That bizarre res happens to be the exact res of most ultra-portable systems; 1024x600. I wonder how it looks on our 8.9 inch widescreens. I dont think its going to play so well with a touchscreen, but thats ok.;)
My question is, how can I get that much data onto my laptop to try it out. I dont think there'll be anything which fits in my PCMCIA slot which can framegrab a 720p stream, encode & feed my system a mpeg. Do they even have pci solutions for that? I suppose I could stream from a desktop, but it'd be far less fun.
Rabbit's done all sorts of insane juju with with the old Z8 core. Its a rather impressive little micro. I'd wager they're not the most power efficient in the world, but thats just a guess.
I seem to remember their dev kits beeing a whole helluva lot cheaper. And I also seem to recall every kit coming with a TCP/IP stack. The wonders of "mature" technologies & companies I guess.
Sounds like you're talking about the P.A. Semi chip recently announced. =]
Well, sans GPU. But given the PCI express interface, a custom one off wouldnt be that hard to tack onto the board. Given the perposterous ram bandwidth on the PA Semi chips, a solution like nVidia's TurboCache would work great: just have one unified ram for processor and GPU.
Thats one other thing I dont expect to see integrated any time soon: RAM. As for storage in general, hopefully flash will continue growing in capacity at a vaguely exponential rate. 32gb flash would almost be sufficient to make the hd unnecessary for most systems.
Just remember though, AMD isnt doing this for integration sakes, they're in it for the I/o. Which, I should note, the P.A. Semi people have in spades. Some smart people are calling the new multi-core age "throughput" computing, which is rather apt; we are going to need a whole new architecture to support our processors.
As wondeful as it makes me feel that someone is sponsoring a warm fuzzy feeling technology like solar energy, the importance of these particulars are dwarfed by the greater issue at hand-- that the responsibility of government is to do what markets cannot. I dont know the numbers, but it certainly feels like government research spending has dwindled. $53mil may be spit in the bucket, but its at least something; it's certainly more than any corporation would invest in a 4 year, two month project.
Government needs to get back in the game of building a future actually worth living in.
To achieve desktop penetration, open source needs one thing. A better metaphor. A new paradigm for how data is stored. Something to replace the file system.
We've oozed eye candy and UI friendliness onto the OS for years. Its no longer what applications, what features or even what games we do or do not have. Linux is competitive with everything else out there for the desktop field. But we're not going to win anything by growing our margin. There's too many network externalities in play such that making the linux desktop "better" will bring it mainstream. We need to leapfrog the world if we are ever to achieve a meaningful victory. Factor ten engineering.
We need to go back to the core, to the heart of the OS and consider where we can truly evolve. What is the purpose of an OS?: secure allocation of resources. Currently all of our resources are files, some files are privledged with having a single function call, these being called programs. Others may be privledged as containers, they hold other files. Anything past this is simply a construct of the application, layers built on top of this base. And this construct is imaginary, because the user will still have to interact with the base layer at some point.
I do not think this is enough. It is not enough. Maybe you can achieve this level of dataware through gnome VFS and KDE's KIO (i think that is there VFS?), but how will this interoperate with a client running Windows or Osx? Can we compell all our applications to be built with GTK? No, we need to rethink from the ground up how data is stored on our systems. Our data is active, alive, by itself, growing.
Right now, we have everything they have, and even when we've built ourselves twice as much, we still wont be any different. Linux got where it is by being disruptive. The only way it will ever continue to excel is by causing further disruption, not by playing the same games everyone else plays. You cannot out Microsoft Microsoft. There are two different metaphors that linux users confuse on a religious basis; the desktop and the desktop. One metaphor describes the feel of the operating system, the other describes a form of market penetration. The success of our desktop environment no longer has bearing on desktop penetration.
Part of the thing about sitting in front of a computer is that often it makes you want to eat. You think of food, you think it would taste good, it creates a sense of hunger, even if your body doesnt really need food.
I suggest learning to drink water as an alternative. You need more water anyways, being mentally active, and it fulfils the oral fixation. I suspect most people do not drink enough water, do yourself the favor and try it. Without fail, everyone I've converted to my h2o cause swears by improved mental acuity, response time, and general mood. Water is my wonder drug.
