It depends on your definition of "shortage". I can think of at least two defintions
1: Where the supply of an item becomes (either through reduced supply, increased demand or some combination of the two) such that even if everyone who wanted it was prepared to pay considerablly above it's production cost some of the people who wanted it would have to lose out. Some things are permanently in this state due to limited natural resources. Other items get into this state from time to time due to bad luck on the production side or mispredicted demand (see hard drives after the flood for example). 2: When an item's availability through regular channels becomes a matter of luck and/or how long a person is prepared to queue rather than a matter of how much they are willing to pay.
Price controls (whether imposed by govnerment or by a manufacturer) have a tendency of turning the first type of shortage into the second type. This in turn creates secondary markets for the goods in question.
I don't think there is a standard for IPv6 NAT just like there weren't ever really formal standard for IPv4 NAT (there were some RFCs written with descriptions and reccomendations but they came AFTER nats started to be widely used)
I pointed out all the IPv4 address exhaustion issues, but was basically told to mind my own business. "No customer demand for this feature." Never mind that a few years down the pipe, customers would be very unhappy they didn't have it.
Honestly for the vast majority of buisnesses there is no need to move their internal communication to IPv6. Afaict the only buisnesses that are likely to have trouble running out of private IPs are those who are both massive and offer IP based services to their customers (e.g. comcast).
Systems that need internet access to servers on the internet may need some way of accessing v6 only servers in the future but my guess is that we are still a long way away from even small-time websites giving up their IPv4 presense. Also remember that in a few years time name based virtual hosting for SSL sites will start to become more practical as windows XP and older versions of andriod gradually fade away.
What will really be hit if v6 is not widely deployed soon is peer to peer communications. Large scale NATs are likely to be far less friendly to peer to peer than the small NATs found in homes today. Unfortunately I suspect there are more than a few in the ISP industry who would love to see bittorrent and it's ilk gone.
NAT is a hack, the client machines and the routers between the client and the NAT have no idea whether the address they are trying to connect to is internal or public.
So if company X uses 9.x.x.x for their internal network with NATs and companies Y and Z use 9.x.x.x for their public servers then users at company X will not be able to access internet services run by company Y because their packets will not reach the NAT at the edge of company X's network.
It is possible to request a direct assignment of provider independent IPv6 prefixes from for example RIPE (fees are around EUR 1300 per year). This is especially tailored for the use-cases you described. Also check: http://en.wikipedia.org/wiki/Provider-independent_address_space
Sure so E1300 per year plus almost certainly having to move up from buisness DSL/Cable to far higher priced dedicated lines. I imagine most small buisnesses reaction to that would be "FUCK IPv6"
Also IIRC one of the teir 1 ISPs refused to route provider independent IPv6 allocations? did that ever get resolved?
Fortunately it seems those miltantly against v6 nat got pushed out of the netfilter team and It looks like v6 nat support got accepted into linux.
Obviously the previous situation where public v4 IPs were virtually free and easy to get had to end.
Over the next few years people will have to evaluate what applications can use v6, what applications can use private v4 and what applications really need and can justify the cost of public v4.
ranges were given out like candy to anyone who asked in the early days of the web. Corporations, Government and Academics made a land grab because they were the only people who could use the resource at the time.
Remember in the early days of the internet there was only Class A, Class B and Class C (equivilent to/8,/16 and/24 nowadays), so if you were too big for a class C then you got a class B and if you were too big for a class B then you got a class A. This lead to many allocations being far bigger than they actually needed to be.
I've heard that Glasgow Uni has a/8 that's never had more than 10 addresses exposed to the Internet.
Sounds like it was either a myth or it was given back years ago. I don't recall ever seeing them on the/8 allocation list.
I could be wrong here but I suspect changing the prefix that is allocated through a router to a LAN is far more likely to cause problems with systems continuing to try and use stale addresses than changing the IP that is allocated to the WAN side of a NAT.
Combine this with the fact that IPv6 prefixes of the size ISPs are likely to hand out (somewhere between/48 and/64 depending on how shitty they are) are plentiful compared to IPv4 addresses and I would expect more ISPs to adopt a "sticky IP"* policy with v6 than with v4.
