The point is that the new model shows that what was previously thought to be sown up, the shortfall between the observed matter, and the amount required to account for the observed behavior is not quite as sown up as we thought.
So while this discovery does not mean that we have now observed all the mass necessary, it does mean that it would be prudent to look again very hard at how we have derived the mass of the universe in case we have left out mass along the line.
There are also other challenges on dark matter. The reason the whole concept exists is that there does not appear to be enough visible matter to explain the rotation of galaxies. However even this has recently being challenged, with the argument that using Newtonian dynamics to model galactic rotation is flawed, and if you do the same modeling using General Relativity (much much harder to do) the missing mass appears to vanish. I am the first to admit that there are issues with the paper that proposes this. However it is an important new avenue of research.
There is also the possibility that we might have gravity wrong, at very low accelerations which would also make dark matter go away.
My personal feeling is that dark matter is about as likely as the ether, and in reality we have not counted the mass accurately and are miss-applying theories.
Then again I think Copenhagen interpretation is hokum as well.
Fedora is not suitable for enterprise deployment. I don't want to have to upgrade my desktops every 12 months to keep getting security updates. Hey even for my personal desktop, having to upgrade every 12 months is just too dam frequent.
I has not always done so. Therefore if at some point in the past you found out it did not work, then you would tend to think it still does not. My Debian Etch installation is using lilo for the very reason that it is pure XFS.
Yes but at the time ReiserFS was brand new. However now in 2008, XFS has been available on Linux for at least six years, and has a much longer pedigree on Irix. The situation is not comparable at all. I would also say that ext3 is an dismal filesystem for a large one. Nobody in their right mind would push ext3 to it's limits.
Problem with AoE is that it is now 20 years and horribly out of date. When talking about microcontrollers it is 68000, no mention of a PIC, AVR or similar. When it talks about programmable logic, as far as it goes is very simple PLD's. So no CPLD's and no FPGA's. The PSU section is also out of date when it comes to SMPS as well.
Don't get me wrong I have a copy, and you won't get me to reliquish it till a third edition comes out.
ext4 is the biggest waste of time and effort in Linux. There are already good extent based filesystems for Linux. Why anyone would consider using what is an experimental filesystem for a multi TB production filesystem is beyond me.
What ever they do XFS and JFS will have way more testing and use than ext4 will ever have. I just don't get the point of ext4. It would be far more useful to fix the one remaining issue with XFS, the inability to shrink the filesystem none destructively, than to flog the dead horse which is ext2/3 even more with ext4, which is not one disk compatible anyway.
Having been the victim of spam backscatter on several occasions in the last five years, it occurred to me some years ago, the solution to bounce issues was to insert random ID into each email as a header. Then track these against the domain they where sent to. Only bounces from matching domains, that contained the magic ID would ever get delivered.
Far more productive to forward it to some email address at Barracuda.
Back when Sobig.F hit, I was on ISDN. At something like 50,000 emails I stopped even counting. However it was the 5000+ spurious bounces that was impossible to filter automatically, as every f&*$ virus scanner sent back different responses. What annoyed me most where the ones that kindly told me it was Sobig.F. So you know that it is a virus that exclusively sends out emails with forged from addresses and you choose to spam innocent victims.
Anyway I had a could of large corporations, who had sent in excess of 1000 bogus warnings each. I proceeded to bounce each and every bogus warning along with an explanation to a suitable contact email at said corporation. One responded with an apology, and said they had already turned this off. The other asked me to stop, but refused to stop sending the bogus bounces, claiming it was "corporate" policy. I told them it would stop when they stopped spaming my inbox. They got something like 2500 emails before Sobig.F was done.
Not quite, the executor of the will, turns up at the bank with proof that he is the executor, and takes posession of the contents of the box to distribute as per the will. If there is no will then the next of kin becomes that person by default.
It sounds likely that this person did not leave a will, so the next of kin gets to do as they please.
The ugly glue code is all but gone now. There was an issue that was fixed maybe five years ago where an unclean shutdown would leave open files truncated to zero bytes. Have not seen that for ages.
Even the issue with xfs_repair requiring unholy amounts of RAM on large filesystems; 2GB of RAM per TB, has been fixed it is now 128MB per TB and 4MB per million inodes.