I dont actually believe in diets (general fitness is far better), but I'd wager to say, any diet where you feel like you are running out of energy is just not a good diet.
In a game like Civ, we have over 80 units, all with different movement rates, strengths, special abilities, experience levels, etc. We also have to decide where to place cities, what to build, who to be nice to and who to make war with. We also have to decide what to research, what religion to spread, what Civics to adopt, etc. All in all, I don't expect to see anything close to true human intelligence any time soon, as long as games continue to get more complex.
I'm gonna hatemyself for saying this, but... *cough* moron *cough*.
Computer intelligence may be going no where, but our ability to gather redonculously huge amounts of data and run signal processing on our data to extract trends we still dont understand... that is ever increasing. As games get more complex and more multi-faceted this trend will come to dominate.
Imagine if there were tactical elements to WoW. You couldd have one WHOPPER (eh? eh?) of a learning system watching PvP battles to exact winning tactics and strategies. Winners tended to heal more. Winners waited around a corner. Trends would begin to emerge.
Metcalf's law states the the value of any system is determined by the square of the number of participants. Obviously Mr. Meier still has a classical hollistic model of AI where each civilization is a monolithic AI playing against other self-same AIs. The problem here is that he is playing a finite game. Whoo! You have 80 units! Thats a very big matrix to determine combat advantage Mr. Meiers. But its inherently a finite system, a simple matrix by nature of each AI element being a monolithic entity.
There will never be emerging ai behaviors for your systems. The best you can hope for is players attributing their Sim's catching on fire to spontaneous combustion. AI only has a chance through the network; it is through the massive rise in complexity that computer AI will come to take hold-- when the AI can crunch huge potential data sets and evolve hideous numbers of very very short term permutations to decide upon tactics that computer AI will ever have a chance. Your AI is dumb because you built it sir, it will always have all the limitations it had the day it was coded, because in truth you never let it decide anything, its nothing more than discrete decision making paths. Of course that will never advance. We can only build our decision making models so complex, only to emulate such a degree of humanity. Thats not AI. Thats just a robot.
All real intelligence is derived from the knowledge of the interconnected nature of realities.
Then again, Armored Core is my game of choice. You'd have to know it to know. Sweet sweet customizability. Luv, LordMyren
Silly cod, you missed the most obvious joke in all eternity!
Verisign to control.com domain name until end of time!
It wasnt just the end of the mayan calender. They werent like, well, this should be enough. It was the end of the 13 eon, where time itself ended and perhaps looped back to the beginning. It is the end of time.
There's no need to make an object just to support an object oriented system. I'd posit that such a system could easily be made cross-platform. Shared memory is a nice-ity, but I dont see that a better implementation thereof compels an entirely new OS to be born. Shared mem aint pretty, but its doable.
Furthermore, many GUI systems are capable of using async callbacks for event notification. AKA generating new threads, as you describe. This process itself comes with (very optimizable) overhead though, as you know well.
These aside, you my friend, are spot on. My heartfelt thanks. I thought I was the only one who was crazy.
My main suspect is that they're distracting our developers. all these UI-centric articles taht keep coming out are simply a red herring to make OSS try to out Redmond MS. Somehow we've fallen for the ploy and decided taht linux will only win by being more user friendly.
Fucking bullshit. Linux community need to get off its ass and start innovating again.
Unless the signal can be predicted, which does not seem wholly unlikely. Most background noises are of a fairly constant nature. Record a couple sample sounds and look at the waveforms very close in; a large number of background sounds are repetitive. These are easy to predict and thus could be cancelled, even given the processing delay.
There's going to be some variation. The technique will not be perfect, but it should be able to achieve rather significant cancellation for many noise sources.
Afaik the real bitch is that the soundcards have different skews, system clocks are easily adjustable and de-skewed with working ntp. Getting the sound card not to drift from the computer--- more problematic. also, low latency interrupts are very problematic to generate to start play.
Please discuss though... I'm very interestd in getting synced audio.
Excellent, thanks so very much, the Hot/Cold is exactly what I need to save myself some perposterous high-availability pains.
