No... what I meant by that was that frequently the way v6 to v4 works is by mapping the space onto a subnet.
For example if my address is 1122:3344:5566:7788:9900:AABB:CCDD:EEFF my subnet is 1122:3344:5566:7788:: and so I might do something like have the entire v4 address space at 1122:3344:5566:7788::1100:0000::
Which is an example of (from the v6 perspective) having 4 billion address inside my subnet. As far as shortage the smallest possible subnet is the size of the internet squared. How can anyone have problems with a hierarchy this big?
I think you are thinking of subnetting in too much of a v4 sense. I think the definition has changed to
first 64 bits = public routing scheme (i.e. regulated) second 64 bits = private routing scheme (unregulated)
So for the large organization the question is whether they want to use normative routing or not.
Yes, exactly after 5 years they are still rolling it out and there have been hiccups. That's what it means for a project to be hard. And that rollout is for home and small business the easiest.
And yes A10 is working with something hard. That's what it is going to be like for enterprise.
___
If you want evidence I gave you evidence to try on your own ISP.
End-users prepared: pretty much (all major operating systems, and all non-shit routers for the last several years)
And their applications?
ISPs prepared: yep, more so than most.
No they aren't, they don't even claim to be. Most ISPs haven't done the work and will find that quite a lot of equipment needs to change. More importantly they don't have the support services in place for various configurations of dual stacks on their side that they are going to be needed to support customers who have transitioning difficulties. That's why even the most foreword thinking ISPs are mainly biting off experiments with consumer and small business.
Your ISP has v6, terrific. Ask them what sorts of QoS services they get when intermixing v6 and v4 traffic both having SIP or video on an MPLS.
Right now the most foreword thinking ISPs are able to handle the simplest clients with the simplest needs.
Backbon infrastructure: Getting there, it'll be ready by the time it's actually needed.
I'd agree.
I'd love for you to provide some of doomsday talk from the "people engaged in v6".
Sure. Listen to any of Comcast's talks. They were floored how many of their management applications used IPv4 related ideas in their logic. How much different routing strategies for IPv4 vs IPv6 were built into their failover plans. They also are transitioning to v6 internally and are finding that many of their administrative applications and hardware don't support v4 nor have obvious v4 replacement.
You want someone further out. Federal agencies. They found most of the equipment they use doesn't support IPv6. So to be compliant they often would have to buy something that is uncompliant in other ways.
A10 which specialized in dual stack work. They've found that most applications including things like server management systems are unable to cope with dual stack transitions.
I don't follow. First off your main post is written in terms of bytes when I think you mean bits. There are only 16 bytes to hand out total, I don't think you want all of them for the global prefix. Regardless, for ISPs the hierarchy is already there for routing the first 64 bits give you 64 different levels. Similarly for subnets, you can choose to route further out. The system allows for a 128 level hierarchy theoretically.
a') I'm not saying how bad the problem is. Right now the problems are still fairly minor the only thing really breaking is geo location on websites with respect to cell phones. What I was saying is how bad the problem will be.
b') No actually all those people behind the scenes doing the work are reporting it has been difficult. It hasn't gone well for them at all. All the people who have started that are not carriers are reporting it very complex. Generally 5-7 years of focused attention.
c') I'm glad your carrier has given you an v6 subnet. But in terms of transitioning difficulties that doesn't prove anything. It isn't complex to throw a server on v6 there have been some servers on v6 nets for 20 years. The fact that you are unaware of v6 problems most likely means your servers don't have the sorts of issues that create transitioning difficulties.
I think maybe you should actually hear from people engaged in v6 transitioning before blithely talking about how smoothly it is going for all of them and how easy it was.
I don't think so. The GP was saying that nothing much happen. Instead costs would rise. Similar to how people in the North East are being transitioned from home heating oil to Natural Gas. So he had a situation where
(a) IPv4 works well for a long time (b) IPv6 works well but their are transitioning costs for individuals (c) The costs for staying on IPv4 gradually increase which leads to people smoothly transitioning over to IPv6 on the basis of cost.
What I was saying was
(a') IPv4 is already starting to fail a little and that will get much worse in many ways over the next 5 years. (b') IPv6 will be a difficult transition and lots of stuff will break. (c') There will be no smooth transition at this point. We've waited too long. It will be blocky with non gradual sudden shifts as carriers move huge blocks of people over and change policies.
The cost of the addresses are going to be a minor factor except at the carrier, ISP level where the desire free them up will lead them to shift over homes and small businesses.
The consumers are being moved over to gateways. That will break geolocation on publicly facing websites. The next thing is v6 resources and security via. tunneling.
We don't need to be that extreme. There is going to be tons of legacy code that only supports v4. We are going to want that to keep working for many years.
