it's all cool to have this stuff, they'll make everyone have it. That is where us people who bitch about not having a plain browser are coming from.
BTW. I am using todays daily build with 5 browser windows open. For the past week that I have been using the daily builds they haven't crashed once. And they do a better job of general browsing than Netscape does!
Ask my friend's wife what she thinks about Moz on Linux. She crashes it almost hourly with just plain-jane web browsing.
She visits a lot of Java/JavaScript sites and Moz blows up bigtime on a lot of them. She doesn't use the email/news/etc. parts, just the browser and she has a worse time with Moz than she does with Netscape.
She went to Moz because of Netscape's problems. You're telling me that she wants a browser that has everything, including the kitchen sink, fifth wheel and ottoman? Give me a break. Mozilla started out as a browser. Now it's an OS with a semi-functional browser.
I believe we've tried Galleon, Opera, etc. Yuck. Or doesn't support the basic Java/JScript. Linux needs a decent browser. Moz was supposed to be that.
How's that phrase go? Get it working, then add on. Not the other way around. They have gecko. All they had to do was put in a basic Java/JScript component and a few widgets to operate the browser and 99% of people would be happy. Don't give me the bullshit statement that they are taking the long hard road to make it better. It's already componentized; it does not take much to take the individual components and encapsulate them with XUL or whatever the fuck it is they're doing now. The could have released Moz months ago in a beautiful fast browser and then started the XUL/MozillaOS extensions to make everyone happy. They didn't, and that's why I agree with suck.
Moz is dead. Long live whoever's next. Hopefully those who put so much time and effort into Moz have learned something and if they get another opportunity for such a high-profile project they won't drop the ball again.
Almost every 386+ OS has not used segments the way Intel intended. So yes, they've had quite a few years (more than a decade) to add an execute bit, if they actually cared.
Intel processors do have an execute bit for each selector. Or did you eman something else?
Nothing nintendo can do about it. They can't force people into a license if you're not using anything of nintendo's actually make the part/addon/rom. You can go out, buy a gameboy, use it to develop your stuff on, and thats it. Nintendo has no authority to make you pay licensing fees, except to use their trademarks and "IP" which you wouldn't need if you did it all on your own.
Actually this is untrue.
The Gameboy processor runs a little routine in ROM which scrolls down the "Nintendo" logo. The bit pattern for the logo is in your ROM. After scrolling it, it then checks the bit pattern against a ROM copy and if they match, will go to execute your code. It is very easy to change the logo to whatever you want, but the game won't run. I guess their arguement is that the bit pattern to display "Nintendo" is copywright/trademarked and will get you on that front.
It is possible, through the use of a microcontroller or little bit of hardware (cap, resistor and maybe a transistor or gate) to display a different logo and still have the game run. This is done by having one area of the ROM appear at the proper location on powerup, but shortly thereafter (after the scroll), flip over to the "Nintendo" bit pattern so the pattern matches that in ROM and the game runs. I've done it myself and it's not hard. How much this protects you legally is another issue, however.
I know Game Genie gets around the licensing by making you have a cartridge. I believe this use of a bit pattern has been tested in court, but I don't know what the outcome was.
The obvious solution is to have a huge store of these images ready to use somewhere, so the web server just has to choose a pre-rendered image.
You don't need a lot of numbers, just say maybe 15 different ones. Make sure all the filesizes are the same (use a noncompressed format maybe?) and either link it to a completely random filename which gets used in the HTML and removed after the page is sent.
It's only "Silly" until your UPS dies (or the card fails or your SCSI bus resets) while there are cached writes.
I would be under the impression that if your UPS going to die it would let you know through the power control protocols. Unless you mean if the UPS explodes unexpectedly, in which case I thought the battery kept cache data, not state data, which was the reason for my "silly" comment. If it keeps state data (a transaction log if you will) then I'm all for it.:-)
Cache will alleviate the performance problem for brief, small transactions.
Which was exactly the context in which I was speaking. The parent to my reply had stated that the bulk of DB transactions were small and that the multiple-write nature of RAID5 made it a performance bottleneck. I had said that a large write cache would allieviate that.