Will it be possible to find this sort of support in "consumer" RAID gear? I'm looking at the Intel SRCU42X and LSI MegaRAID 320-2e (both $700) and dont know how it would work. The main thing I dont understand is what form the utilities for these kinds of cards will be in... how arrays are setup, how hotswap is performed, &all... I doubt the firmware interfaces are open enough to allow for GPL kernel modules. BIOS isnt an option since they need to be configured and reconfigured on the fly.
Good advice on the local boot drives. Thanks again! Myren
I ask because its either this or using drbd to replicate the entire file system over multi GigE lines while having to use twice the number of hard drives. I'd much prefer to avoid these interconnections altogether and simply have SCSI itself be the common communication bus, at least such that either controller can access the raid array should the other fail. I'm not/totally/ OT. I was looking at 4 gige or a 10gb solutions which would have pounded cpu usage to death... I'd much rather just make sure I can always access the drives.
I'm aware that it's feasible to have multiple SCSI cards in a SCSI chain, is there any way to do this with hardware RAID cards? It would be a great boon to reliability & costs were this feasible for me. Systems fail but RAID-5 and RAID-6 far less so.
I vote for the first person who will vow to stab verisign in the eyeballs^H^H^H^H^H^H^H^Hstop sitefinder. Once or twice a week I get redirected to a Network Solutions sitefinder-a-like.
For the most part the internet is governed by the corps, who basically understand its only useful because it is a free for all. For the most part, this works fine, with occasional DNS hijinx. I'm rather scared of what would happen if a government or governing agent decided it actually wanted some form of control. Kiss those privacy rights good bye.
I'm waiting for the M-5000u to come out before I buy a media player box. Although I suspect the morons will botch what would be an almost perfect media player by omitting wifi. It does have an internal hard drive, and I have enough extra OpenWRT boxes lying around that this isnt a HUGE problem, but its still an awfully glaring omission.
Hope you like the Squeezebox. I'm considering getting one while I wait. -Myren
I'm hoping we see some new alternatives for remote controls crop up. I'm still using two ATI TV Wonder remotes and a pile of USB recievers (all first gen) to control my medianet. The range is beyond pathetic, the control pad is 4-directional (what I wouldnt give for analog), the remote is huge but it does have decent linux support. And its RF. The advantage is that the remotes and dongles were $6 a pop on ebay and hus cheap enough that I can litter the house with dongles for decent coverage.
I'd like to see Zigbee or Bluetooth solutions crop up. But I'll take anything better if its at a good price. People dont want to pay $100 for a damned remote control, there's little reason for it.
Linux or BSD boxes are free on the curb. Either is fine if you dont dick with it constantly, put it in a closet and go. Gigabit is $100 for the switch alone, I havent found any in the trash yet. Radio stations often operate on shoestring budgets. We have the largest record collection on the east coast and the station still operates on pennies.
Besides, there's no reason you need gigabit for such a piddling task. If someone's going to be saturating a 100bt such that even a 1.4mbps stream is underbuffering, there's a good chance of under run on gigabit too.
QoS is the only acceptable solution for a radio station. If you have to find two boxes on the street for absolute redunancy, go for it. Get redunant $5 switches too. While you are at it, the most likely thing to fail will be the hard drive; if you need bullet proof do netboot to hard drive less routers or use one of the distro's on floppy or distro on cd's. There are some good distro's made explicitly for routing for this very purpose.
Really, the big problem with this is that if you are on a switch all packets now have to flow through the single box. In effect, its much like running a full duplex hub from there since all communication is flowing through the single point.
That aside, I'd like to thank you for at least providing a valid argument against my post. The other two people who replied were a bit more challeneged. I like to think I at least made some sense. Aside from DHCP relays I dont think anything I mentioned wasnt straight out of LARTC. And you do have an excellent point that should definately be taken into consideration.
Setup a linux box to do all routing for the small mini network. By giving each system a netmask of 255.255.255.255 you can make all traffic go through the linux router. If you've got static ip's for the machines, use em as normal, otherwise setup the router to do DHCP and use NAT connection sharing. If you want to get fancy you can setup some hijinx with DHCP relay systems. Easy enough. From there you can either mangle the gateway address to point to your router or make the linux box a transperent gateway- very clean, probably the way to go but you still have to mangle whatever netmask the actual DHCP srever is providing. Details details. Basically get all traffic going through the linux box. I suggest just using an internal DHCP network, its the simplest most exandable.
