Tiny base OS (linux), booted from PXE/TFTP server running from ramdisk. networking, storage, snmp, ssh, grid engine, "botnet" client and bugger all else. On top of that you run a VM host, Xen, VMware, vserver or whatever fits your requirements. This is the infrastructure platform. It can be rolled out to anything which supports PXE. Tens of thousands of machines if required. They can be functional literally as fast as machines can be fitted into racks.
Basically you don't touch a machine till it comes up and reports that it's ready. At that point you know you've got a host there waiting to provide whichever service you're interested in. If it's an application server you simply add it to the list and the load distribution system handles the migration of applications. when the average load is too low, below 1 for each CPU for instance, you pop a machine off the list, the controller tells the machine to get rid of it's applications. Then you switch it off at the socket. It's all just scripting, there is no real coding required.
In terms of application scalability and redundancy, well that depends on the application itself and how it's configured. DB clusters for instance are a bit of a pain to manage because the physical location and remoteness from other cluster machines severely reduces performance so sometimes you have to specify which machines are able to run which applications.
It's not suitable to handle 5min peaks in load, it takes between 15 and 30 mins for a machine to configure itself, but over night, weekends it works.
People seem to think that geniuses are simply more intelligent than the rest of us, I hear talk of IQs of 200, 250 etc. Which is utter bullshit, there aren't enough people on the planet for that, never mind the validity of IQ tests. What you really see when you take a look at the life of a genius is damned near monomania. The drive, ability and desire to focus on a single thing for years, decades, to the exclusion of almost everything else. To the point that they finally see "the truth" or at least, closer to the truth than the rest of us who are more distracted by daily life.
Not to say that geniuses aren't spectacularly talented people, obviously they are, but what really makes the difference is focus.
We do something like this, but from scratch rather than using LTSP. It's really not difficult, just a slightly different way of looking at how an operating system and server application should work. Think botnet. It's a fundamental shift in the mathematics of computing infrastructure, from linear or worse to logarithmic.
o that means that servers should be built the Gentoo way, from scratch, using just the things you need, no more, no less. How much does it cost to deploy such a server? How much does it cost to pay someone qualified enough to do it properly? Frankly anyone with half a brain can pretty much use mkinitrd to make such a server.
How much does it cost to hire 500 admins for thousands of machines rather than half a dozen? How much does electricity and AC cost?
Meh, no point explaining. The price of oil and the economics will do that job.
Switch the machines off at the the socket. You can do it using SNMP. Monitor the average load on your machines, if too low, migrate everything off it and switch a machine off. If too high, switch one on.
Course it assumes you know how to create highly available load balanced clusters. Automatic installations, network booting and all that. Not so difficult.
LTO4 drives can't read DLT tapes. You need to maintain a DLT drive to read the "archived" data on DLT media. You may also need to maintain the backup software and possibly even the server the software is running on if it isn't cross platform.
My job is not in jeopardy:) Only if, nobody where you work understands the nature and problems associated with long term archival, and most likely they won't have given it a moment's thought. And of course by the time they hit the problems you will have moved on anyway, though if you have "archives" on DLT, you already have a problem.
Say that again? Do you mean it's abysmal because of access latency (which I suppose would be valid), or from a degradation perspective? No, it's abysmal because tape separates the data access technology from the media. CD & DVD are just as bad. It means that 20 years later, you have no way to read the information... or...
You have to maintain a mainframe for 20 years with 45mb reel to reel tape units, with version N.NN of backup system Blahblah, and hire someone who understands the hardware and software in order to read off the "archived" data.
Archival is a process, not a technology. Archive constantly. Plan for it and consciously convert your data into formats you will always be able to use and move it onto the new "double the size" storage system every 5 years when you move everything else.
In 10 years you won't have anything which can read those tapes. You're going to have to go to a specialist to find the hardware and software to get the information back... Or rather, your successor will.
