You know, that reminds me of when I went to work at a computer camp one summer ('00) during college. When I went there, I brought my recently aquired "purple" computer. (yeah, it looked cool, and had cool-looking screensavers everyone noticed)
When looking at it, one of the counselors (ok, he was the lazy guy who ran the R/C cars stuff) asked if it was running '98 (as in Windows '98):-)
Of course it wasn't. It was an SGI Indigo2 running IRIX 6.5, with 4Dwm as the X window manager.
I'd love to have some average person ask me about my home "computer" (probably referring to my main desktop, as I have several systems) sometime these days, if only to confuse them. At the moment it is a Sun box running Solaris 9. (and tech people assume Linux, when they see KDE and all that other OSS stuff running on the screen, hehe)
Yeah, I really wish that wasn't such an uncommon feature on x86 machines. I've net-booted Sun (SPARC and 68k), IBM RS/6000, SGI, VAX, and who knows what other kinds of machines before, and not needing any removable-media devices for an OS install does kinda spoil you.
If anyone is interested in learning about how the netboot process actually works, I highly recommend the
NetBSD Diskless HOWTO. It gives a good walkthrough for just about any platform, and you can apply most of the knowledge to netbooting any OS.
So what exactly is stopping you from porting/writing drives for Solaris now? Sun has a full development kit and developer documentation freely available on your website. One merely has to want to code the things, and it can happen. Heck, 3rd party opensource Solaris drives do exist, but they're just not that common.
However, I'm not sure how easy it would be to port Linux drivers to Solaris.
Though I have looked at both FreeBSD and Solaris driver code side-by-side, and I must say that they appear quite structurally similar in design. (Different names for the kernel call functions, but they tend to do the same things)
Yeah, and I'm posting this from a Solaris 9 machine running KDE.
Honestly, one thing that bugs me is how so many people think of all this stuff as "Linux" software, when it'll compile and run just perfectly fine on any *nix. Linux is just an open-source POSIX-style operating system, and it certanly isn't the only game in town.
I find it kinda interesting how I basically use all the same software your average Linux user runs, and use a multitude of non-Windows operating systems, but I don't actually use Linux myself.
Actually, Solaris has had support for this feature for some time, and OpenBSD now has it as well. It works by making the stack marked as non-executable memory, and stops normal buffer overflow attacks dead in their tracks.
However, a system with this is still exploitable. It is just much harder. One writeup on this can be found here.
Reminds me when I was on a summer internship in San Diego, working about 10 miles north of MCAS Miramar. I would hear F/A-18's flying overhead on a daily basis. Too bad I never got any pictures of those, because they had always passed by the time I got my camera out. They also flew choppers, but those were nowhere near as loud, but much easier to snap a photo of.
Exactly... The Sun philosophy towards "openness" is open standards, not open implementations.
Now frankly, open standards are far more important, because it allows anyone to interoperate with their products.
What's a bigger proiblem with MS? Closed-source or closed-standards? If their file formats were open standards, do you think we'd have so many problems with MS Office interoperability with our non MS systems?
What amazes me is just how often people take their professors to be experts, even when they get it wrong. I've run into a few situations where someone I know learned something from a prof, and I found myself in an argument trying to convince them that they heard wrong. These were technical situations, so I might as well give some examples:
One person I was talking to was convinced that the different classes of IP addresses (A, B, C, D, etc.) defined what types of organizations they were assigned to. (as opposed to the first few bits of the address, and the size of the network blocks)
Another person was convinced that the character generator on a PC was in the monitor, not the video card.
Actually, these days those games need one more huge thing... A high-resolution display, where you can see the little details as well as the big things. TVs just aren't designed for this.
Agreed... Like water, or even gasoline (yes, all automotive gas IS a generic product, and is all transported over a common shared pipeline system. only difference is additives at the truck-fill station, and most of those are similar too), there is no difference between power from one company or another. The only differentiation infrastructure would be financial, where the metering of power generation, usage, and billing creates the illusion of being able to choose your power company.
Honestly, there is nothing special about those machines, beyond that they are just weird PC boxes.