Then just setup traffic policing. Cap all connections to like 50% of the 100bt's TESTED capbilities except for the couple ports doing streaming. Dead fsckign simple QoS stuff, just a Token Bucket Filter. Traffic policing is also known as Ingress queueing, and is made possible by Intermediate queueing devices (IMQ).
Basically the Linux Advanced Routing and Traffic Control HOWTO will become your new bible. It has everything you need.
Now that I think about it, I dont suppose theres any reason you couldnt just do packet queueing/egress filtering too once you get the linux system acting as a router.
I dunno, I really think we need an effective space plane, not just better and better rockets. Something like White Knight to 40,000 feet, accelerated launch to 120,000 feet, start skipping across the atmosphere doing periodic burns when atmosphere is available. Build up speed, get near double digit mach before turning on the boosters.
I guess we've just got to rely on NASA to work on scramjet and hope they get somewhere first.
Dont worry, the capitalist system of intellectual property is trying as hard as it can to legislate and litigate around this potential.
Capitalism has been errected upon efficiencies of industry, not around innovation. And most dangerously, its legislated protection on high for itself. The singularity will be built around rouge elements [F/OSS] doing projects in their basement, around remixed and re-remixed and re-re-remixed technologies that build massive network externalities. The singularity is the rapid cascade of disruptive technologies that rebuilds the entire technological platform we've built underfoot, a complete architectural revamp that encompasses the loosely coupled distributed nature of information and processes.
By its very definition, the technologies which drive the singularity will be so flexible that they will cause clash after clash with existing IP, simply because these new technologies are so malleable, so capable of outperforming the existing although still infringing. The revolution will be litigated-- but the corporations will loose. IP is not a sustainable edge.
As always, http://ccs.mit.edu/21c/21CWP001.html">Two Scenarios for 21st Century Organizations.
Slashdot Mirror
That bizarre res happens to be the exact res of most ultra-portable systems; 1024x600. I wonder how it looks on our 8.9 inch widescreens. I dont think its going to play so well with a touchscreen, but thats ok. ;)
My question is, how can I get that much data onto my laptop to try it out. I dont think there'll be anything which fits in my PCMCIA slot which can framegrab a 720p stream, encode & feed my system a mpeg. Do they even have pci solutions for that? I suppose I could stream from a desktop, but it'd be far less fun.
Myren
Rabbit's done all sorts of insane juju with with the old Z8 core. Its a rather impressive little micro. I'd wager they're not the most power efficient in the world, but thats just a guess.
I seem to remember their dev kits beeing a whole helluva lot cheaper. And I also seem to recall every kit coming with a TCP/IP stack. The wonders of "mature" technologies & companies I guess.
Myren
Sounds like you're talking about the P.A. Semi chip recently announced. =]
Well, sans GPU. But given the PCI express interface, a custom one off wouldnt be that hard to tack onto the board. Given the perposterous ram bandwidth on the PA Semi chips, a solution like nVidia's TurboCache would work great: just have one unified ram for processor and GPU.
Thats one other thing I dont expect to see integrated any time soon: RAM. As for storage in general, hopefully flash will continue growing in capacity at a vaguely exponential rate. 32gb flash would almost be sufficient to make the hd unnecessary for most systems.
Just remember though, AMD isnt doing this for integration sakes, they're in it for the I/o. Which, I should note, the P.A. Semi people have in spades. Some smart people are calling the new multi-core age "throughput" computing, which is rather apt; we are going to need a whole new architecture to support our processors.
As wondeful as it makes me feel that someone is sponsoring a warm fuzzy feeling technology like solar energy, the importance of these particulars are dwarfed by the greater issue at hand-- that the responsibility of government is to do what markets cannot. I dont know the numbers, but it certainly feels like government research spending has dwindled. $53mil may be spit in the bucket, but its at least something; it's certainly more than any corporation would invest in a 4 year, two month project.
Government needs to get back in the game of building a future actually worth living in.
Myren
Fundamental limitations on plasma fusion systems not in thermodynamic equilibrium.
Keep fucking dreaming kids.
