We don't need crazy ideas the major regulatory agencies are onboard. And their idea is simple. They are going to allow v4 addresses to be sold. So ISPs which have tons of v4 addresses now can convert them to cash for people who don't convert and pool up their addresses.
They are going to be allocated a full/64. And no in 20 years we won't be close to exhausting the first 64 bits. That's the size of the internet squared.
Give 1,000,000 full/64 subnets to every man woman and child on the planet and another million to every network enabled device in existence and you are still wouldn't be close to exhausting the v6 address space.
We have no idea if people would want to have all sorts of easy to use global stuff it became obvious. Your coffee machine is running a start server and when you set your alarm on your cell phone to wakeup it sets the coffee to start 10 minutes before.
I'm not sure how this is a problem. This is defined in v6.
I can setup whole virtualized subnets with translation on my routing equipment. That's how 6to4 works.
. With v4 and NAT this is easy you run your internal network on private IPs, use NAT to access the internet and use tunnels to communicate between sites. How a subnet is addressed and how it connects to the internet can be completely inde[pendent.]
Whereas without NAT if you want to change how a portion of your network connects to the internet you have to readdreess it and while you may try to enforce use of private addresses for internal communication in practice it's likely that some stuff will get inadvertantly set up to point to public IPs.
So let me give you the v6 equivalent.
I have say 3 sites with their own/64 subnets and 10 machines on each. I mirror each others subnets and setup static route conversions on the 3 routers. Done.
So for example sites A, B and C look like A] 1122:3344:5566:7788:0AXX:: B] 1234:5678:90AB:CDEF:OBXX:: C] 2345:6789:0ABC:DEF1:OCXX::
I think ISPs will be required to give customers a/64. You either give a/0 a/1 (P2P) or a/64. There are no variable sized subnets in v6. The routing tables are much simpler which makes lower latency possible.
Sure but that will be very late in the process. Companies are going to spend years on dual stack. When IPv4 addresses are say $250 / yr / each companies are going to sell them fast.
Lots of Android devices are on v6 already, possibly the majority are. Android is fine. Mac, Windows and Linux (assuming last 10 years OS) should be fine. And your ISP will provide you with a modem / home network.
And your/. user number is far too low to be thinking you should be able to use tech equipment forever.
The vast majority of home users get their router from their ISP. Compared to the cost of reconfiguring their entire network at every stage the price of the home router ain't much.
I agree though this should have been fixed. So do the carriers themselves, most are already going to be forcing this issue this year.
Sure there is. The market rate for IPs is now $9 each (Microsoft purchase). If someone offered to buy MIT's addresses it might be worth $100m to change.
Now assume the situation is getting worse and we re up at $40 / each.
What could have been a smooth transition is now going to be a massive messy expensive rush crisis of a transition because ISPs couldn't see past their next quarter profit results
When I was a kid we had this thing called government regulation which was designed to move companies to act in the common interest even when not acting was in their individual interest.
Not for very long. And redistributing addresses more evenly means making routing tables more complex and they are already so complex that they are introducing latency.
The vast, VAST majority of us only need IPv6 on the Internet. We are perfectly fine with IPv4 internally.
No you aren't. v4 security devices don't understand what's happening with v6 traffic tunneling. So in that kind of setup you end up with lots of v6 tunnels running over v4 which you can't control.
That is exactly what the carriers are planning for. They are going to pool all their phone customers so they have static IPV6 addresses, and share a pool of IPV4 for when they are on the network. They then are going to do the same thing with home users, except that home users will get an entire/64 subnet for all their devices. And the entire subnet will have to get a pooled address.
Then if that gets too crowded next step is pooling at the port level using NAT.
They minimum sized subnet is/64. Think about taking each IP address on the internet and replacing it with an entire internet. Now imagine doing a port scan of that.
Ah you missed the key strategy security. Most IPV4 OSes (including Windows and OSX) setup IPV6 tunnels over IPV4. IPV4 security stuff doesn't understand tunnelled traffic, they just ignore it. Ignoring it ain't blocking it.
Mary Jane in accounting getting onto IPV6 Facebook and Sam Smith legal getting on his favorite porn site at work is what's going to force networks to make the switch.
SunOS was always BSD based, there was a complete switchover in OS. Anyway, Solaris came out for end users in 1992. I certainly was using SunOS as a primary in 93 and still had to use it a bit in 95/96 still. But.... the time when the switchover to LInux happened for Solaris users was '94. Remember the original Linux came from the hobbyist community (Minix). Solaris in '94-95 and Windows people (and some SCO) around 97.