If you're moving more than 256MB through the controller (in either direction, remember that reads consume that cache, too) in less time than the disks can service it, then your I/O's become as slow as the disks. This is unavoidable and unfixable.
Agreed. But then you're back to square one anyway, with the system (usually) being faster than the bulk storage, which is why you have a small but fast disk cache, a slower but bigger controller cache, and a slower yet but bigger filesystem cache on the OS. Each time you step back from the hardware you get a larger cache. System memory is slower than the fast SRAM on the disk cache, but if the memory has it it's a ton faster than actually waiting to get the drive to give you the data (and waiting to get it over a 16/32-bit bus
RAID5 is best-suited for read-intensive environments, or cost-sensitive customers. It is not a high-performance solution. As others have said, RAID0+1 (striped mirrors) are the answer if you want fast and safe instead of cheap and safe.
I'll state again that it depends on your situation. No need to spend a pile on 30G SCSI-II UW disks for a database when you're doing many small transactions. Better to get a few smaller SCSI-II UW disks and RAID-5 with a large cache. There's the ultimate, then there's the practical.:-) The lines between which depend on the pocketbook and the application.
RAID 5 is real slow for small writes common with a database. You have to first read the whole stripesize (much bigger than the oftentimes single block to write) from all disks, calculate parity and write the small changes back (data + parity). What you want is mirroring, RAID 1, which won't decrease write performance to a crawl but keep your data safe.
Personally I don't like having two very large disks around. Give me a half dozen or so smaller ones.
Also, Most hardware RAID controllers have a decent amount of cache with them. The DPT controllers I use can have up to (I think) 256M of ECC cache RAM and optionally battery back it up (silly IMO). That'll fix your performance issues on RAID5.
Slashdot Mirror
It's just another example of security through obscurity, that's what's wrong with it. IP-based authentication is worthless.
I've said it before and I'll say it again and again. Security through obscurity is a perfectly fine LAYER to add to a security regimen.
Sure you've got everything tuned up pretty good but there is nothing wrong and I strongly encourage the use of an obscurity layer in order to build up defences a little more. Relying on obscurity to protect you is one thing, and it's a very bad thing. But having the attacker have to guess what the hell he's looking at before he can apply his standard toolkits and procedures is always a good thing. Always. You don't have a little sticker on your house door which says "Dudley EX-145 model lock" now do you?
Now bringing this back to the topic at hand... If I have a trigger port that will open up all my services (or a selected service) when tickled just right is a fine way to keep the scanners at bay. Even better is if you put that trigger port on auth and your auth demon is tied in with your inetd server: If you get an auth request for port 12345 followed by a request for port 54321 within a 5 second window, it tells inetd to start listening on whichever ports you specify for the next 1 minute (or something).
Tell me, what's so insecure about that, if you've got all your other defences built up?
Its /possible/ for someone at the other end of the wire to send packets to you with a destination-IP of 127.0.0.1, and your box will happily accept them. Sure, this is a long-shot, and there's no way the hacker at the central-office will get a response, but there are a lot of attacks that dont need responses.
That's why my IPCHAINS input chain looks a little like this:
I've played a bit with deny policies on the input chain but they're really too restrictive and your chain starts to get a little long, especially if you're doing more than just telnet/ssh/ftp/web. or have more than just two interfaces (think incoming dialup, DSL, local network and VPN) The forward policy, however, is always deny. Always.
I wonder how long it will take until cable companies start to hire outsiders to scan their networks...
What's so hard about using a trigger port/ports to open everything else up?
Given that a regular modem involves a CPU response to almost every single character (a DSL interface won't require that)
I believe that some cheap-ass NICs are almost as bad. 3COM's Parallel tasking chipset (in the 3C905B) is very good about not using your CPU to bring in data.
That's right, a Pentium 90. Not as bad as a 486, but no great shakes.
A pox on both your P90 and the other guy's 486!
I'm doing the masq/firewall mambo with a 12 (10?) year old 80386DX/40. This poor machine was purchased new by my father for a small fortune and a few years ago I claimed it.
It sits headless in a corner and does nothing but work. An old 400MB hard drive (the original 120 went screwy) and 8MB of memory and it's pretty much at its limit. But it doesn't complain at all.
but this isn't all good, it's a question of quantity vs. quality :P obviously. but i thought it needed to be said... some people don't realize, but if you listen... at least, i can't help but hear the difference. mp3s soudn awful to me.