To achieve desktop penetration, open source needs one thing. A better metaphor. A new paradigm for how data is stored. Something to replace the file system.
We've oozed eye candy and UI friendliness onto the OS for years. Its no longer what applications, what features or even what games we do or do not have. Linux is competitive with everything else out there for the desktop field. But we're not going to win anything by growing our margin. There's too many network externalities in play such that making the linux desktop "better" will bring it mainstream. We need to leapfrog the world if we are ever to achieve a meaningful victory. Factor ten engineering.
We need to go back to the core, to the heart of the OS and consider where we can truly evolve. What is the purpose of an OS?: secure allocation of resources. Currently all of our resources are files, some files are privledged with having a single function call, these being called programs. Others may be privledged as containers, they hold other files. Anything past this is simply a construct of the application, layers built on top of this base. And this construct is imaginary, because the user will still have to interact with the base layer at some point.
I do not think this is enough. It is not enough. Maybe you can achieve this level of dataware through gnome VFS and KDE's KIO (i think that is there VFS?), but how will this interoperate with a client running Windows or Osx? Can we compell all our applications to be built with GTK? No, we need to rethink from the ground up how data is stored on our systems. Our data is active, alive, by itself, growing.
Right now, we have everything they have, and even when we've built ourselves twice as much, we still wont be any different. Linux got where it is by being disruptive. The only way it will ever continue to excel is by causing further disruption, not by playing the same games everyone else plays. You cannot out Microsoft Microsoft. There are two different metaphors that linux users confuse on a religious basis; the desktop and the desktop. One metaphor describes the feel of the operating system, the other describes a form of market penetration. The success of our desktop environment no longer has bearing on desktop penetration.
Myren
Part of the thing about sitting in front of a computer is that often it makes you want to eat. You think of food, you think it would taste good, it creates a sense of hunger, even if your body doesnt really need food.
I suggest learning to drink water as an alternative. You need more water anyways, being mentally active, and it fulfils the oral fixation. I suspect most people do not drink enough water, do yourself the favor and try it. Without fail, everyone I've converted to my h2o cause swears by improved mental acuity, response time, and general mood. Water is my wonder drug.
I dont actually believe in diets (general fitness is far better), but I'd wager to say, any diet where you feel like you are running out of energy is just not a good diet.
Myren
I never liked tables. Too much impedance mismatch. I would get rid of tables.
In a game like Civ, we have over 80 units, all with different movement rates, strengths, special abilities, experience levels, etc. We also have to decide where to place cities, what to build, who to be nice to and who to make war with. We also have to decide what to research, what religion to spread, what Civics to adopt, etc. All in all, I don't expect to see anything close to true human intelligence any time soon, as long as games continue to get more complex.
I'm gonna hatemyself for saying this, but... *cough* moron *cough*.
Computer intelligence may be going no where, but our ability to gather redonculously huge amounts of data and run signal processing on our data to extract trends we still dont understand... that is ever increasing. As games get more complex and more multi-faceted this trend will come to dominate.
Imagine if there were tactical elements to WoW. You couldd have one WHOPPER (eh? eh?) of a learning system watching PvP battles to exact winning tactics and strategies. Winners tended to heal more. Winners waited around a corner. Trends would begin to emerge.
Metcalf's law states the the value of any system is determined by the square of the number of participants. Obviously Mr. Meier still has a classical hollistic model of AI where each civilization is a monolithic AI playing against other self-same AIs. The problem here is that he is playing a finite game. Whoo! You have 80 units! Thats a very big matrix to determine combat advantage Mr. Meiers. But its inherently a finite system, a simple matrix by nature of each AI element being a monolithic entity.
There will never be emerging ai behaviors for your systems. The best you can hope for is players attributing their Sim's catching on fire to spontaneous combustion. AI only has a chance through the network; it is through the massive rise in complexity that computer AI will come to take hold-- when the AI can crunch huge potential data sets and evolve hideous numbers of very very short term permutations to decide upon tactics that computer AI will ever have a chance. Your AI is dumb because you built it sir, it will always have all the limitations it had the day it was coded, because in truth you never let it decide anything, its nothing more than discrete decision making paths. Of course that will never advance. We can only build our decision making models so complex, only to emulate such a degree of humanity. Thats not AI. Thats just a robot.