*That is IPs don't normally change but the ISP doesn't make any particular commitment about them not changing.
Not really. Diamond is a crystaline form of carbon. The carbon itself isn't worth much so the value must be the form it is in.
When carbon is put under sufficient pressure it crystalises into diamond. However the vast majority of those diamonds end up small and/or full of flaws. These small flawed diamonds are cheap enough (whether formed naturally or aritficially) that we coat drills and cutting discs with them.
Gemstone diamonds are much rarer naturally and harder to produce artificially (though CVD is starting to make them easier to produce artificially).
It seems to vary a lot. Some drives are so quiet you can't hear them unless you hold your ear to the machine. Others clunk like hell. Newer drives in general seem to be quieter than old drives
Indeed they can't, OTOH i'm not sure I want the government (or worse a consortium of representives of governments) designing my networks for me either.
That's still going to require NAT if they want to use less than one public IPv4 address per user. There may still be use cases for it, but I am not convinced.
True, you need a special NAT for the ISP end of ds-lite but you don't need to manage private v4 addresses in your acess network. This avoids the risk of address collision with your users private IP addresses and means that very large ISPS (like comcast who wrote the ds-lite spec) don't need to federate their access networks to conserve private v4 addresses. It also scales horizonally nicely (just route different users to different ds-lite AFTR units) and is easy to implient in clients (from the clients point of view it's just a tunnel providing a route to a NAT)
Can't comment on 4rd as I haven't looked through that spec myself.
In my opinion, if there's a legitimate way to make the vehicle move, there's a way to make the vehicle move. If you don't want the vehicle to move then you need to remove something from it that makes it move, preferably something that a thief wouldn't normally bring with them, like a coil wire on a vehicle with a distributor, or a fuel pump relay or ASD relay, or something like that. Come to think of it, one could probably relocate such a relay to the passenger compartment to allow one to use the relay itself like a key, removing it to immobilize the vehicle
Which works fine if it's just the one vehicle but if an automaker started doing that theives would just start carrying relays.
An incomplete list of people who will be affected:
1: admins/moderators of interactive websites who find it harder to identify/ban users because of the inevitable rise of ISP level NAT (granted this is already a problem to some degree but is likely to get much worse). 2: users hit by bans aimed at thier shared IP either because the website owner didn't know it was shared or because they decide that the collateral damage is acceptable. 3: users who use software that needs to accept incoming connections on packages that no longer receive a public v4 IP 4: users who need public v4 IPs for services they are hosting and see the prices rise to reflect the market value of IP addresses.
Because that's the only Internet connection they can offer soon.
Bullshit, they can offer you:
v4 only with a private IP and ISP level NAT v6+v4 dual stack with a public v6 IP a private v4 IP and ISP level v4 NAT v6 with a public v6 IP and ds-lite v6 with a public v6 IP with NAT64 and DNS64
They will also be able to offer public v4 as a premium service once they push their bottom tier of users onto one of the above.
Remember your ISP's existing v4 IPs won't dissapear, they will just have to reduce the average number used per customer (or buy IP addresses on the market) if they want to increase the number of customers they serve.
Any ISP/corp that doesn't want to pay for IPv4 addresses for their public services and pay for some mechanism to allow their users to access servers on the IPv4 internet is free to sit back and watch as they lose the ability to sell to many of their customers and buy from many of their suppliers over the internet. Of course by chosing this path they have a good chance of going broke (or getting overridden in a shareholder revolt) sooner rather than later but it is their choice to make.
Now personally I will offer any services i'm responsible for v6 where the provider offers it because I belive it is the right thing to do but i'm not going to pretend there is any strong reason for a buisness to do so.
The trouble is there has been a chicken and egg problem.
The internet is mostly a network of buisnesses (with the occasional academic network thrown in). Some of those buisnesses sell service to other buisnesses and consumers, some just use it to support their main buisness.