However, it would be nice if open-source OSes did support SGI's real MIPS-based machines a bit better. All the ports I've seen so far do not provide very good support for the hardware beyond the basics. And frankly, if they can't support the graphics hardware, what is even the point of running something other than IRIX?
Another platform I'd like to see support for would be the older microchannel-based non-CHRP IBM RS/6000 machines. They are frankly quite nice boxes, but the only OS they'll run is AIX. (which isn't bad, but tends to confuse hobbyists who aren't multi-platform savvy)
Actually, I was sorely disappointed with my only recent OpenBSD/sparc64 experience. However, I will say that OpenBSD/sparc runs beautifully.
I had been using an old SPARCstation 5 for something, under OpenBSD, and wanted to upgrade to a more powerful (and rackmount) machine. So I got myself a nice Netra t1 105.
Initially I tried OpenBSD/sparc64 3.3 on it, of course. Well, getting it to work with a mirrored root disk was very painful, and resulted in 30+ minute boot times whenever there was an impropper shutdown. Beyond that, it did work, though the machine was performing much more slowly than I had expected. Furthermore, I had to do some digging and manually apply a kernel patch so that one of the network interfaces wouldn't randomly crap out every once in a while.
Then, given that this box was doing some network stuff, I wanted more ethernet ports beyond the two "hme" controllers built-in. So I picked up a QFE (4-port hme 10/100 NIC) and stuck it in the PCI slot. OpenBSD totally hated that, and I never had any luck getting it to work. Sure it detected the card, but forget about forwarding packets. A friend of mine was using a similar card in his box, though he only had one on-board hme. Only one port on the card worked for him. So I'll guess OpenBSD doesn't like more than 2 hme controllers on a machine.
Eventually I had enough, and just switched the machine over to Solaris 9. I found some good guides for locking it down, and all the software I needed would run on that OS. So after a little bit of effort, I managed to convert the box. It now runs perfectly, all my ethernet ports work, the on-board watchdog timer is supported, and software mirroring of the root disk works painlessly. Oh, and the machine is also performing MUCH faster now.
At this point, I will still say OpenBSD/sparc works wonderfully on the older 32-bit SPARC boxes, but I've lost all faith in the sparc64 port. I'm still eager to see how FreeBSD/sparc64 turns out, though, but right now Solaris just "works better" for all the UltraSPARC machines (and their associated hardware) I have and use.
The U5 and the U10 are the same exact machine. The only difference is the case, PCI riser card, and the CPU speeds they shipped with. The motherboard is identical.
First of all, the OSI model needs to be taken with a grain of salt. It does not directly map to the way things really work. Furthermore, there is flexibility allowed for layers of indirection.
However, from the host's view down, TCP, UDP, or BIC-TCP is layer 4, IP is layer 3, and all that DSL/ATM/etc. stuff fits under the guise of Layer 2 and below. (of course with layers of indirection, you can simulate the ATM cloud over IP/w MPLS as well, but that is all transparent beyond the provider network)
Actually it isn't cheaper than SCSI (well, unless you're comparing to over-bloated VAR pricing). It's basically the same price as a 36GB 10K SCSI drive. So why even bother with the IDE interface? Get the real thing!
You know, everyone likes to say how bad it was in high school. But actually, I didn't think it was the case. Everything I've seen leads me to believe that middle school is actually the worst time of all. By high school, it felt like everyone had pretty much "grown out of it" and really just left you alone if you didn't bother them.
Yeah, we all like to tell outselves this, but I'm not sure it is quite as true anymore as it used to be. Back in high school, I noticed the emergence of what could possibly be called "smart jocks." They were into school athletics and social circles, yet also were in the more advanced classes and had good grades. These people are something we should all be afraid of, because the only reason the world puts up with our "eccentricity" is because of what else we can offer. Once the "normal" can offer the same, what good are we anymore?
(sure, our technical creativity, versus just academic ability, is still a unique skill, but it really isn't one outwardly noticed by the rest of the world as much)
I was once in a class where a foreign TA (are domestics even allowed?) was giving the lecture that day, and he referred to something called "ell-hospital's rule".;-)
Thankfully, I knew it was really pronounced "low-pital's rule", like most students in the class, and got a little chuckle about it.