All real intelligence is derived from the knowledge of the interconnected nature of realities.
Then again, Armored Core is my game of choice. You'd have to know it to know. Sweet sweet customizability.
Luv,
LordMyren
Silly cod, you missed the most obvious joke in all eternity!
.com domain name until end of time!
Verisign to control
It wasnt just the end of the mayan calender. They werent like, well, this should be enough. It was the end of the 13 eon, where time itself ended and perhaps looped back to the beginning. It is the end of time.
Erm, no need to make an operating system. Bork.
There's no need to make an object just to support an object oriented system. I'd posit that such a system could easily be made cross-platform. Shared memory is a nice-ity, but I dont see that a better implementation thereof compels an entirely new OS to be born. Shared mem aint pretty, but its doable.
Furthermore, many GUI systems are capable of using async callbacks for event notification. AKA generating new threads, as you describe. This process itself comes with (very optimizable) overhead though, as you know well.
These aside, you my friend, are spot on. My heartfelt thanks. I thought I was the only one who was crazy.
My main suspect is that they're distracting our developers. all these UI-centric articles taht keep coming out are simply a red herring to make OSS try to out Redmond MS. Somehow we've fallen for the ploy and decided taht linux will only win by being more user friendly.
Fucking bullshit. Linux community need to get off its ass and start innovating again.
Unless the signal can be predicted, which does not seem wholly unlikely. Most background noises are of a fairly constant nature. Record a couple sample sounds and look at the waveforms very close in; a large number of background sounds are repetitive. These are easy to predict and thus could be cancelled, even given the processing delay.
There's going to be some variation. The technique will not be perfect, but it should be able to achieve rather significant cancellation for many noise sources.
Afaik the real bitch is that the soundcards have different skews, system clocks are easily adjustable and de-skewed with working ntp. Getting the sound card not to drift from the computer--- more problematic. also, low latency interrupts are very problematic to generate to start play.
Please discuss though... I'm very interestd in getting synced audio.
Excellent, thanks so very much, the Hot/Cold is exactly what I need to save myself some perposterous high-availability pains.
Will it be possible to find this sort of support in "consumer" RAID gear? I'm looking at the Intel SRCU42X and LSI MegaRAID 320-2e (both $700) and dont know how it would work. The main thing I dont understand is what form the utilities for these kinds of cards will be in... how arrays are setup, how hotswap is performed, &all... I doubt the firmware interfaces are open enough to allow for GPL kernel modules. BIOS isnt an option since they need to be configured and reconfigured on the fly.
Good advice on the local boot drives. Thanks again!
Myren
I ask because its either this or using drbd to replicate the entire file system over multi GigE lines while having to use twice the number of hard drives. I'd much prefer to avoid these interconnections altogether and simply have SCSI itself be the common communication bus, at least such that either controller can access the raid array should the other fail. I'm not /totally/ OT. I was looking at 4 gige or a 10gb solutions which would have pounded cpu usage to death... I'd much rather just make sure I can always access the drives.
Myren
I'm aware that it's feasible to have multiple SCSI cards in a SCSI chain, is there any way to do this with hardware RAID cards? It would be a great boon to reliability & costs were this feasible for me. Systems fail but RAID-5 and RAID-6 far less so.
Myren
I vote for the first person who will vow to stab verisign in the eyeballs^H^H^H^H^H^H^H^Hstop sitefinder. Once or twice a week I get redirected to a Network Solutions sitefinder-a-like.
For the most part the internet is governed by the corps, who basically understand its only useful because it is a free for all. For the most part, this works fine, with occasional DNS hijinx. I'm rather scared of what would happen if a government or governing agent decided it actually wanted some form of control. Kiss those privacy rights good bye.
-Myren
I'm waiting for the M-5000u to come out before I buy a media player box. Although I suspect the morons will botch what would be an almost perfect media player by omitting wifi. It does have an internal hard drive, and I have enough extra OpenWRT boxes lying around that this isnt a HUGE problem, but its still an awfully glaring omission.
Hope you like the Squeezebox. I'm considering getting one while I wait.
