[Translation] Three reasons not to worry about the absence of IPv4 addresses, and three reasons to worry about it
IPv4 Internet addresses are near exhaustion. For most companies, the question is, how much will this affect them? Do I need to immediately transfer the entire infrastructure to IPv6 , or can I use other options? The following is a reflection on when to worry about it, and when — what is more important — is not worth it.
In February of this year, the Internet Assigned Numbers Authority ( IANA ) Internet Administration identified the last two pre-transition IPv4 addresses of the Asia-Pacific Region Address Space Administration (ARNIC). Somewhere in early August [2011], the four regional Address Space Administrations (ARINC, ARIN, RIPE and LACNIC) will begin to distribute their remaining blocks to the clients that are not reserved for the transition.
From this point on, the Internet will enter a transition phase, the outcome of which will be the widespread use of IPv6 . However, some experts see situations where IPv4 will remain valid for many years.
How destructive will the transition be? This is a big question to discuss. The range of predictions is from “this is not the biggest problem” to “this will be the end of the open Internet as we know it”; from “most people and companies will not notice anything but small roughness” to “inevitable gigantic and long-term costs”; and from “ IPv6 will work very well” to “ IPv6 will kill performance for many classes of applications until new generations of silicon come onto the scene.”
')
Who to believe?
As everyone develops his own sober plans for transition to IPv6 with his suppliers of equipment and communication services, we offer several topics for reflection.
Do not worry : This is the same as the “Problem 2000”. A lot of noise, and the almost complete lack of problems in the end.
Worry : On the other hand, when solving “Problems of the Year 2000”, a lot of working code was reworked, and some obsolete equipment was put into junk. In this case, you need to do quite a bit - just be ready. You will need to make changes to the code or port it to new systems in order to implement new APIs and move to IPv6 applications and databases (sometimes by using add-ons over IPv4 ). One thing brightens up this work - most of these systems are written in more relevant languages ​​compared to COBOL. In addition, many companies are planning to migrate or reconfigure network equipment to dual-support equipment ( IPv4 / IPv6 ) in combination with “creative” using NAT settings and firewalls (possibly using tools such as Firewall Builder ) to smooth out rough edges. All this does not necessarily require large expenditures, but it is rather difficult and time consuming. Further, the most optimistic estimates of the transition to the “IPv6 world” give a period of at least five years with potential incompatibility problems, loss of access to network resources, “white death screens” and other stress factors that will cause confusion until the last device IPv4 will not be out of use.
Do not worry : All our personal computers are translated to Windows 7, which means they will be compatible with IPv6 . Most network equipment is also ready. As soon as the time comes, we will simply switch the settings to IPv6 , and everything will work without problems.
Worry : Oops ... We need to stop and think about popular technologies and convergence. Most mobile devices with WiFi (Android, iOS, etc.) will be able to use IPv6 over WiFi only if IPv4 will encapsulate it for the organization of the tunnel. Network printers and scanners will require at least a firmware upgrade, as well as most VoIP equipment. Most companies (and consumers) will not be able to overcome the leapfrog of the era of dual stack, tunneling and NAT.
Do not worry : After the transition, IPv6 routers will be 18 times more productive, will have more options for point-to-point communication in high-load modes. Internet performance will only grow.
Worry : In theory, IPv6 performance is higher (at least, not lower) than IPv4 . In practice, people see that performance is lower than IPv4 for several systemic reasons. Most likely, the main culprit is not yet worked out and not implemented centrally transition mechanisms (such as IPv6 tunneling through IPv4 ). Then, failures in the mechanisms of distribution of network content (for example, caching is closer to the consumer) with IPv6 support. Finally, the lack of experience in performance management in accordance with the requirements of IPv6 .
Read more about the transition ( transition ) can be read here . A brief translation of the quote:
UPD. Thank you all for the fruitful discussion. Here I will try to summarize and add useful links from the comments.
In general, the next bell rang - it's time to really think, if not about a quick transition to a new standard, then at least about its readiness.
In February of this year, the Internet Assigned Numbers Authority ( IANA ) Internet Administration identified the last two pre-transition IPv4 addresses of the Asia-Pacific Region Address Space Administration (ARNIC). Somewhere in early August [2011], the four regional Address Space Administrations (ARINC, ARIN, RIPE and LACNIC) will begin to distribute their remaining blocks to the clients that are not reserved for the transition.