If we are going to tell personal stories... I'm a good example. I remember being excited about 386BSD. I had seriously considered buying Dell Unix (a OEM version of SCO) but a lot more for the computer and an extra $900 for the OS... However the reports about 386BSD were it was too raw. I was an end user of Unix not a system admin. I needed easy installation and a full software suite. That didn't even start to exist until 1995. I learned Unix administration from Linux. I think that or slightly more is typical of the Solaris users that switched. I certainly knew about BSD. I heard good things about BSD. I even tried BSD. I believe my first open source Unix was a BSD. But the Unix I actually got to do what I wanted a Unix for was RedHat.
Even today, Linuxes do a much much better job of bringing in new people than BSDs.
And it happened at exactly the wrong time, when i386 systems were growing in popularity, and people wanted some Unix-like OS to run on it, and really wanted it for free.
Not really. If you look at the usage totals during those crucial years they weren't very high. That is not and was not a dominant marketshare. The issue was simple, the early adopters of Linux were from two groups:
a) Solaris users, who knew Sys-V b) Windows power users that required lots of hand holding and education.
The BSD community wasn't interested in supporting either of those communities.
The issue is whether they used not why didn't they use it. Asserting that you can't think of a good reason is not proof they didn't, especially when they were very explicit they did a complete rewrite.
And the reason they rewrote it for Vista from scratch is that Microsoft has created certain intermediate layers in their networking stack that offer features which are Windows specific. For example an advanced exception handling mechanism which goes beyond standard TCP/IP
There's a reason Apple used FreeBSD as their basis for OSX and not Linux.
Yeah the reason was that NeXT was founded in 1985, the very first Linux was released in 1991. And some ideas were taken from Apple Lisa, where design started in 1978.
Can you imagine when essentially all of the Windows server applications are going to work well enough GUI less that Microsoft is going to feel comfortable removing the GUI? Microsoft is not Apple. They don't announce a new direction at WWDC and all the developers fall in line. And they aren't willing to make large breaking changes to ram through new ways of doing things.
Windows may be moving towards more stuff being able to be run without a GUI. I doubt there are going to be any major apps without a GUI interface. I think it is essentially impossible there will be a GUI-less version of Windows server within a generation.
Mandriva is not 3rd tier. That is Mandrake which goes 98 and connectiva to 95. Connectiva was part of the whole Open Linux initiative which history had turned out differently could have been the 1st tier distribution.
Mandriva wasn't born from Mandrake code, Mandrake acquired other distribution companies like Lycos and added their stuff.
Slashdot Mirror
We don't need crazy ideas the major regulatory agencies are onboard. And their idea is simple. They are going to allow v4 addresses to be sold. So ISPs which have tons of v4 addresses now can convert them to cash for people who don't convert and pool up their addresses.
They are going to be allocated a full /64. And no in 20 years we won't be close to exhausting the first 64 bits. That's the size of the internet squared.
Give 1,000,000 full /64 subnets to every man woman and child on the planet and another million to every network enabled device in existence and you are still wouldn't be close to exhausting the v6 address space.
We have no idea if people would want to have all sorts of easy to use global stuff it became obvious. Your coffee machine is running a start server and when you set your alarm on your cell phone to wakeup it sets the coffee to start 10 minutes before.
I'm not sure how this is a problem. This is defined in v6.
I can setup whole virtualized subnets with translation on my routing equipment. That's how 6to4 works.
So let me give you the v6 equivalent.
I have say 3 sites with their own /64 subnets and 10 machines on each. I mirror each others subnets and setup static route conversions on the 3 routers. Done.
So for example sites A, B and C look like
A] 1122:3344:5566:7788:0AXX::
B] 1234:5678:90AB:CDEF:OBXX::
C] 2345:6789:0ABC:DEF1:OCXX::
And on A's site I represent B's addresses as:
1122:3344:5566:7788:OBXX
and the router translates. That's perfectly safe.
Once routing tables are simplified you are going to start having protocols that are more sensitive to latency.
It is likely that NAT may ever work right on v6 in practice. In theory you can of course do anything.
I think ISPs will be required to give customers a /64. You either give a /0 a /1 (P2P) or a /64. There are no variable sized subnets in v6. The routing tables are much simpler which makes lower latency possible.
Sure but that will be very late in the process. Companies are going to spend years on dual stack. When IPv4 addresses are say $250 / yr / each companies are going to sell them fast.
Phones are the first to go to v6. Many smartphones are already on v6.
Lots of Android devices are on v6 already, possibly the majority are. Android is fine. Mac, Windows and Linux (assuming last 10 years OS) should be fine. And your ISP will provide you with a modem / home network.
And your /. user number is far too low to be thinking you should be able to use tech equipment forever.
The vast majority of home users get their router from their ISP. Compared to the cost of reconfiguring their entire network at every stage the price of the home router ain't much.