There is basically no benefit and significant cost to an buisness in deploying dual stack while v6 only nodes are basically unheard of. You can't really deploy v6 only nodes while there are a significant number of v4 only nodes*
So for each individual buisness (whether ISP or user) that makes up the internet the rational thing to do was to sit and wait and hope others would take the costs of being early adoptors of IPv6.
*unless you use something like NAT64 or DS-lite but if nodes were ok with being behind those they would probablly also be ok with being behind NAT44 so the only real benefit to doing it would be if your internal network was massive.
Smart meters have the advantage of being on-grid pretty much by definition. On-grid electricty is cheap.
off-grid small scale electricity is far more expensive. You either use primary batteries (only practical for very low consumption levels) or you use rechargable batteries and try to find some way to recharge them (not cheap to set up).
Radio losses are a bitch, only a tiny fraction of the original energy is left after a typical radio path.
So if you try and power your device over radio you end up with a tiny fraction of the original energy, then losses in conversion and then a tiny fraction of what is left after those losses getting back to the base-station. I'ts just not practical except for very short distance links to very low power devices (think: rfid tags).
Afaict supermarkets do it by leaving enough space arround the door that being near the door implies wanting to go through the door. Which works ok for them but is far too space-consuming for more general use.
The ideal RNG collects as much entropy from the real world as there is information in it's output. Second best is a cryptographically secure PRNG. To be cryptographically secure given an arbitary sized sample of the outut it must be computationally infeasible to predict the next bit with an accuracy greater than random chance. This requires both an algorithm that is resistant to reversal and sufficient seed data and internal state to prevent brute forcing of the random number genertor's state.
A lot of such devices are built round PCs (some use special embedded form factor PCs, others just have a normal PC tower sitting inside them) despite them being overkill and not particularly reliable. I guess it's because windows devs are easier to find than devs who can handle an arm linux board.
I do wonder why an i7 though, a celeron would be more than sufficent.
Slashdot Mirror
It depends on your definition of "shortage". I can think of at least two defintions
1: Where the supply of an item becomes (either through reduced supply, increased demand or some combination of the two) such that even if everyone who wanted it was prepared to pay considerablly above it's production cost some of the people who wanted it would have to lose out. Some things are permanently in this state due to limited natural resources. Other items get into this state from time to time due to bad luck on the production side or mispredicted demand (see hard drives after the flood for example).
2: When an item's availability through regular channels becomes a matter of luck and/or how long a person is prepared to queue rather than a matter of how much they are willing to pay.
Price controls (whether imposed by govnerment or by a manufacturer) have a tendency of turning the first type of shortage into the second type. This in turn creates secondary markets for the goods in question.
I don't think there is a standard for IPv6 NAT just like there weren't ever really formal standard for IPv4 NAT (there were some RFCs written with descriptions and reccomendations but they came AFTER nats started to be widely used)
I pointed out all the IPv4 address exhaustion issues, but was basically told to mind my own business. "No customer demand for this feature." Never mind that a few years down the pipe, customers would be very unhappy they didn't have it.
Honestly for the vast majority of buisnesses there is no need to move their internal communication to IPv6. Afaict the only buisnesses that are likely to have trouble running out of private IPs are those who are both massive and offer IP based services to their customers (e.g. comcast).
Systems that need internet access to servers on the internet may need some way of accessing v6 only servers in the future but my guess is that we are still a long way away from even small-time websites giving up their IPv4 presense. Also remember that in a few years time name based virtual hosting for SSL sites will start to become more practical as windows XP and older versions of andriod gradually fade away.
What will really be hit if v6 is not widely deployed soon is peer to peer communications. Large scale NATs are likely to be far less friendly to peer to peer than the small NATs found in homes today. Unfortunately I suspect there are more than a few in the ISP industry who would love to see bittorrent and it's ilk gone.
No
NAT is a hack, the client machines and the routers between the client and the NAT have no idea whether the address they are trying to connect to is internal or public.
So if company X uses 9.x.x.x for their internal network with NATs and companies Y and Z use 9.x.x.x for their public servers then users at company X will not be able to access internet services run by company Y because their packets will not reach the NAT at the edge of company X's network.
Having a public IP that changes from time to time is mildly annoying but can be worked arround with stuff like dyndns.
