Re:No need for NAT router for open socket migratio
on
OpenMosix
·
· Score: 1
Switches have finite-sized ARP caches, so by artificially inflating them you increase the number of misses
no, switches doesn't have ARP caches, they have MAC-Port tables. the ARP tables reside on the hosts.
Also, each ARP packet will itself be broadcast, causing more traffic up front.
the ARP-update packets (yet to be invented, i think) would include the new MAC address (or a special, host-bound IP address), that would be broadcast (bad, but only one packet), there wouldn't be a new ARP query.
in any case it woulbe FAR less traffic than the current moel of sending everything to the original node, just to be retransmitted. since in most scenarios every process would originate at the same node, it would easily saturate one leg of the switch.
a process-bound IP would allow network migration with just an extra broadcast packet for each migration (usually the migration needs to transfer a few megs per se)
Re:No need for NAT router for open socket migratio
on
OpenMosix
·
· Score: 1
Since Mosix clusters usually are in a closed network, could there be a unique IP number for EVERY Process? If so, the IP address could be tied to the process, instead of tied to the machine
i think Mosix tries to get a cluster-wide process id, so if it's 16bit long, it could be in the class-B form [net.net.pid.pid], instead of [net.net.host.host]
when a process is migrated, it would send a broadcast message to change ARP tables, so the same IP address would be sent to a different MAC address
it would be better than setting every ethernet port in promiscous mode, since that woudn't work in a switched network (or in myrinet, i think)
But the truth still is that the IEEE standard defines both 6pin and 4pin connectors (with or without power, respectively)
iLink and FireWire are only the brand names. Of course, Sony uses only 4pin ports, and Apple sports 6pins everywhere... (althought I think the original documents Apple sent to IEEE did include both types of port)
sorry, but if I am pissed off at the service I receive at a restaurant and I feel the need to stand in the parking lot telling people what I think I have all the rights in the world to do so. How the hell would this be any different?
all the rights in the world? right in all the world?
i'm not so sure... you can say a all you want 'bout freedom of speech, but it can be interpreted differently on different parts of the world. if you tried to do this inside the restaurant, you would be shown the exit, no doubt. that's why they had this sign with the "we reserve the rights of admission" (not sure of the fraseology, i only know the spanish version).
would that include the parking lot?
would that include just out of the door?
could you post a sign on their front wall?
could you post a GPS sign on their space?
the exact limits are hazy, and different in different legislation; even on countries that pretend to preseve freedom of speech
Don't you know anything about human lazyness and self-deception? the cost analisys is irrelevant!
The meme that will stay in the market droid's head is "we'll have a favorable cost study" after a week, he'll think "we have a favorable cost study" he'll never check it!
And, at the next sales meeting with clients, he'll say "And of course you know all real cost studies show linux is far more expensive!"
And if you regularly hang of your fingernails over a 50m freefall, you'll really appreciate trimming down any non-vital weight. Trust me, you'll feel the difference day to day (instant gratification)
(of course, i'm 1.84m tall and have never been over 60kg, that's the s+:-- part of my geek code)
In Quantum teleportation, states aren't copied, since de original measured state is destroyed by the measurement. The resulting 'copy' is not a copy, but a 'reconstruction' of the original; the same down to quantum indeterminancies.
And the FTL part isn't correct either, since the teleportation needs a 'classical component' via normal information transfer. The 'spooky action at a distance' (as Einstein called it (have you read his article on that? very interesting reading)) is FTL, but can't be measured without destroying the quantum state.
Obviously, that means they can control which program you use on windows! of course, it's not your machine, it's their OS!
it's supposed to first unload 'sensitive data' and/or stop 'trusted applications'... but the trusted application will be your mouse driver, and the sensitive data is the page swap table
that's it! the perfect excuse so nobody can play on their backyard
The main advice seems to be "get rid of IDE drives, use SCSI", mostly because of the lower CPU utilization (the SCSI cards are far more intelligent than the IDE chips on the mainboard)... but SCSI drives are so much more expensive per GB!
so... check the Arena Array, it's a harware RAID box, with a SCSI interface, but with IDE bays! it uses a separate IDE channel for each drive, so the througput is as high as possible, and the host interface is SCSI, so you can use those low-CPU-demmanding cards.
In a word: EditDV, from DigitalOrigin (now owned by Avid)
Longer: Mac G3/300, 30 GB IDE HD, 128MB RAM (a litthe short), Canon DV camcorders.
If you considered the iMac DV, look at it again. (no, don't think iMovie, that's just a toy) EditDV works on it, and there are fast firewire hard drives for external storage. If you work mainly on DV format (as I do) it's all as seamless as uncompressed video, and you don't need a great bandwidth from the drives (6 Meg/sec will do easily).