Furthermore, the UltraSPARC has this nice feature you can enable where the stack space is non-executable memory. (a feature easily enabled in Solaris, and now OpenBSD as well) While it is still possible to exploit a buffer overflow with this feature, it us MUCH more difficult (google around, and you may find some writeups)
It may seem cool at first, but after a while it gets really boring. You have no sense of accomplishment, and just spend your whole day figuring out how you're gonna kill the time until it is over.
I'm there right now, and I'm jumping after any opportunity I can find for any change that might bring me closer to doing something useful. At least I'm starting grad school via distance-learning, so it gives me something to do some of the time.
Well, any basic standardized and commoditized integrated I/O is a good thing. I'd also be fine with on-board Ethernet (good chipset, of course) and on-board SCSI/FC/SATA/etc. But yeah, I'd rather add my own cards for video/sound.
Re:Buy a new fridge, and other suggestions.
on
DIY HVAC
·
· Score: 1
Bah! As computer geeks, we should be willing to just accept the cost of powering our computers as a condition of our lifestyle. After all, its not like we're going to spend that money as readily on fancy cars or anything (which are a lot more expensive).
I did some measurements a few months ago, and my rack of computer equipment has a continuous draw of around 1000W. That's the stuff that runs 24/7, and doesn't include monitors (which I do shut off when I'm not home). And guess what? Soon I'm planning on putting some more power hungry stuff into use, and that number will shoot up quite a bit.
If you really care about power, though, I strongly recommend thin clients. I currently have a bunch of Sun Ray units (yes, they require Solaris/Sparc on the server, but they're just one example of a possible solution). These things use less power than a laptop, make no noise, and basically let you stick a "desktop setup" anywhere you can snake an ethernet connection to. So for the power draw of single decent desktop (I'm using a dual-proc Ultra 60, which has a 350W power supply and is quieter than the PC sitting next to it), you can have as many desktops as you feel able to share CPU power with.
Slashdot Mirror
You know, that reminds me of when I went to work at a computer camp one summer ('00) during college. When I went there, I brought my recently aquired "purple" computer. (yeah, it looked cool, and had cool-looking screensavers everyone noticed)
:-)
When looking at it, one of the counselors (ok, he was the lazy guy who ran the R/C cars stuff) asked if it was running '98 (as in Windows '98)
Of course it wasn't. It was an SGI Indigo2 running IRIX 6.5, with 4Dwm as the X window manager.
I'd love to have some average person ask me about my home "computer" (probably referring to my main desktop, as I have several systems) sometime these days, if only to confuse them. At the moment it is a Sun box running Solaris 9. (and tech people assume Linux, when they see KDE and all that other OSS stuff running on the screen, hehe)
Yeah, I really wish that wasn't such an uncommon feature on x86 machines. I've net-booted Sun (SPARC and 68k), IBM RS/6000, SGI, VAX, and who knows what other kinds of machines before, and not needing any removable-media devices for an OS install does kinda spoil you.
If anyone is interested in learning about how the netboot process actually works, I highly recommend the
NetBSD Diskless HOWTO. It gives a good walkthrough for just about any platform, and you can apply most of the knowledge to netbooting any OS.
So what exactly is stopping you from porting/writing drives for Solaris now? Sun has a full development kit and developer documentation freely available on your website. One merely has to want to code the things, and it can happen. Heck, 3rd party opensource Solaris drives do exist, but they're just not that common.
However, I'm not sure how easy it would be to port Linux drivers to Solaris.
Though I have looked at both FreeBSD and Solaris driver code side-by-side, and I must say that they appear quite structurally similar in design. (Different names for the kernel call functions, but they tend to do the same things)
Yeah, and I'm posting this from a Solaris 9 machine running KDE.
Honestly, one thing that bugs me is how so many people think of all this stuff as "Linux" software, when it'll compile and run just perfectly fine on any *nix. Linux is just an open-source POSIX-style operating system, and it certanly isn't the only game in town.
I find it kinda interesting how I basically use all the same software your average Linux user runs, and use a multitude of non-Windows operating systems, but I don't actually use Linux myself.
Yeah, I started with the steel wire shelves. But eventually I went rackmount anyways. It is just plain cooler looking.
