IPv4 is over. What will happen? IPv6?
The last blocks of / 8 are printed , the Internet is boiling, experts argue, industry leaders are sharpening their teeth on each other, anticipating the turn of the struggle for markets in the conditions of a new reality.
This story resembles at least three different stories at once:
Everyone blows in his tune (well, what else?)
Some, the ideologists of IPv6 and the manufacturers of telecommunications hardware, who are awaiting the expansion of markets, offer to drop everything immediately and start implementing v6. Others - most of them technical experts - are about the idea of a full IPv6 transition as crazy IANA crazy. Third - the majority of participants in the industry, including almost all Russian operators - are not at all steaming (see the paragraph above about the global financial crisis).
')
I will try to give the argument of each of the groups.
First, let's call them this way, the man in the street: operating services in providers, admins in companies and other ordinary IT specialists. It looks like they are most afraid of the appearance of IPv6-addresses. This is really an unusual and scary sight. In addition, their motivation is doubtful - why do they need it? Finally, people are just too lazy to understand something that is not very interesting, they do not really need, and in general.
Merchants, managers and owners of telecom operators, companies with large corporate networks, etc. - today in our country they have no incentives to invest in the introduction of IPv6 (however, this is discussed below).
Ideologues. Technically there really is no other real way out. Most ways to circumvent the shortage of IPv4-addresses (actually, NAT) - allows, at best, only to relieve the acute pain (not at all for free), delaying the need for surgery and the burden of history. “IPv6 is the worst of solutions, but there is no other,” say the ideologues. The most frostbitten ones paint an idyllic picture of a complete transition to the v6 of the entire Internet with tiny v4 islands, which NAT64 saves and back. This picture looks utopian, but there is no other one.
Vendors echo ideologues, vying with one another that their equipment is IPv6 ready. The only incident is that the news of the form “well, now exactly ready,” has been distributed from time to time for a dozen years.
Experts twist fingers at a temple and predict "an epoch of big NAT'a". Like, well, you citizens, what kind of circus? Firstly, a v6-ready router is not a tricky business. And show us, for example, v6-ready IPS, eh? Even with firewalls, not everything is so smooth.
Yes, and with routers, if you think about it. The size of hardware switching tables (FIB) for IPv6 in most boxes is several times smaller than for IPv4 - you can’t do anything special with this because the IPv6 address is 4 times longer. This is despite the fact that for v4 the size of the FIB is a problem. The complete routing table of the v4 Internet today is 350 thousand entries, and the number is growing exponentially. For IPv6 - just a little less than 4.5 thousand. Now imagine that the entire Internet got bogged down with IPv6 addresses a year later, and Internet routers need to keep 400 thousand v4 entries and the same number of v6. All this at speeds of tens of gigabits. The iron that can do this is very, very expensive, and few have it today.
Secondly. There are a lot of technologies and protocols that are tightly bound to IPv4. All control processes in MPLS protocols and applications today use IPv4, and implementations for IPv6 are still not widely supported, sometimes there are no standards for this, etc. That is, providers still cannot get away from IPv4 after a year or after all, in five.
Finally - and why are all providers (and mobile operators)? What is their economic motivation for implementing IPv6? Many of them have just built packet networks, put up with the ubiquity of IP, learned the word MPLS, and — hello, please go back into battle? What for? There are two and a half sites on the Internet that respond to IPv6, which also work on v4. Users are important that classmates work. Yes, they are. That is, there is no demand, there is no technological need. And why spend a lot of money on the introduction of this IPv6, if you can not? Is it possible?
Still not sure.
Most broadband access providers and (especially) mobile operators that provide Internet to individuals-subscribers now spend huge amounts of money on building NAT infrastructure. This very NAT is already strangling them today. They have no way out, of course, and for the time being they will increase these capacities, but even with naty sooner or later there will come a time when there will be little real IPv4 addresses. It's not about the notorious number of ports, which is unlikely to ever become a problem. The fact is that when too many users are broadcast to one address, this address begins to look like the source of the attack, and popular resources begin to block it.
In the field of the Internet for companies is still worse. Firstly, NAT is traditionally not popular there, because legal entities need considerable technological freedom: publishing their Internet services, forwarding various kinds of tunnels, etc. And the further, the more: companies try to be more and more independent from providers (why, in fact, the size of the routing table is growing so rapidly). That is, everyone wants to get their own address block and connect to several providers at once, change them as you please, etc.
These companies will ask for addresses and get to the server: “Sorry, only IPv6”, - and the companies will begin to put pressure on providers. Moreover, there is an important driver: for a part of non-public applications working via the Internet, such as client-banking, various kinds of tunnel VPN-networks, etc., it is not at all required for the rest of the world (favorite sites) to support IPv6.
And here there is one moment. Generally speaking, a properly built network, especially a backbone network, although aggregation-access too, is not too difficult to become v6-ready. A much bigger problem is a different kind of bodywork, such as billing, control systems, etc. And another problem, of course, is that there are not so many well-built networks in our country, much more collected on our knees. And their operators can be understood. How such a feature of the industry will show itself further is a very interesting question.
Forecast. The era of large NAT'a really will be. She, in fact, has already begun. Maybe someone still raskulachat a little bit or unpack the blocks at the end of the list, which will even delay the problem. But as long as no other proposed picture was offered, IPv6 will be there anyway. If only because it is reluctant to re-invent something fundamentally new. But it will be shameful not to be in IPv6 in three years, not earlier.
This story resembles at least three different stories at once:
- 2000 problem (does anyone remember this?)