Not having a public IP at all is much worse.
It is possible to request a direct assignment of provider independent IPv6 prefixes from for example RIPE (fees are around EUR 1300 per year). This is especially tailored for the use-cases you described. Also check: http://en.wikipedia.org/wiki/Provider-independent_address_space
Sure so E1300 per year plus almost certainly having to move up from buisness DSL/Cable to far higher priced dedicated lines. I imagine most small buisnesses reaction to that would be "FUCK IPv6"
Also IIRC one of the teir 1 ISPs refused to route provider independent IPv6 allocations? did that ever get resolved?
Fortunately it seems those miltantly against v6 nat got pushed out of the netfilter team and It looks like v6 nat support got accepted into linux.
Obviously the previous situation where public v4 IPs were virtually free and easy to get had to end.
Over the next few years people will have to evaluate what applications can use v6, what applications can use private v4 and what applications really need and can justify the cost of public v4.
ranges were given out like candy to anyone who asked in the early days of the web. Corporations, Government and Academics made a land grab because they were the only people who could use the resource at the time.
Remember in the early days of the internet there was only Class A, Class B and Class C (equivilent to /8, /16 and /24 nowadays), so if you were too big for a class C then you got a class B and if you were too big for a class B then you got a class A. This lead to many allocations being far bigger than they actually needed to be.
I've heard that Glasgow Uni has a /8 that's never had more than 10 addresses exposed to the Internet.
Sounds like it was either a myth or it was given back years ago. I don't recall ever seeing them on the /8 allocation list.
I could be wrong here but I suspect changing the prefix that is allocated through a router to a LAN is far more likely to cause problems with systems continuing to try and use stale addresses than changing the IP that is allocated to the WAN side of a NAT.
Combine this with the fact that IPv6 prefixes of the size ISPs are likely to hand out (somewhere between /48 and /64 depending on how shitty they are) are plentiful compared to IPv4 addresses and I would expect more ISPs to adopt a "sticky IP"* policy with v6 than with v4.
*That is IPs don't normally change but the ISP doesn't make any particular commitment about them not changing.
Not really. Diamond is a crystaline form of carbon. The carbon itself isn't worth much so the value must be the form it is in.
When carbon is put under sufficient pressure it crystalises into diamond. However the vast majority of those diamonds end up small and/or full of flaws. These small flawed diamonds are cheap enough (whether formed naturally or aritficially) that we coat drills and cutting discs with them.
Gemstone diamonds are much rarer naturally and harder to produce artificially (though CVD is starting to make them easier to produce artificially).
hardware will move back to expensive labor markets,
The question is will the jobs move back with them or will they be moved back to highly automated factories that employ as few people as possible.
It seems to vary a lot. Some drives are so quiet you can't hear them unless you hold your ear to the machine. Others clunk like hell. Newer drives in general seem to be quieter than old drives
Indeed they can't, OTOH i'm not sure I want the government (or worse a consortium of representives of governments) designing my networks for me either.
That's still going to require NAT if they want to use less than one public IPv4 address per user. There may still be use cases for it, but I am not convinced.
True, you need a special NAT for the ISP end of ds-lite but you don't need to manage private v4 addresses in your acess network. This avoids the risk of address collision with your users private IP addresses and means that very large ISPS (like comcast who wrote the ds-lite spec) don't need to federate their access networks to conserve private v4 addresses. It also scales horizonally nicely (just route different users to different ds-lite AFTR units) and is easy to implient in clients (from the clients point of view it's just a tunnel providing a route to a NAT)
Can't comment on 4rd as I haven't looked through that spec myself.
In my opinion, if there's a legitimate way to make the vehicle move, there's a way to make the vehicle move. If you don't want the vehicle to move then you need to remove something from it that makes it move, preferably something that a thief wouldn't normally bring with them, like a coil wire on a vehicle with a distributor, or a fuel pump relay or ASD relay, or something like that. Come to think of it, one could probably relocate such a relay to the passenger compartment to allow one to use the relay itself like a key, removing it to immobilize the vehicle
Which works fine if it's just the one vehicle but if an automaker started doing that theives would just start carrying relays.