I agree though this should have been fixed. So do the carriers themselves, most are already going to be forcing this issue this year.
Sure there is. The market rate for IPs is now $9 each (Microsoft purchase). If someone offered to buy MIT's addresses it might be worth $100m to change.
Now assume the situation is getting worse and we re up at $40 / each.
When I was a kid we had this thing called government regulation which was designed to move companies to act in the common interest even when not acting was in their individual interest.
Sure.
Imagine two subnets.
(A) has static IP addresses and a firewall configured to block all unrequested traffic.
(B) has NAT.
now lets handle fixed addresses
(C) is like A but has a few preset open ports on some machines.
(D) is like B but has some static routes.
What's the difference?
Not for very long. And redistributing addresses more evenly means making routing tables more complex and they are already so complex that they are introducing latency.
No you aren't. v4 security devices don't understand what's happening with v6 traffic tunneling. So in that kind of setup you end up with lots of v6 tunnels running over v4 which you can't control.
That is exactly what the carriers are planning for. They are going to pool all their phone customers so they have static IPV6 addresses, and share a pool of IPV4 for when they are on the network. They then are going to do the same thing with home users, except that home users will get an entire /64 subnet for all their devices. And the entire subnet will have to get a pooled address.
Then if that gets too crowded next step is pooling at the port level using NAT.
They minimum sized subnet is /64. Think about taking each IP address on the internet and replacing it with an entire internet. Now imagine doing a port scan of that.
Ah you missed the key strategy security. Most IPV4 OSes (including Windows and OSX) setup IPV6 tunnels over IPV4. IPV4 security stuff doesn't understand tunnelled traffic, they just ignore it. Ignoring it ain't blocking it.
Mary Jane in accounting getting onto IPV6 Facebook and Sam Smith legal getting on his favorite porn site at work is what's going to force networks to make the switch.
Terry
SunOS was always BSD based, there was a complete switchover in OS. Anyway, Solaris came out for end users in 1992. I certainly was using SunOS as a primary in 93 and still had to use it a bit in 95/96 still. But.... the time when the switchover to LInux happened for Solaris users was '94. Remember the original Linux came from the hobbyist community (Minix). Solaris in '94-95 and Windows people (and some SCO) around 97.
If we are going to tell personal stories... I'm a good example. I remember being excited about 386BSD. I had seriously considered buying Dell Unix (a OEM version of SCO) but a lot more for the computer and an extra $900 for the OS... However the reports about 386BSD were it was too raw. I was an end user of Unix not a system admin. I needed easy installation and a full software suite. That didn't even start to exist until 1995. I learned Unix administration from Linux. I think that or slightly more is typical of the Solaris users that switched. I certainly knew about BSD. I heard good things about BSD. I even tried BSD. I believe my first open source Unix was a BSD. But the Unix I actually got to do what I wanted a Unix for was RedHat.
Even today, Linuxes do a much much better job of bringing in new people than BSDs.
Not really. If you look at the usage totals during those crucial years they weren't very high. That is not and was not a dominant marketshare. The issue was simple, the early adopters of Linux were from two groups:
a) Solaris users, who knew Sys-V
b) Windows power users that required lots of hand holding and education.
The BSD community wasn't interested in supporting either of those communities.
Almost all the protocols predate the GPL and certainly predate the rise of open source. This might be a better question in 20 years. But...
WebDAV is the big one. Gnutella is a less important one.
The issue is whether they used not why didn't they use it. Asserting that you can't think of a good reason is not proof they didn't, especially when they were very explicit they did a complete rewrite.
And the reason they rewrote it for Vista from scratch is that Microsoft has created certain intermediate layers in their networking stack that offer features which are Windows specific. For example an advanced exception handling mechanism which goes beyond standard TCP/IP
Yeah the reason was that NeXT was founded in 1985, the very first Linux was released in 1991. And some ideas were taken from Apple Lisa, where design started in 1978.
Can you imagine when essentially all of the Windows server applications are going to work well enough GUI less that Microsoft is going to feel comfortable removing the GUI? Microsoft is not Apple. They don't announce a new direction at WWDC and all the developers fall in line. And they aren't willing to make large breaking changes to ram through new ways of doing things.
Windows may be moving towards more stuff being able to be run without a GUI. I doubt there are going to be any major apps without a GUI interface. I think it is essentially impossible there will be a GUI-less version of Windows server within a generation.
Mandriva is not 3rd tier. That is Mandrake which goes 98 and connectiva to 95. Connectiva was part of the whole Open Linux initiative which history had turned out differently could have been the 1st tier distribution.
Mandriva wasn't born from Mandrake code, Mandrake acquired other distribution companies like Lycos and added their stuff.