Actually, Solaris has had support for this feature for some time, and OpenBSD now has it as well. It works by making the stack marked as non-executable memory, and stops normal buffer overflow attacks dead in their tracks.
However, a system with this is still exploitable. It is just much harder. One writeup on this can be found
here.
Reminds me when I was on a summer internship in San Diego, working about 10 miles north of MCAS Miramar. I would hear F/A-18's flying overhead on a daily basis. Too bad I never got any pictures of those, because they had always passed by the time I got my camera out. They also flew choppers, but those were nowhere near as loud, but much easier to snap a photo of.
Exactly... The Sun philosophy towards "openness" is open standards, not open implementations.
Now frankly, open standards are far more important, because it allows anyone to interoperate with their products.
What's a bigger proiblem with MS? Closed-source or closed-standards? If their file formats were open standards, do you think we'd have so many problems with MS Office interoperability with our non MS systems?
What amazes me is just how often people take their professors to be experts, even when they get it wrong. I've run into a few situations where someone I know learned something from a prof, and I found myself in an argument trying to convince them that they heard wrong. These were technical situations, so I might as well give some examples:
One person I was talking to was convinced that the different classes of IP addresses (A, B, C, D, etc.) defined what types of organizations they were assigned to. (as opposed to the first few bits of the address, and the size of the network blocks)
Another person was convinced that the character generator on a PC was in the monitor, not the video card.
Actually, these days those games need one more huge thing... A high-resolution display, where you can see the little details as well as the big things. TVs just aren't designed for this.
Agreed... Like water, or even gasoline (yes, all automotive gas IS a generic product, and is all transported over a common shared pipeline system. only difference is additives at the truck-fill station, and most of those are similar too), there is no difference between power from one company or another. The only differentiation infrastructure would be financial, where the metering of power generation, usage, and billing creates the illusion of being able to choose your power company.
Honestly, there is nothing special about those machines, beyond that they are just weird PC boxes.
However, it would be nice if open-source OSes did support SGI's real MIPS-based machines a bit better. All the ports I've seen so far do not provide very good support for the hardware beyond the basics. And frankly, if they can't support the graphics hardware, what is even the point of running something other than IRIX?
Another platform I'd like to see support for would be the older microchannel-based non-CHRP IBM RS/6000 machines. They are frankly quite nice boxes, but the only OS they'll run is AIX. (which isn't bad, but tends to confuse hobbyists who aren't multi-platform savvy)
Actually, I was sorely disappointed with my only recent OpenBSD/sparc64 experience. However, I will say that OpenBSD/sparc runs beautifully.
I had been using an old SPARCstation 5 for something, under OpenBSD, and wanted to upgrade to a more powerful (and rackmount) machine. So I got myself a nice Netra t1 105.
Initially I tried OpenBSD/sparc64 3.3 on it, of course. Well, getting it to work with a mirrored root disk was very painful, and resulted in 30+ minute boot times whenever there was an impropper shutdown. Beyond that, it did work, though the machine was performing much more slowly than I had expected. Furthermore, I had to do some digging and manually apply a kernel patch so that one of the network interfaces wouldn't randomly crap out every once in a while.
Then, given that this box was doing some network stuff, I wanted more ethernet ports beyond the two "hme" controllers built-in. So I picked up a QFE (4-port hme 10/100 NIC) and stuck it in the PCI slot. OpenBSD totally hated that, and I never had any luck getting it to work. Sure it detected the card, but forget about forwarding packets. A friend of mine was using a similar card in his box, though he only had one on-board hme. Only one port on the card worked for him. So I'll guess OpenBSD doesn't like more than 2 hme controllers on a machine.
Eventually I had enough, and just switched the machine over to Solaris 9. I found some good guides for locking it down, and all the software I needed would run on that OS. So after a little bit of effort, I managed to convert the box. It now runs perfectly, all my ethernet ports work, the on-board watchdog timer is supported, and software mirroring of the root disk works painlessly. Oh, and the machine is also performing MUCH faster now.
At this point, I will still say OpenBSD/sparc works wonderfully on the older 32-bit SPARC boxes, but I've lost all faith in the sparc64 port. I'm still eager to see how FreeBSD/sparc64 turns out, though, but right now Solaris just "works better" for all the UltraSPARC machines (and their associated hardware) I have and use.