At the same time we are going to want to move the bulk of the world to v6. The shortage is creating the right mix of incentives and no chaos. We don't need to criminalize routing.
The scarcity is what is finally getting carriers and ISPs to do the billions of dollars worth of work to make v6 possible. The scarcity is what is going to move consumers off v4 to v6. I wish they would be more scarce.
Yes they are going to do that. The carriers don't want carrier based NAT. The regulators don't want carrier based NAT. Consumers will be the easiest to switch and will free up tons of v4 addresses for business users.
You are forgetting that you can split IPv4 that finely. Router table fragmentation is already a serious problem. There are also problems of partial transitioning. Things like geo location fail if your customers are using v6 to v4 gateways. Things like session maintenance also fail if v4 addresses are expensive and carriers pool them more tightly. Public facing websites are already starting to fail that will get worse.
The next area is security. v4 security equipment doesn't understand v6 tunnelled traffic. As resources are available on v6 and carriers will move v4 traffic you'll start seeing lots of tunnels.
So no, this will be very messy. v4 will have substantial functional problems by 2015 with companies that aren't ready for the transition.
Looking at tables of meaningful IPv6 address is just as easy as IPv4. Most of the digits don't change so you treat the whole block as a big graphic and look at the few meaningful digits.
What you are describing is pretty much how v4 to v6 gateways work. But that to happen locally, as part of your own subnet it can't be done globally.
1) How does a v4 device respond to a request from a v6 device with an address that isn't 0001? 2) There are no routing tables how does a 0001... get delivered everywhere in the world?
I wouldn't use windows networking at all. Just file share using NFS and don't use Windows protocols. Most locked down windows machines can use an NFS mount.
If you want to be cute. Grab any piece of junk machine and set Samba as a Primary Domain Controller for your home network with the same name as the domain controller your laptop wants. You'll trick the laptop out.
The comment about/etc/hosts is right. That's what I use for my home network. What IPv6 offers you for your home is the end of having an external and an internal numbering scheme. Every box has its own IP, and you can setup the scheme in a way that makes sense for total addressability. Which means you can use someone else's external DNS if you like.
First off because all the tricks other than carrier level NAT will only buy a few extra months.
But also, because ARIN wants to encourage the switch to IPv6. They worked hard in the 1990s to buy extra years and the carriers the ISPs and the corporations wasted those extra years doing nothing. No one is going to do anything until there is a serious shortage.
As far as I can tell LMDE is targeted to desktop users as well, just as Mint was designed as a slight improvement over Ubunutu, LMDE is designed for those who run Debian Testing. The Mint desktop, the codecs and the software that Debian won't touch are the big advantages over just running Debian Testing all of which effect desktops. So installing Skype on LMDE is not a hassle.
The big advantage LMDE has over Mint (using Ubuntu) is they don't pick up all the heavy weight add ons of Ubuntu. At startup that amounts to 30m of RAM less stuff. With a light desktop LMDE is still usable on a 128m machine and even with Mint on a 512m machine. And there are speed differences as well. Potentially Mint might make the jump to LMDE and become an alternative to Ubuntu rather than just Ubuntu with a few upgrades.
As for what they do with other desktops... I think it makes sense to release everything. They don't do anything complex for the other DEs so I don't know why they can't just roll them out on the same DVD as options when they release the main distribution. That being said they don't seem to be treating those as casual rollouts but rather do additional work.
Let me correct my previous answer. It appears the XFCE... (excluding KDE) are grouped with the LMDE edition which is based on Debian Testing not Ubuntu. Essentially Mint has two totally different distributions Ubuntu based for Gnome variants and KDE and Debian based LMDE which is a rolling distribution.
Yep they do. Main page already has the 8 listed for version 13: http://www.linuxmint.com/release.php?id=18 AFAIK the last time they put out an XFCE was version 9, when they also had a Fluxbox. I assume that's been completely dropped. The version 13 user's guide is written in terms of MATE. Are you sure they aren't going Gnome variants only?
Mint isn't targeting the masses. Mint seems to be targeting the disgruntled Ubuntu users, so people with a 1/2 dozen years of Unix experience who like the features. They all know whether they have a 32 or 64 bit system.
You don't run a local copy of quickbooks off a remote database, you run quickbooks or another accounting solution in the cloud and pass HTML back and forth. More likely you break the solution up a bit.
Slashdot Mirror
No... what I meant by that was that frequently the way v6 to v4 works is by mapping the space onto a subnet.