Where I usually hear the difference between CD and MP3 is when there are a lot of cymballs, tambourine or hi hats and not much else. The encoders just can't seem to get the high frequency percussion right without using a lot of bits in the encode.
That being said, I was absolutely amazed when I heard an MP3 of Melissa Etheridge's "Like the Way I Do". I heard the MP3, threw in my CD and listened. I honestly cannot tell the difference in the opening bars at all!
For those who don't know the song, it starts out with a cymbal and hi hat laying down the rhythm and a tambourine "filling in". That's it. Nothing but highs and "noise" which would normally be completely mangled by most encoders.
I want to do some more testing when I have some free time, because I can't remember if it's an MP3 I ripped or if I had downloaded it. I usually use LAME and its VBR, although this MP3 is a plain jane 128kbit MP3 and it sounds absolutely incredible. If I didn't encode it, I want to know whose encoder it is because that encoder is worth a lot of money in my eyes (ears).
I listen to a lot of MP3s and I can usually tell when it's been encoded with a poor encoder. Most MP3 encoders do a fair job. It's good enough for me in the car or at a party. But if I want to keep the quality I usually set up a 128kbit mininmum with VBR and get acceptable results, most of the time. Whoever did the Etheridge track I mentioned has one KICK ASS encoder.
Thirteen? Is that how old I am? I must be doing something right to have a home, two cars, two kids and a wife and a career to pay for it all considering I'm only 13.
If you don't like what I have to say, ignore it. Everyone is entitled to my opinion.
There's an obligatory Richard Gere joke somewhere in this....
A relative of mine is a 17+ year Toronto General Hospital nurse. Many moons ago when she was in emergency she swears to god she pulled hamsters out of his arse. And this woman never ever lies. Ever.
She didn't say it was him for sure but she is willing to put money on it. She's got quite a number of interesting stories about her emergency room days, in fact.
Of course, one coudl always revert to a standard news reader, get onto usenet, and forgo the middleman altogether, couldn't they???
The thing I use dejanews most for is its ability to do a search of all their carried newsgroups for terms I want and give me only the results I want. I dont' want to have to have the bandwidth required to do that on my own machine.
Are there any services (or newsreader software) which allow me to ask a server to search and give me the newsgroups and articles which contain the terms I asked for?
My favorite is the tic tac toe game that is both the game and the code! (recompile to play next move...)
Hmmm, I can't seem to compile it. Lots of errors about undeclared variables. :-(
Resistor color codes - you know Victory Garden Walls - are just unfathomable to me.
Victory Garden Walls??! Sounds like you had a sissy shop teacher. :-)
The correct colour code memory aid goes thusly:
Bad Boys Rape Our Young Girls But Violet Goes Willingly
My first year electronics teacher (I took grade 11 electronics in grade 9) taught us that way, and I've heard it from almost every old fart electronics guy (teacher or no). The new ones teach you something like "Better Be Right Or Your Whole Big Venture Goes Wrong" or some such namby-pamby thing. :-)
Those are the one working on the module. Don't want it, don't compile it in. There are very few interdependency bugs, and that's why mozilla kicks all the other browsers asses.
How the hell can it kick "the other browser's asses" when it doesn't work?!?!!
That's just a short list of problems under Linux. You've got to be smoking some serious crack if you think Moz kicks any other browser's ass right now. About the only part of Mozilla that does rock is the renderer. Too bad it's buried in 18 megs of shit and the access points are on the other side of broken bridges.
C is an unsafe language (i.e. bugs in one part of a program can cause unforseen errors in totally unrelated parts of the same or different programs).
If you scribble memory, yes you'll have problems. Is C bad because it allows you to scribble memory? Not in my eyes (with power comes danger) but the jury is still out.
int *iptr = malloc(sizeof(int) * 2), i;
for(i = 2; i>0 ; i--)
iptr[i] = i;
Maybe I'm a C zealot then, or maybe I've become cynical enough to realize that computers do exactly what you tell them. :-)
Zero is a number. That was the hardest lesson I've ever learned with computers. Zero Is A Number; All Things Start With Zero.