The U5 and the U10 are the same exact machine. The only difference is the case, PCI riser card, and the CPU speeds they shipped with. The motherboard is identical.
First of all, the OSI model needs to be taken with a grain of salt. It does not directly map to the way things really work. Furthermore, there is flexibility allowed for layers of indirection.
/w MPLS as well, but that is all transparent beyond the provider network)
However, from the host's view down, TCP, UDP, or BIC-TCP is layer 4, IP is layer 3, and all that DSL/ATM/etc. stuff fits under the guise of Layer 2 and below. (of course with layers of indirection, you can simulate the ATM cloud over IP
Actually it isn't cheaper than SCSI (well, unless you're comparing to over-bloated VAR pricing). It's basically the same price as a 36GB 10K SCSI drive. So why even bother with the IDE interface? Get the real thing!
The title is a bit misleading, though. The article is really on the philosophical design differences between Windows NT and the UNIX world.
You know, everyone likes to say how bad it was in high school. But actually, I didn't think it was the case. Everything I've seen leads me to believe that middle school is actually the worst time of all. By high school, it felt like everyone had pretty much "grown out of it" and really just left you alone if you didn't bother them.
Yeah, we all like to tell outselves this, but I'm not sure it is quite as true anymore as it used to be. Back in high school, I noticed the emergence of what could possibly be called "smart jocks." They were into school athletics and social circles, yet also were in the more advanced classes and had good grades. These people are something we should all be afraid of, because the only reason the world puts up with our "eccentricity" is because of what else we can offer. Once the "normal" can offer the same, what good are we anymore?
(sure, our technical creativity, versus just academic ability, is still a unique skill, but it really isn't one outwardly noticed by the rest of the world as much)
I was once in a class where a foreign TA (are domestics even allowed?) was giving the lecture that day, and he referred to something called "ell-hospital's rule". ;-)
Thankfully, I knew it was really pronounced "low-pital's rule", like most students in the class, and got a little chuckle about it.
Furthermore, the UltraSPARC has this nice feature you can enable where the stack space is non-executable memory. (a feature easily enabled in Solaris, and now OpenBSD as well) While it is still possible to exploit a buffer overflow with this feature, it us MUCH more difficult (google around, and you may find some writeups)
It may seem cool at first, but after a while it gets really boring. You have no sense of accomplishment, and just spend your whole day figuring out how you're gonna kill the time until it is over.
I'm there right now, and I'm jumping after any opportunity I can find for any change that might bring me closer to doing something useful. At least I'm starting grad school via distance-learning, so it gives me something to do some of the time.
Well, any basic standardized and commoditized integrated I/O is a good thing. I'd also be fine with on-board Ethernet (good chipset, of course) and on-board SCSI/FC/SATA/etc. But yeah, I'd rather add my own cards for video/sound.
Bah! As computer geeks, we should be willing to just accept the cost of powering our computers as a condition of our lifestyle. After all, its not like we're going to spend that money as readily on fancy cars or anything (which are a lot more expensive).
I did some measurements a few months ago, and my rack of computer equipment has a continuous draw of around 1000W. That's the stuff that runs 24/7, and doesn't include monitors (which I do shut off when I'm not home). And guess what? Soon I'm planning on putting some more power hungry stuff into use, and that number will shoot up quite a bit.
If you really care about power, though, I strongly recommend thin clients. I currently have a bunch of Sun Ray units (yes, they require Solaris/Sparc on the server, but they're just one example of a possible solution). These things use less power than a laptop, make no noise, and basically let you stick a "desktop setup" anywhere you can snake an ethernet connection to. So for the power draw of single decent desktop (I'm using a dual-proc Ultra 60, which has a 350W power supply and is quieter than the PC sitting next to it), you can have as many desktops as you feel able to share CPU power with.
Well, you can buy 64-bit laptops today. They're just not x86 or Mac/PPC based.
t computing.com/
http://www.tadpolecomputer.com/
http://www.nex
(They're mainly UltraSPARC based. I think TadPole used to make PA-RISC laptops as well.)