For example if my address is 1122:3344:5566:7788:9900:AABB:CCDD:EEFF my subnet is
1122:3344:5566:7788:: and so I might do something like have the entire v4 address space at 1122:3344:5566:7788::1100:0000::
Which is an example of (from the v6 perspective) having 4 billion address inside my subnet. As far as shortage the smallest possible subnet is the size of the internet squared. How can anyone have problems with a hierarchy this big?
I think you are thinking of subnetting in too much of a v4 sense. I think the definition has changed to
first 64 bits = public routing scheme (i.e. regulated)
second 64 bits = private routing scheme (unregulated)
So for the large organization the question is whether they want to use normative routing or not.
Yes, exactly after 5 years they are still rolling it out and there have been hiccups. That's what it means for a project to be hard. And that rollout is for home and small business the easiest.
And yes A10 is working with something hard. That's what it is going to be like for enterprise.
___
If you want evidence I gave you evidence to try on your own ISP.
End-users prepared: pretty much (all major operating systems, and all non-shit routers for the last several years)
And their applications?
ISPs prepared: yep, more so than most.
No they aren't, they don't even claim to be. Most ISPs haven't done the work and will find that quite a lot of equipment needs to change. More importantly they don't have the support services in place for various configurations of dual stacks on their side that they are going to be needed to support customers who have transitioning difficulties. That's why even the most foreword thinking ISPs are mainly biting off experiments with consumer and small business.
Your ISP has v6, terrific. Ask them what sorts of QoS services they get when intermixing v6 and v4 traffic both having SIP or video on an MPLS.
Right now the most foreword thinking ISPs are able to handle the simplest clients with the simplest needs.
Backbon infrastructure: Getting there, it'll be ready by the time it's actually needed.
I'd agree.
I'd love for you to provide some of doomsday talk from the "people engaged in v6".
Sure. Listen to any of Comcast's talks. They were floored how many of their management applications used IPv4 related ideas in their logic. How much different routing strategies for IPv4 vs IPv6 were built into their failover plans. They also are transitioning to v6 internally and are finding that many of their administrative applications and hardware don't support v4 nor have obvious v4 replacement.
You want someone further out. Federal agencies. They found most of the equipment they use doesn't support IPv6. So to be compliant they often would have to buy something that is uncompliant in other ways.
A10 which specialized in dual stack work. They've found that most applications including things like server management systems are unable to cope with dual stack transitions.
I don't follow. First off your main post is written in terms of bytes when I think you mean bits. There are only 16 bytes to hand out total, I don't think you want all of them for the global prefix. Regardless, for ISPs the hierarchy is already there for routing the first 64 bits give you 64 different levels. Similarly for subnets, you can choose to route further out. The system allows for a 128 level hierarchy theoretically.
a') I'm not saying how bad the problem is. Right now the problems are still fairly minor the only thing really breaking is geo location on websites with respect to cell phones. What I was saying is how bad the problem will be.
b') No actually all those people behind the scenes doing the work are reporting it has been difficult. It hasn't gone well for them at all. All the people who have started that are not carriers are reporting it very complex. Generally 5-7 years of focused attention.
c') I'm glad your carrier has given you an v6 subnet. But in terms of transitioning difficulties that doesn't prove anything. It isn't complex to throw a server on v6 there have been some servers on v6 nets for 20 years. The fact that you are unaware of v6 problems most likely means your servers don't have the sorts of issues that create transitioning difficulties.
I think maybe you should actually hear from people engaged in v6 transitioning before blithely talking about how smoothly it is going for all of them and how easy it was.
I don't think so.
The GP was saying that nothing much happen. Instead costs would rise. Similar to how people in the North East are being transitioned from home heating oil to Natural Gas. So he had a situation where
(a) IPv4 works well for a long time
(b) IPv6 works well but their are transitioning costs for individuals
(c) The costs for staying on IPv4 gradually increase which leads to people smoothly transitioning over to IPv6 on the basis of cost.
What I was saying was
(a') IPv4 is already starting to fail a little and that will get much worse in many ways over the next 5 years.
(b') IPv6 will be a difficult transition and lots of stuff will break.
(c') There will be no smooth transition at this point. We've waited too long. It will be blocky with non gradual sudden shifts as carriers move huge blocks of people over and change policies.
The cost of the addresses are going to be a minor factor except at the carrier, ISP level where the desire free them up will lead them to shift over homes and small businesses.
The consumers are being moved over to gateways. That will break geolocation on publicly facing websites.
The next thing is v6 resources and security via. tunneling.
We don't need to be that extreme. There is going to be tons of legacy code that only supports v4. We are going to want that to keep working for many years.
At the same time we are going to want to move the bulk of the world to v6. The shortage is creating the right mix of incentives and no chaos. We don't need to criminalize routing.
Why would we want even more fragmented routing tables? What's the upside of doing all this work to hunt down addresses for the v4 space?