I assume you know why that code doesn't work, but it all comes down to that premise. You told malloc() to give you enough mem to hold two ints. Then you only use 1/2 your allocated memory and write to memory beyond what you allocated because you forgot that All Things Start At Zero.
That error will bite you in the ass in Assembly and perl too, will it not? You don't need to C to have logical errors. C may not be "safe" but I'd rather have a language which allows you to shoot your head off than one which won't let you try.
Mozilla DOES give you the option to install or not install everything except the browser.
That's not the point; The browser is still buggy as hell because they haven't finished it because they've been pissing away all their time with the other components.
When you write code, do you build each module and test it, or do you build all the modules, throw them together and try to debug the whole thing? One way is much easier, and inter-module problems can be solved much faster if you know how each module responds on its own.
Wow, you have absolutly no clue, did you ever try NSbeta1??
Netscape 1.0 or Netscape 6 beta 1?
I can't download just a browser from Netscape. The last browser-only release was 4.01 IIRC.
Netscape 6.0 Beta 1 I have not tried. However I have tried several milestone relases of Mozilla. You can't get away from the kitchen sink syndrome with Moz.
WHY C SUCKS
-----------
int i =0;
i = i + 1;
Segmentation Fault(Core Dumped)
Care to elaborate?
For example, I use ICQ, telnet, IRC, FTP, and email. Why would anyone expect all of their standard daily tools be built into one, huge, bloated, takes-forever-to-load-because-of-all-this-crap "internet" tool?
Mirabilis does. Have you taken a look at their ICQ2000a client? Looks like they're making room to give you ads too. Too bad KICQ doesn't appear to be actively developed anymore, it was one of the best.
And, as a side note, why do so many people think that mail/news are usuless. When I'm on a new system, or borrowing someone's laptop, mail becomes extremely useful, even with the HTML-based access to my e-mail that my ISP, Earthlink/Mindspring provides. Again, I have to ask the question, "who are you people to decide what's useful for a browser client?" You are not the only bloody users, are you? There are a large number of people who do use the mail/news client from the 4.x series.
Right back at you
Who are YOU to decide that everyone NEEDS the email/news client? Give me a plain browser and the option to download the mail/news client (and IRC client, and XUL templating, and the MozillaOS extensions, and that kitchen sink module while you're at it) and I'll (as would be 99% of the people bitching about Mozilla) be happier than you can imagine.
But no... Mozilla developers figure that since it's all cool to have this stuff, they'll make everyone have it. That is where us people who bitch about not having a plain browser are coming from.
BTW. I am using todays daily build with 5 browser windows open. For the past week that I have been using the daily builds they haven't crashed once. And they do a better job of general browsing than Netscape does!
Ask my friend's wife what she thinks about Moz on Linux. She crashes it almost hourly with just plain-jane web browsing.
She visits a lot of Java/JavaScript sites and Moz blows up bigtime on a lot of them. She doesn't use the email/news/etc. parts, just the browser and she has a worse time with Moz than she does with Netscape.
She went to Moz because of Netscape's problems. You're telling me that she wants a browser that has everything, including the kitchen sink, fifth wheel and ottoman? Give me a break. Mozilla started out as a browser. Now it's an OS with a semi-functional browser.
I believe we've tried Galleon, Opera, etc. Yuck. Or doesn't support the basic Java/JScript. Linux needs a decent browser. Moz was supposed to be that.
How's that phrase go? Get it working, then add on. Not the other way around. They have gecko. All they had to do was put in a basic Java/JScript component and a few widgets to operate the browser and 99% of people would be happy. Don't give me the bullshit statement that they are taking the long hard road to make it better. It's already componentized; it does not take much to take the individual components and encapsulate them with XUL or whatever the fuck it is they're doing now. The could have released Moz months ago in a beautiful fast browser and then started the XUL/MozillaOS extensions to make everyone happy. They didn't, and that's why I agree with suck.
Moz is dead. Long live whoever's next. Hopefully those who put so much time and effort into Moz have learned something and if they get another opportunity for such a high-profile project they won't drop the ball again.