- traditionally a sudden onset of winter for housing and communal services at the end of November,
- the global financial crisis: everyone was waiting for the bubbles to burst once, but no one understood where and why in this regard to run - perhaps somehow very, too big to fail, and all that.
Everyone blows in his tune (well, what else?)
Some, the ideologists of IPv6 and the manufacturers of telecommunications hardware, who are awaiting the expansion of markets, offer to drop everything immediately and start implementing v6. Others - most of them technical experts - are about the idea of a full IPv6 transition as crazy IANA crazy. Third - the majority of participants in the industry, including almost all Russian operators - are not at all steaming (see the paragraph above about the global financial crisis).
')
I will try to give the argument of each of the groups.
First, let's call them this way, the man in the street: operating services in providers, admins in companies and other ordinary IT specialists. It looks like they are most afraid of the appearance of IPv6-addresses. This is really an unusual and scary sight. In addition, their motivation is doubtful - why do they need it? Finally, people are just too lazy to understand something that is not very interesting, they do not really need, and in general.
Merchants, managers and owners of telecom operators, companies with large corporate networks, etc. - today in our country they have no incentives to invest in the introduction of IPv6 (however, this is discussed below).
Ideologues. Technically there really is no other real way out. Most ways to circumvent the shortage of IPv4-addresses (actually, NAT) - allows, at best, only to relieve the acute pain (not at all for free), delaying the need for surgery and the burden of history. “IPv6 is the worst of solutions, but there is no other,” say the ideologues. The most frostbitten ones paint an idyllic picture of a complete transition to the v6 of the entire Internet with tiny v4 islands, which NAT64 saves and back. This picture looks utopian, but there is no other one.
Vendors echo ideologues, vying with one another that their equipment is IPv6 ready. The only incident is that the news of the form “well, now exactly ready,” has been distributed from time to time for a dozen years.
Experts twist fingers at a temple and predict "an epoch of big NAT'a". Like, well, you citizens, what kind of circus? Firstly, a v6-ready router is not a tricky business. And show us, for example, v6-ready IPS, eh? Even with firewalls, not everything is so smooth.
Yes, and with routers, if you think about it. The size of hardware switching tables (FIB) for IPv6 in most boxes is several times smaller than for IPv4 - you can’t do anything special with this because the IPv6 address is 4 times longer. This is despite the fact that for v4 the size of the FIB is a problem. The complete routing table of the v4 Internet today is 350 thousand entries, and the number is growing exponentially. For IPv6 - just a little less than 4.5 thousand. Now imagine that the entire Internet got bogged down with IPv6 addresses a year later, and Internet routers need to keep 400 thousand v4 entries and the same number of v6. All this at speeds of tens of gigabits. The iron that can do this is very, very expensive, and few have it today.
Secondly. There are a lot of technologies and protocols that are tightly bound to IPv4. All control processes in MPLS protocols and applications today use IPv4, and implementations for IPv6 are still not widely supported, sometimes there are no standards for this, etc. That is, providers still cannot get away from IPv4 after a year or after all, in five.
Finally - and why are all providers (and mobile operators)? What is their economic motivation for implementing IPv6? Many of them have just built packet networks, put up with the ubiquity of IP, learned the word MPLS, and — hello, please go back into battle? What for? There are two and a half sites on the Internet that respond to IPv6, which also work on v4. Users are important that classmates work. Yes, they are. That is, there is no demand, there is no technological need. And why spend a lot of money on the introduction of this IPv6, if you can not? Is it possible?
Still not sure.
Most broadband access providers and (especially) mobile operators that provide Internet to individuals-subscribers now spend huge amounts of money on building NAT infrastructure. This very NAT is already strangling them today. They have no way out, of course, and for the time being they will increase these capacities, but even with naty sooner or later there will come a time when there will be little real IPv4 addresses. It's not about the notorious number of ports, which is unlikely to ever become a problem. The fact is that when too many users are broadcast to one address, this address begins to look like the source of the attack, and popular resources begin to block it.
In the field of the Internet for companies is still worse. Firstly, NAT is traditionally not popular there, because legal entities need considerable technological freedom: publishing their Internet services, forwarding various kinds of tunnels, etc. And the further, the more: companies try to be more and more independent from providers (why, in fact, the size of the routing table is growing so rapidly). That is, everyone wants to get their own address block and connect to several providers at once, change them as you please, etc.
These companies will ask for addresses and get to the server: “Sorry, only IPv6”, - and the companies will begin to put pressure on providers. Moreover, there is an important driver: for a part of non-public applications working via the Internet, such as client-banking, various kinds of tunnel VPN-networks, etc., it is not at all required for the rest of the world (favorite sites) to support IPv6.
And here there is one moment. Generally speaking, a properly built network, especially a backbone network, although aggregation-access too, is not too difficult to become v6-ready. A much bigger problem is a different kind of bodywork, such as billing, control systems, etc. And another problem, of course, is that there are not so many well-built networks in our country, much more collected on our knees. And their operators can be understood. How such a feature of the industry will show itself further is a very interesting question.
Forecast. The era of large NAT'a really will be. She, in fact, has already begun. Maybe someone still raskulachat a little bit or unpack the blocks at the end of the list, which will even delay the problem. But as long as no other proposed picture was offered, IPv6 will be there anyway. If only because it is reluctant to re-invent something fundamentally new. But it will be shameful not to be in IPv6 in three years, not earlier.
Source: https://habr.com/ru/post/113056/
All Articles