An incomplete list of people who will be affected:
1: admins/moderators of interactive websites who find it harder to identify/ban users because of the inevitable rise of ISP level NAT (granted this is already a problem to some degree but is likely to get much worse).
2: users hit by bans aimed at thier shared IP either because the website owner didn't know it was shared or because they decide that the collateral damage is acceptable.
3: users who use software that needs to accept incoming connections on packages that no longer receive a public v4 IP
4: users who need public v4 IPs for services they are hosting and see the prices rise to reflect the market value of IP addresses.
Because that's the only Internet connection they can offer soon.
Bullshit, they can offer you:
v4 only with a private IP and ISP level NAT
v6+v4 dual stack with a public v6 IP a private v4 IP and ISP level v4 NAT
v6 with a public v6 IP and ds-lite
v6 with a public v6 IP with NAT64 and DNS64
They will also be able to offer public v4 as a premium service once they push their bottom tier of users onto one of the above.
Remember your ISP's existing v4 IPs won't dissapear, they will just have to reduce the average number used per customer (or buy IP addresses on the market) if they want to increase the number of customers they serve.
Any ISP/corp that doesn't want to pay for IPv4 addresses for their public services and pay for some mechanism to allow their users to access servers on the IPv4 internet is free to sit back and watch as they lose the ability to sell to many of their customers and buy from many of their suppliers over the internet. Of course by chosing this path they have a good chance of going broke (or getting overridden in a shareholder revolt) sooner rather than later but it is their choice to make.
Now personally I will offer any services i'm responsible for v6 where the provider offers it because I belive it is the right thing to do but i'm not going to pretend there is any strong reason for a buisness to do so.
The trouble is there has been a chicken and egg problem.
The internet is mostly a network of buisnesses (with the occasional academic network thrown in). Some of those buisnesses sell service to other buisnesses and consumers, some just use it to support their main buisness.
There is basically no benefit and significant cost to an buisness in deploying dual stack while v6 only nodes are basically unheard of.
You can't really deploy v6 only nodes while there are a significant number of v4 only nodes*
So for each individual buisness (whether ISP or user) that makes up the internet the rational thing to do was to sit and wait and hope others would take the costs of being early adoptors of IPv6.
*unless you use something like NAT64 or DS-lite but if nodes were ok with being behind those they would probablly also be ok with being behind NAT44 so the only real benefit to doing it would be if your internal network was massive.
You are a bit behind the times, there is a service for ships on the high seas that offers borderline broadband data rates.
http://www.inmarsat.com/Services/Maritime/FleetBroadband/default.aspx
At $18.00 / MByte you'd have to be pretty rich to use it for anything but essential communications though.
Smart meters have the advantage of being on-grid pretty much by definition. On-grid electricty is cheap.
off-grid small scale electricity is far more expensive. You either use primary batteries (only practical for very low consumption levels) or you use rechargable batteries and try to find some way to recharge them (not cheap to set up).
Radio losses are a bitch, only a tiny fraction of the original energy is left after a typical radio path.
So if you try and power your device over radio you end up with a tiny fraction of the original energy, then losses in conversion and then a tiny fraction of what is left after those losses getting back to the base-station. I'ts just not practical except for very short distance links to very low power devices (think: rfid tags).
Afaict supermarkets do it by leaving enough space arround the door that being near the door implies wanting to go through the door. Which works ok for them but is far too space-consuming for more general use.
The ideal RNG collects as much entropy from the real world as there is information in it's output. Second best is a cryptographically secure PRNG. To be cryptographically secure given an arbitary sized sample of the outut it must be computationally infeasible to predict the next bit with an accuracy greater than random chance. This requires both an algorithm that is resistant to reversal and sufficient seed data and internal state to prevent brute forcing of the random number genertor's state.
A lot of such devices are built round PCs (some use special embedded form factor PCs, others just have a normal PC tower sitting inside them) despite them being overkill and not particularly reliable. I guess it's because windows devs are easier to find than devs who can handle an arm linux board.
I do wonder why an i7 though, a celeron would be more than sufficent.