Almost every 386+ OS has not used segments the way Intel intended. So yes, they've had quite a few years (more than a decade) to add an execute bit, if they actually cared.
Intel processors do have an execute bit for each selector. Or did you eman something else?
Nothing nintendo can do about it. They can't force people into a license if you're not using anything of nintendo's actually make the part/addon/rom. You can go out, buy a gameboy, use it to develop your stuff on, and thats it. Nintendo has no authority to make you pay licensing fees, except to use their trademarks and "IP" which you wouldn't need if you did it all on your own.
Actually this is untrue.
The Gameboy processor runs a little routine in ROM which scrolls down the "Nintendo" logo. The bit pattern for the logo is in your ROM. After scrolling it, it then checks the bit pattern against a ROM copy and if they match, will go to execute your code. It is very easy to change the logo to whatever you want, but the game won't run. I guess their arguement is that the bit pattern to display "Nintendo" is copywright/trademarked and will get you on that front.
It is possible, through the use of a microcontroller or little bit of hardware (cap, resistor and maybe a transistor or gate) to display a different logo and still have the game run. This is done by having one area of the ROM appear at the proper location on powerup, but shortly thereafter (after the scroll), flip over to the "Nintendo" bit pattern so the pattern matches that in ROM and the game runs. I've done it myself and it's not hard. How much this protects you legally is another issue, however.
I know Game Genie gets around the licensing by making you have a cartridge. I believe this use of a bit pattern has been tested in court, but I don't know what the outcome was.
The obvious solution is to have a huge store of these images ready to use somewhere, so the web server just has to choose a pre-rendered image.
You don't need a lot of numbers, just say maybe 15 different ones. Make sure all the filesizes are the same (use a noncompressed format maybe?) and either link it to a completely random filename which gets used in the HTML and removed after the page is sent.
It's only "Silly" until your UPS dies (or the card fails or your SCSI bus resets) while there are cached writes.
I would be under the impression that if your UPS going to die it would let you know through the power control protocols. Unless you mean if the UPS explodes unexpectedly, in which case I thought the battery kept cache data, not state data, which was the reason for my "silly" comment. If it keeps state data (a transaction log if you will) then I'm all for it. :-)
Cache will alleviate the performance problem for brief, small transactions.
Which was exactly the context in which I was speaking. The parent to my reply had stated that the bulk of DB transactions were small and that the multiple-write nature of RAID5 made it a performance bottleneck. I had said that a large write cache would allieviate that.
If you're moving more than 256MB through the controller (in either direction, remember that reads consume that cache, too) in less time than the disks can service it, then your I/O's become as slow as the disks. This is unavoidable and unfixable.
Agreed. But then you're back to square one anyway, with the system (usually) being faster than the bulk storage, which is why you have a small but fast disk cache, a slower but bigger controller cache, and a slower yet but bigger filesystem cache on the OS. Each time you step back from the hardware you get a larger cache. System memory is slower than the fast SRAM on the disk cache, but if the memory has it it's a ton faster than actually waiting to get the drive to give you the data (and waiting to get it over a 16/32-bit bus
RAID5 is best-suited for read-intensive environments, or cost-sensitive customers. It is not a high-performance solution. As others have said, RAID0+1 (striped mirrors) are the answer if you want fast and safe instead of cheap and safe.
I'll state again that it depends on your situation. No need to spend a pile on 30G SCSI-II UW disks for a database when you're doing many small transactions. Better to get a few smaller SCSI-II UW disks and RAID-5 with a large cache. There's the ultimate, then there's the practical. :-) The lines between which depend on the pocketbook and the application.
RAID 5 is real slow for small writes common with a database. You have to first read the whole stripesize (much bigger than the oftentimes single block to write) from all disks, calculate parity and write the small changes back (data + parity). What you want is mirroring, RAID 1, which won't decrease write performance to a crawl but keep your data safe.
Personally I don't like having two very large disks around. Give me a half dozen or so smaller ones.
Also, Most hardware RAID controllers have a decent amount of cache with them. The DPT controllers I use can have up to (I think) 256M of ECC cache RAM and optionally battery back it up (silly IMO). That'll fix your performance issues on RAID5.
I think that RAID5 is a good idea, but YMMV.