-Myren
I'm hoping we see some new alternatives for remote controls crop up. I'm still using two ATI TV Wonder remotes and a pile of USB recievers (all first gen) to control my medianet. The range is beyond pathetic, the control pad is 4-directional (what I wouldnt give for analog), the remote is huge but it does have decent linux support. And its RF. The advantage is that the remotes and dongles were $6 a pop on ebay and hus cheap enough that I can litter the house with dongles for decent coverage.
I'd like to see Zigbee or Bluetooth solutions crop up. But I'll take anything better if its at a good price. People dont want to pay $100 for a damned remote control, there's little reason for it.
Linux or BSD boxes are free on the curb. Either is fine if you dont dick with it constantly, put it in a closet and go. Gigabit is $100 for the switch alone, I havent found any in the trash yet. Radio stations often operate on shoestring budgets. We have the largest record collection on the east coast and the station still operates on pennies.
Besides, there's no reason you need gigabit for such a piddling task. If someone's going to be saturating a 100bt such that even a 1.4mbps stream is underbuffering, there's a good chance of under run on gigabit too.
QoS is the only acceptable solution for a radio station. If you have to find two boxes on the street for absolute redunancy, go for it. Get redunant $5 switches too. While you are at it, the most likely thing to fail will be the hard drive; if you need bullet proof do netboot to hard drive less routers or use one of the distro's on floppy or distro on cd's. There are some good distro's made explicitly for routing for this very purpose.
Really, the big problem with this is that if you are on a switch all packets now have to flow through the single box. In effect, its much like running a full duplex hub from there since all communication is flowing through the single point.
That aside, I'd like to thank you for at least providing a valid argument against my post. The other two people who replied were a bit more challeneged. I like to think I at least made some sense. Aside from DHCP relays I dont think anything I mentioned wasnt straight out of LARTC. And you do have an excellent point that should definately be taken into consideration.
Setup a linux box to do all routing for the small mini network. By giving each system a netmask of 255.255.255.255 you can make all traffic go through the linux router. If you've got static ip's for the machines, use em as normal, otherwise setup the router to do DHCP and use NAT connection sharing. If you want to get fancy you can setup some hijinx with DHCP relay systems. Easy enough. From there you can either mangle the gateway address to point to your router or make the linux box a transperent gateway- very clean, probably the way to go but you still have to mangle whatever netmask the actual DHCP srever is providing. Details details. Basically get all traffic going through the linux box. I suggest just using an internal DHCP network, its the simplest most exandable.
Then just setup traffic policing. Cap all connections to like 50% of the 100bt's TESTED capbilities except for the couple ports doing streaming. Dead fsckign simple QoS stuff, just a Token Bucket Filter. Traffic policing is also known as Ingress queueing, and is made possible by Intermediate queueing devices (IMQ).
Basically the Linux Advanced Routing and Traffic Control HOWTO will become your new bible. It has everything you need.
Now that I think about it, I dont suppose theres any reason you couldnt just do packet queueing/egress filtering too once you get the linux system acting as a router.
I dunno, I really think we need an effective space plane, not just better and better rockets. Something like White Knight to 40,000 feet, accelerated launch to 120,000 feet, start skipping across the atmosphere doing periodic burns when atmosphere is available. Build up speed, get near double digit mach before turning on the boosters.
I guess we've just got to rely on NASA to work on scramjet and hope they get somewhere first.
Dont worry, the capitalist system of intellectual property is trying as hard as it can to legislate and litigate around this potential.
Capitalism has been errected upon efficiencies of industry, not around innovation. And most dangerously, its legislated protection on high for itself. The singularity will be built around rouge elements [F/OSS] doing projects in their basement, around remixed and re-remixed and re-re-remixed technologies that build massive network externalities. The singularity is the rapid cascade of disruptive technologies that rebuilds the entire technological platform we've built underfoot, a complete architectural revamp that encompasses the loosely coupled distributed nature of information and processes.
By its very definition, the technologies which drive the singularity will be so flexible that they will cause clash after clash with existing IP, simply because these new technologies are so malleable, so capable of outperforming the existing although still infringing. The revolution will be litigated-- but the corporations will loose. IP is not a sustainable edge.
As always, http://ccs.mit.edu/21c/21CWP001.html">Two Scenarios for 21st Century Organizations.