The one remaining issue with XFS is that you cannot shrink it. Not that you often do this, but when you need to the fact that you cannot is a show stopper. It's absence from what is otherwise the only enterprise filesystem that checks all the boxes on Linux is a strange omission (the DMAPI implementation in JFS seems to be bit rotted).
There is even a cross platform cluster version in CXFS. So Linux, Mac OSX, Solaris, Irix, Windows and AIX can all mount the same filesystem at the same time. Though this does cost you $$$
Why on earth would IBM be interested in ReiserFS. They have JFS, which is their own and has a much better pedigree than ReiserFS, and they also have their high performance cluster filesystem GPFS, though admittedly this does cost $$$.
Assuming it is an IBM mainframe, it will be LTO or 3952, anything else and it might also be a DLT/SDLT variant. The market is rapidly converging on LTO, and the latest iteration LTO4 offers on drive encryption of your data. As has any decent enterprise backup software for like a decade.
I would add to this that every enterprise backup system that I know of has had the ability to encrypt the backup for ages. It's number six on the Tao of Backup, and that is 11 years old.
If the contents of your tapes are encrypted it matters not if they go missing.
And the Velociraptor or even the old 150GB Raptor will wipe the floor against any 7200RPM setup in the real world. It is all about IO's per second. In my last job I had a 150GB Raptor in my workstation, fastest drive I have ever had. I could boot four VM's at once and do other stuff without a problem. Before the upgrade doing that sort of thing was painful.
Why on earth everyone concentrates on throughput is beyond me. It is like measuring processor performance by looking at clock speed, and as we all know now that the lower speed Opteron's and PentiumM where besting the power hungry NetBurst P4's.
Assuming that your home does not need more than 55m cable runs (that's a big house if it does and you can probably afford to go on a holiday for a month while you have the house recabled), then Cat6 is good for at least 20 years. With 10GbE you can send 1080p uncompressed video at a whopping 200fps. You need serious high end server kit with mutli spindle SAS or FC RAID arrays to saturate a 10GbE link.
Lets face it the original 10Mbps which is now 28 years old is still faster than the vast majority of peoples internet connections, and 10BaseT is around 20 years old now.
Since when has 17.6 trillion USD been practically equal to 13.8 trillion USD? That is IMF estimated EU for 2008, and actual USA for 2007, but the IMF are predicting a mild recession in 2008 for the USA so that 13.8 is not going to improve. I believe Eurozone is a bit smaller, but that is mostly because the UK is not in the Euro, and we are the second largest economy in Europe after Germany.
All air conditioning has filters to remove the crap in it. However if you look at the max air inlet temperatures quoted by manufactures there are plenty of places in the UK that could provide it from outside air 365 days a year, as the maximum recorded historical temperature (over a couple hundred years) is below that of the max air inlet temperature.
A quick search shows that the maximum recorded air temperature in Reykjavik is just under 25 Celsius, so no air conditioning need even on the hottest of days. Given that the main power source in Iceland is hydroelectric and geothermal both of which have zero carbon foot print and the price of which does not depend on any external sources. With an estimated capacity of 50TW/h per year of hydro and geothermal yet undeveloped, looks like a good place to locate a datacentre to me.
Add to that list tape drives and libraries. I know there are a few SATA tape drives out their, but they are mostly SCSI (increasingly SAS) and Fibre Channel.
Before anyone says tape is dead, you replace my 1.5PB library which could run of a 3kVA UPS and needs no air con to speak of, with spinning disk at anywhere near the same upfront cost let alone the running cost.
USB drives at the moment might just be SATA or IDE interfaces with a USB bridge. However there is nothing stopping someone developing a direct FLASH to USB or Firewire or SAS or SCSI, or FC device. Should Seagate push this, I am quite sure that some other interface will come to the fore.
Also their claims of silent are bogus. Simple fact is the much of the noise from fans comes from the noise as the air blows over the heatsink etc. and not from the fan itself. As this design still has air being blown over heatsinks it will make noise.
Slashdot Mirror
The point is that the new model shows that what was previously thought to be sown up, the shortfall between the observed matter, and the amount required to account for the observed behavior is not quite as sown up as we thought.
So while this discovery does not mean that we have now observed all the mass necessary, it does mean that it would be prudent to look again very hard at how we have derived the mass of the universe in case we have left out mass along the line.