You archive your information in some open, standardised format to an immediately accessible system, like an array of disks. That is your archive, it can be easily and automatically moved to new systems as technology changes. Then you back it up to prevent short term loss of the archive in the case of robbery, fire etc. The backups are not meant to be the archive. The format of the backup can then be the backup system flavour of the month.
Backups are NOT archives and archives are NOT backups. The requirements are entirely different.
How come they never shipped any good thing to USA market? Cos the US telcos won't let them. Nokia supplies the mobile telcos, they can only sell what the telcos will buy.
The US mobile market is owned and you are their bitch.
Every year... They are basically doing what the OLPC people would love to do in their wettest dreams.
These guys can be rendered redundant by simple pieces of hardware placed into a critical mass of hands. How many handsets do you make and sell?
You see, putting a "simple piece of hardware" into a critical mass of hands is not the same as copying a piece of software. It is a linear process, you need an infrastructure which can produce and distribute that critical mass of handsets and that requires a huge investment.
Getting Linux onto Nokia phones is a huge leap forward, it is a step past the desktop which is now largely irrelevant. As long as they stick to the GPL (and they will, their lawyers and developers will be perfectly aware of the issues) what they actually do with it is up to them. That is almost certainly going to include DRM, locked down hardware and patented software because that is what their customers (the mobile networks) demand of them.
But you know what? That phone is still a Linux box.
i find it ironic that the Democratic party in the US is accused of being liberal when in fact they are social democrats. Ron Paul's policies are closer to true liberalism than either of the larger parties.
"Hence to fight and conquer in all your battles is not supreme excellence; supreme excellence consists in breaking the enemy's resistance without fighting." Said over 2500 years ago. That's before Jesus Christ superstar, btw.
You should all read Sun Tzu and have a bash at playing "Go" as well.
Haven't for decades. The kinetic energy involved doesn't get much higher.
Fabric can be made rigid and extremely durable simply by doping it with an epoxy. Think cellulose reinforced plastic if you will, y'know, like motorcycle helmets.
Slashdot Mirror
Tiny base OS (linux), booted from PXE/TFTP server running from ramdisk. networking, storage, snmp, ssh, grid engine, "botnet" client and bugger all else. On top of that you run a VM host, Xen, VMware, vserver or whatever fits your requirements. This is the infrastructure platform. It can be rolled out to anything which supports PXE. Tens of thousands of machines if required. They can be functional literally as fast as machines can be fitted into racks.
Basically you don't touch a machine till it comes up and reports that it's ready. At that point you know you've got a host there waiting to provide whichever service you're interested in. If it's an application server you simply add it to the list and the load distribution system handles the migration of applications. when the average load is too low, below 1 for each CPU for instance, you pop a machine off the list, the controller tells the machine to get rid of it's applications. Then you switch it off at the socket. It's all just scripting, there is no real coding required.
In terms of application scalability and redundancy, well that depends on the application itself and how it's configured. DB clusters for instance are a bit of a pain to manage because the physical location and remoteness from other cluster machines severely reduces performance so sometimes you have to specify which machines are able to run which applications.
It's not suitable to handle 5min peaks in load, it takes between 15 and 30 mins for a machine to configure itself, but over night, weekends it works.
News at 11.
People seem to think that geniuses are simply more intelligent than the rest of us, I hear talk of IQs of 200, 250 etc. Which is utter bullshit, there aren't enough people on the planet for that, never mind the validity of IQ tests. What you really see when you take a look at the life of a genius is damned near monomania. The drive, ability and desire to focus on a single thing for years, decades, to the exclusion of almost everything else. To the point that they finally see "the truth" or at least, closer to the truth than the rest of us who are more distracted by daily life.
Not to say that geniuses aren't spectacularly talented people, obviously they are, but what really makes the difference is focus.
We do something like this, but from scratch rather than using LTSP. It's really not difficult, just a slightly different way of looking at how an operating system and server application should work. Think botnet. It's a fundamental shift in the mathematics of computing infrastructure, from linear or worse to logarithmic.