Slashdot Mirror
no, switches doesn't have ARP caches, they have MAC-Port tables. the ARP tables reside on the hosts.
the ARP-update packets (yet to be invented, i think) would include the new MAC address (or a special, host-bound IP address), that would be broadcast (bad, but only one packet), there wouldn't be a new ARP query.
in any case it woulbe FAR less traffic than the current moel of sending everything to the original node, just to be retransmitted. since in most scenarios every process would originate at the same node, it would easily saturate one leg of the switch.
a process-bound IP would allow network migration with just an extra broadcast packet for each migration (usually the migration needs to transfer a few megs per se)
Since Mosix clusters usually are in a closed network, could there be a unique IP number for EVERY Process? If so, the IP address could be tied to the process, instead of tied to the machine
i think Mosix tries to get a cluster-wide process id, so if it's 16bit long, it could be in the class-B form [net.net.pid.pid], instead of [net.net.host.host]
when a process is migrated, it would send a broadcast message to change ARP tables, so the same IP address would be sent to a different MAC address
it would be better than setting every ethernet port in promiscous mode, since that woudn't work in a switched network (or in myrinet, i think)
Maybe that's what the article writer means
But the truth still is that the IEEE standard defines both 6pin and 4pin connectors (with or without power, respectively)
iLink and FireWire are only the brand names. Of course, Sony uses only 4pin ports, and Apple sports 6pins everywhere... (althought I think the original documents Apple sent to IEEE did include both types of port)
sorry, but if I am pissed off at the service I receive at a restaurant and I feel the need to stand in the parking lot telling people what I think I have all the rights in the world to do so. How the hell would this be any different?
all the rights in the world? right in all the world?
i'm not so sure... you can say a all you want 'bout freedom of speech, but it can be interpreted differently on different parts of the world. if you tried to do this inside the restaurant, you would be shown the exit, no doubt. that's why they had this sign with the "we reserve the rights of admission" (not sure of the fraseology, i only know the spanish version).
would that include the parking lot?
would that include just out of the door?
could you post a sign on their front wall?
could you post a GPS sign on their space?
the exact limits are hazy, and different in different legislation; even on countries that pretend to preseve freedom of speech
Don't you know anything about human lazyness and self-deception? the cost analisys is irrelevant!
The meme that will stay in the market droid's head is "we'll have a favorable cost study" after a week, he'll think "we have a favorable cost study" he'll never check it!
And, at the next sales meeting with clients, he'll say "And of course you know all real cost studies show linux is far more expensive!"
It's lots of fun!
And if you regularly hang of your fingernails over a 50m freefall, you'll really appreciate trimming down any non-vital weight. Trust me, you'll feel the difference day to day (instant gratification)
(of course, i'm 1.84m tall and have never been over 60kg, that's the s+:-- part of my geek code)
In Quantum teleportation, states aren't copied, since de original measured state is destroyed by the measurement. The resulting 'copy' is not a copy, but a 'reconstruction' of the original; the same down to quantum indeterminancies.
And the FTL part isn't correct either, since the teleportation needs a 'classical component' via normal information transfer. The 'spooky action at a distance' (as Einstein called it (have you read his article on that? very interesting reading)) is FTL, but can't be measured without destroying the quantum state.
Obviously, that means they can control which program you use on windows! of course, it's not your machine, it's their OS!
it's supposed to first unload 'sensitive data' and/or stop 'trusted applications'... but the trusted application will be your mouse driver, and the sensitive data is the page swap table
that's it! the perfect excuse so nobody can play on their backyard
The main advice seems to be "get rid of IDE drives, use SCSI", mostly because of the lower CPU utilization (the SCSI cards are far more intelligent than the IDE chips on the mainboard)... but SCSI drives are so much more expensive per GB!
so... check the Arena Array, it's a harware RAID box, with a SCSI interface, but with IDE bays! it uses a separate IDE channel for each drive, so the througput is as high as possible, and the host interface is SCSI, so you can use those low-CPU-demmanding cards.
In a word: EditDV, from DigitalOrigin (now owned by Avid)
Longer: Mac G3/300, 30 GB IDE HD, 128MB RAM (a litthe short), Canon DV camcorders.
If you considered the iMac DV, look at it again. (no, don't think iMovie, that's just a toy) EditDV works on it, and there are fast firewire hard drives for external storage. If you work mainly on DV format (as I do) it's all as seamless as uncompressed video, and you don't need a great bandwidth from the drives (6 Meg/sec will do easily).
How can they patent it? the PowerPC has it since the 604