There are also other challenges on dark matter. The reason the whole concept exists is that there does not appear to be enough visible matter to explain the rotation of galaxies. However even this has recently being challenged, with the argument that using Newtonian dynamics to model galactic rotation is flawed, and if you do the same modeling using General Relativity (much much harder to do) the missing mass appears to vanish. I am the first to admit that there are issues with the paper that proposes this. However it is an important new avenue of research.
There is also the possibility that we might have gravity wrong, at very low accelerations which would also make dark matter go away.
My personal feeling is that dark matter is about as likely as the ether, and in reality we have not counted the mass accurately and are miss-applying theories.
Then again I think Copenhagen interpretation is hokum as well.
Fedora is not suitable for enterprise deployment. I don't want to have to upgrade my desktops every 12 months to keep getting security updates. Hey even for my personal desktop, having to upgrade every 12 months is just too dam frequent.
I has not always done so. Therefore if at some point in the past you found out it did not work, then you would tend to think it still does not. My Debian Etch installation is using lilo for the very reason that it is pure XFS.
Yes but at the time ReiserFS was brand new. However now in 2008, XFS has been available on Linux for at least six years, and has a much longer pedigree on Irix. The situation is not comparable at all. I would also say that ext3 is an dismal filesystem for a large one. Nobody in their right mind would push ext3 to it's limits.
Problem with AoE is that it is now 20 years and horribly out of date. When talking about microcontrollers it is 68000, no mention of a PIC, AVR or similar. When it talks about programmable logic, as far as it goes is very simple PLD's. So no CPLD's and no FPGA's. The PSU section is also out of date when it comes to SMPS as well.
Don't get me wrong I have a copy, and you won't get me to reliquish it till a third edition comes out.
ext4 is the biggest waste of time and effort in Linux. There are already good extent based filesystems for Linux. Why anyone would consider using what is an experimental filesystem for a multi TB production filesystem is beyond me.
What ever they do XFS and JFS will have way more testing and use than ext4 will ever have. I just don't get the point of ext4. It would be far more useful to fix the one remaining issue with XFS, the inability to shrink the filesystem none destructively, than to flog the dead horse which is ext2/3 even more with ext4, which is not one disk compatible anyway.
Having been the victim of spam backscatter on several occasions in the last five years, it occurred to me some years ago, the solution to bounce issues was to insert random ID into each email as a header. Then track these against the domain they where sent to. Only bounces from matching domains, that contained the magic ID would ever get delivered.
Far more productive to forward it to some email address at Barracuda.
Back when Sobig.F hit, I was on ISDN. At something like 50,000 emails I stopped even counting. However it was the 5000+ spurious bounces that was impossible to filter automatically, as every f&*$ virus scanner sent back different responses. What annoyed me most where the ones that kindly told me it was Sobig.F. So you know that it is a virus that exclusively sends out emails with forged from addresses and you choose to spam innocent victims.
Anyway I had a could of large corporations, who had sent in excess of 1000 bogus warnings each. I proceeded to bounce each and every bogus warning along with an explanation to a suitable contact email at said corporation. One responded with an apology, and said they had already turned this off. The other asked me to stop, but refused to stop sending the bogus bounces, claiming it was "corporate" policy. I told them it would stop when they stopped spaming my inbox. They got something like 2500 emails before Sobig.F was done.
Not quite, the executor of the will, turns up at the bank with proof that he is the executor, and takes posession of the contents of the box to distribute as per the will. If there is no will then the next of kin becomes that person by default.
It sounds likely that this person did not leave a will, so the next of kin gets to do as they please.
The ugly glue code is all but gone now. There was an issue that was fixed maybe five years ago where an unclean shutdown would leave open files truncated to zero bytes. Have not seen that for ages.
Even the issue with xfs_repair requiring unholy amounts of RAM on large filesystems; 2GB of RAM per TB, has been fixed it is now 128MB per TB and 4MB per million inodes.
The one remaining issue with XFS is that you cannot shrink it. Not that you often do this, but when you need to the fact that you cannot is a show stopper. It's absence from what is otherwise the only enterprise filesystem that checks all the boxes on Linux is a strange omission (the DMAPI implementation in JFS seems to be bit rotted).