How much does it cost to deploy such a server?
How much does it cost to pay someone qualified enough to do it properly? Frankly anyone with half a brain can pretty much use mkinitrd to make such a server.
How much does it cost to hire 500 admins for thousands of machines rather than half a dozen? How much does electricity and AC cost?
Meh, no point explaining. The price of oil and the economics will do that job.
Switch the machines off at the the socket. You can do it using SNMP.
Monitor the average load on your machines, if too low, migrate everything off it and switch a machine off. If too high, switch one on.
Course it assumes you know how to create highly available load balanced clusters. Automatic installations, network booting and all that. Not so difficult.
LTO4 drives can't read DLT tapes. You need to maintain a DLT drive to read the "archived" data on DLT media. You may also need to maintain the backup software and possibly even the server the software is running on if it isn't cross platform.
My job is not in jeopardyYou have to maintain a mainframe for 20 years with 45mb reel to reel tape units, with version N.NN of backup system Blahblah, and hire someone who understands the hardware and software in order to read off the "archived" data.
Archival is a process, not a technology. Archive constantly. Plan for it and consciously convert your data into formats you will always be able to use and move it onto the new "double the size" storage system every 5 years when you move everything else.
In 10 years you won't have anything which can read those tapes. You're going to have to go to a specialist to find the hardware and software to get the information back... Or rather, your successor will.
You archive your information in some open, standardised format to an immediately accessible system, like an array of disks. That is your archive, it can be easily and automatically moved to new systems as technology changes. Then you back it up to prevent short term loss of the archive in the case of robbery, fire etc. The backups are not meant to be the archive. The format of the backup can then be the backup system flavour of the month.
Backups are NOT archives and archives are NOT backups. The requirements are entirely different.
HTH.
And tape is an abysmal archival medium.
How long have you worked in the industry?
HTH.
Work out how the mathematics of software development works.
The UK NHS and ID card systems promise to completely blow tens of billions of GBP. But hey, it's only tax.... That's like, free, right.
One of the reasons i'm not too worried about ID cards.
You just keep slowing down.
Very cool bit of kit, shame it seems to be dead now. Highly scriptable too.
The US mobile market is owned and you are their bitch.
Not the most rational or careful bunch of people. Best thing you could do is give them an A-bomb to play with.
It's basically the equivalent to a $5 cable to them.
You see, putting a "simple piece of hardware" into a critical mass of hands is not the same as copying a piece of software. It is a linear process, you need an infrastructure which can produce and distribute that critical mass of handsets and that requires a huge investment.
Getting Linux onto Nokia phones is a huge leap forward, it is a step past the desktop which is now largely irrelevant. As long as they stick to the GPL (and they will, their lawyers and developers will be perfectly aware of the issues) what they actually do with it is up to them. That is almost certainly going to include DRM, locked down hardware and patented software because that is what their customers (the mobile networks) demand of them.
But you know what? That phone is still a Linux box.
I say good luck to them.
They've pulled 500 million people out of grinding poverty in 20 years. Just another 500 million to go...
i find it ironic that the Democratic party in the US is accused of being liberal when in fact they are social democrats. Ron Paul's policies are closer to true liberalism than either of the larger parties.
Z
"Hence to fight and conquer in all your battles is not supreme excellence; supreme excellence consists in breaking the enemy's resistance without fighting." Said over 2500 years ago. That's before Jesus Christ superstar, btw.
You should all read Sun Tzu and have a bash at playing "Go" as well.
And I wanted to know the fastest way to level up.
Haven't for decades. The kinetic energy involved doesn't get much higher.
Fabric can be made rigid and extremely durable simply by doping it with an epoxy. Think cellulose reinforced plastic if you will, y'know, like motorcycle helmets.
The irony being that the Trabants here in Germany still look exactly as they did 30 years ago. Steel cars have long since rusted to scrap.