The scarcity is what is finally getting carriers and ISPs to do the billions of dollars worth of work to make v6 possible. The scarcity is what is going to move consumers off v4 to v6. I wish they would be more scarce.
No your ISP gives you an entire /64 subnet. You don't get an entire v4 internet of v4 internets worth of addresses.
Yes they are going to do that. The carriers don't want carrier based NAT. The regulators don't want carrier based NAT. Consumers will be the easiest to switch and will free up tons of v4 addresses for business users.
You are forgetting that you can split IPv4 that finely. Router table fragmentation is already a serious problem. There are also problems of partial transitioning. Things like geo location fail if your customers are using v6 to v4 gateways. Things like session maintenance also fail if v4 addresses are expensive and carriers pool them more tightly. Public facing websites are already starting to fail that will get worse.
The next area is security. v4 security equipment doesn't understand v6 tunnelled traffic. As resources are available on v6 and carriers will move v4 traffic you'll start seeing lots of tunnels.
So no, this will be very messy. v4 will have substantial functional problems by 2015 with companies that aren't ready for the transition.
One thing that creates 4 billion users is IPv4 to IPv6 done locally. You have to throw in the entire IPv4 address space into your subnet.
But excluding that. We aren't likely to use the first 64 bytes either. The v6 space is massive.
Looking at tables of meaningful IPv6 address is just as easy as IPv4. Most of the digits don't change so you treat the whole block as a big graphic and look at the few meaningful digits.
What you are describing is pretty much how v4 to v6 gateways work. But that to happen locally, as part of your own subnet it can't be done globally.
1) How does a v4 device respond to a request from a v6 device with an address that isn't 0001?
2) There are no routing tables how does a 0001... get delivered everywhere in the world?
I wouldn't use windows networking at all. Just file share using NFS and don't use Windows protocols. Most locked down windows machines can use an NFS mount.
If you want to be cute. Grab any piece of junk machine and set Samba as a Primary Domain Controller for your home network with the same name as the domain controller your laptop wants. You'll trick the laptop out.
The comment about /etc/hosts is right. That's what I use for my home network. What IPv6 offers you for your home is the end of having an external and an internal numbering scheme. Every box has its own IP, and you can setup the scheme in a way that makes sense for total addressability. Which means you can use someone else's external DNS if you like.
First off because all the tricks other than carrier level NAT will only buy a few extra months.
But also, because ARIN wants to encourage the switch to IPv6. They worked hard in the 1990s to buy extra years and the carriers the ISPs and the corporations wasted those extra years doing nothing. No one is going to do anything until there is a serious shortage.
As far as I can tell LMDE is targeted to desktop users as well, just as Mint was designed as a slight improvement over Ubunutu, LMDE is designed for those who run Debian Testing. The Mint desktop, the codecs and the software that Debian won't touch are the big advantages over just running Debian Testing all of which effect desktops. So installing Skype on LMDE is not a hassle.
The big advantage LMDE has over Mint (using Ubuntu) is they don't pick up all the heavy weight add ons of Ubuntu. At startup that amounts to 30m of RAM less stuff. With a light desktop LMDE is still usable on a 128m machine and even with Mint on a 512m machine. And there are speed differences as well. Potentially Mint might make the jump to LMDE and become an alternative to Ubuntu rather than just Ubuntu with a few upgrades.
As for what they do with other desktops... I think it makes sense to release everything. They don't do anything complex for the other DEs so I don't know why they can't just roll them out on the same DVD as options when they release the main distribution. That being said they don't seem to be treating those as casual rollouts but rather do additional work.
Let me correct my previous answer. It appears the XFCE ... (excluding KDE) are grouped with the LMDE edition which is based on Debian Testing not Ubuntu. Essentially Mint has two totally different distributions Ubuntu based for Gnome variants and KDE and Debian based LMDE which is a rolling distribution.
Yep they do. Main page already has the 8 listed for version 13: http://www.linuxmint.com/release.php?id=18
AFAIK the last time they put out an XFCE was version 9, when they also had a Fluxbox. I assume that's been completely dropped. The version 13 user's guide is written in terms of MATE. Are you sure they aren't going Gnome variants only?
Mint isn't targeting the masses. Mint seems to be targeting the disgruntled Ubuntu users, so people with a 1/2 dozen years of Unix experience who like the features. They all know whether they have a 32 or 64 bit system.
2 desktops * 2 bit sizes * with/without codecs = 8. However there are also 2 OEM version so 10 total.
You don't run a local copy of quickbooks off a remote database, you run quickbooks or another accounting solution in the cloud and pass HTML back and forth. More likely you break the solution up a bit.