There is even a cross platform cluster version in CXFS. So Linux, Mac OSX, Solaris, Irix, Windows and AIX can all mount the same filesystem at the same time. Though this does cost you $$$
Why on earth would IBM be interested in ReiserFS. They have JFS, which is their own and has a much better pedigree than ReiserFS, and they also have their high performance cluster filesystem GPFS, though admittedly this does cost $$$.
Assuming it is an IBM mainframe, it will be LTO or 3952, anything else and it might also be a DLT/SDLT variant. The market is rapidly converging on LTO, and the latest iteration LTO4 offers on drive encryption of your data. As has any decent enterprise backup software for like a decade.
Nobody uses 9" real to real tapes these days.
I would add to this that every enterprise backup system that I know of has had the ability to encrypt the backup for ages. It's number six on the Tao of Backup, and that is 11 years old.
If the contents of your tapes are encrypted it matters not if they go missing.
If you want to boost your IO's per second, and stretching to SAS is too pricey, then the Raptor is the best drive around.
And the Velociraptor or even the old 150GB Raptor will wipe the floor against any 7200RPM setup in the real world. It is all about IO's per second. In my last job I had a 150GB Raptor in my workstation, fastest drive I have ever had. I could boot four VM's at once and do other stuff without a problem. Before the upgrade doing that sort of thing was painful.
Why on earth everyone concentrates on throughput is beyond me. It is like measuring processor performance by looking at clock speed, and as we all know now that the lower speed Opteron's and PentiumM where besting the power hungry NetBurst P4's.
What I would like to see is holographic technology in an LTO tape cartridge. Now we would be talking, more in the region of 100TB a tape...
Assuming that your home does not need more than 55m cable runs (that's a big house if it does and you can probably afford to go on a holiday for a month while you have the house recabled), then Cat6 is good for at least 20 years. With 10GbE you can send 1080p uncompressed video at a whopping 200fps. You need serious high end server kit with mutli spindle SAS or FC RAID arrays to saturate a 10GbE link.
Lets face it the original 10Mbps which is now 28 years old is still faster than the vast majority of peoples internet connections, and 10BaseT is around 20 years old now.
Since when has 17.6 trillion USD been practically equal to 13.8 trillion USD? That is IMF estimated EU for 2008, and actual USA for 2007, but the IMF are predicting a mild recession in 2008 for the USA so that 13.8 is not going to improve. I believe Eurozone is a bit smaller, but that is mostly because the UK is not in the Euro, and we are the second largest economy in Europe after Germany.
You would need four 10Gbps ethernet cards, which is going to set you back about 6000USD, then a suitable switch, probably double that.
All air conditioning has filters to remove the crap in it. However if you look at the max air inlet temperatures quoted by manufactures there are plenty of places in the UK that could provide it from outside air 365 days a year, as the maximum recorded historical temperature (over a couple hundred years) is below that of the max air inlet temperature.
A quick search shows that the maximum recorded air temperature in Reykjavik is just under 25 Celsius, so no air conditioning need even on the hottest of days. Given that the main power source in Iceland is hydroelectric and geothermal both of which have zero carbon foot print and the price of which does not depend on any external sources. With an estimated capacity of 50TW/h per year of hydro and geothermal yet undeveloped, looks like a good place to locate a datacentre to me.
Add to that list tape drives and libraries. I know there are a few SATA tape drives out their, but they are mostly SCSI (increasingly SAS) and Fibre Channel.
Before anyone says tape is dead, you replace my 1.5PB library which could run of a 3kVA UPS and needs no air con to speak of, with spinning disk at anywhere near the same upfront cost let alone the running cost.
Try the System Rescue CD, has ClamAV and ntfs-3g on it, and will do a live ClamAV over the net update for the very latest signatures.
USB drives at the moment might just be SATA or IDE interfaces with a USB bridge. However there is nothing stopping someone developing a direct FLASH to USB or Firewire or SAS or SCSI, or FC device. Should Seagate push this, I am quite sure that some other interface will come to the fore.
Also their claims of silent are bogus. Simple fact is the much of the noise from fans comes from the noise as the air blows over the heatsink etc. and not from the fan itself. As this design still has air being blown over heatsinks it will make noise.
Wake me up when I can embed my XML based SVG graphics into my XML documents.