From this point on, the Internet will enter a transition phase, the outcome of which will be the widespread use of IPv6 . However, some experts see situations where IPv4 will remain valid for many years.
How destructive will the transition be? This is a big question to discuss. The range of predictions is from “this is not the biggest problem” to “this will be the end of the open Internet as we know it”; from “most people and companies will not notice anything but small roughness” to “inevitable gigantic and long-term costs”; and from “ IPv6 will work very well” to “ IPv6 will kill performance for many classes of applications until new generations of silicon come onto the scene.”
')
Who to believe?
As everyone develops his own sober plans for transition to IPv6 with his suppliers of equipment and communication services, we offer several topics for reflection.
Do not worry : This is the same as the “Problem 2000”. A lot of noise, and the almost complete lack of problems in the end.
Worry : On the other hand, when solving “Problems of the Year 2000”, a lot of working code was reworked, and some obsolete equipment was put into junk. In this case, you need to do quite a bit - just be ready. You will need to make changes to the code or port it to new systems in order to implement new APIs and move to IPv6 applications and databases (sometimes by using add-ons over IPv4 ). One thing brightens up this work - most of these systems are written in more relevant languages ​​compared to COBOL. In addition, many companies are planning to migrate or reconfigure network equipment to dual-support equipment ( IPv4 / IPv6 ) in combination with “creative” using NAT settings and firewalls (possibly using tools such as Firewall Builder ) to smooth out rough edges. All this does not necessarily require large expenditures, but it is rather difficult and time consuming. Further, the most optimistic estimates of the transition to the “IPv6 world” give a period of at least five years with potential incompatibility problems, loss of access to network resources, “white death screens” and other stress factors that will cause confusion until the last device IPv4 will not be out of use.
Do not worry : All our personal computers are translated to Windows 7, which means they will be compatible with IPv6 . Most network equipment is also ready. As soon as the time comes, we will simply switch the settings to IPv6 , and everything will work without problems.
Worry : Oops ... We need to stop and think about popular technologies and convergence. Most mobile devices with WiFi (Android, iOS, etc.) will be able to use IPv6 over WiFi only if IPv4 will encapsulate it for the organization of the tunnel. Network printers and scanners will require at least a firmware upgrade, as well as most VoIP equipment. Most companies (and consumers) will not be able to overcome the leapfrog of the era of dual stack, tunneling and NAT.
Do not worry : After the transition, IPv6 routers will be 18 times more productive, will have more options for point-to-point communication in high-load modes. Internet performance will only grow.
Worry : In theory, IPv6 performance is higher (at least, not lower) than IPv4 . In practice, people see that performance is lower than IPv4 for several systemic reasons. Most likely, the main culprit is not yet worked out and not implemented centrally transition mechanisms (such as IPv6 tunneling through IPv4 ). Then, failures in the mechanisms of distribution of network content (for example, caching is closer to the consumer) with IPv6 support. Finally, the lack of experience in performance management in accordance with the requirements of IPv6 .
Read more about the transition ( transition ) can be read here . A brief translation of the quote:
The main goal of a successful transition is to preserve the ability of hosts to interact with IPv4 and IPv6 addresses. The second goal is to allow IPv6 hosts and routers to be installed as widely and independently as possible. The third goal is maximum transparency for end users, system and network administrators.
IPv6 transition mechanisms are a set of internally implemented tools for hosts and routers with specific procedures for addressing and deployment so that the transition takes place with minimal problems. This gives confidence that IPv6 hosts will be able to interact with IPv4 hosts until the full disappearance of IPv4 from the scene.
UPD. Thank you all for the fruitful discussion. Here I will try to summarize and add useful links from the comments.
- This article is a translation, the author and the topic-starter are different people. By structure, it is provocative - it is clearly visible that the “Do not worry” sections are written with great sarcasm.
- The main pool of IPv4 root registrar addresses has run out and is fully distributed. Regional registrars will begin to distribute their reserves from August 2011. The transition period is at least five years to say the least “instability”.
- IPv6 is announced as a new IPv4 replacement standard. At the moment, it is poorly tested in "adult" networks, the stated compatibility and support is still theoretical - the difficulties of real life have not yet begun. Theoretical discussions have been going on for a decade, but there is no real transition ... Very similar to many failed ERP implementation projects - it’s like, the system is ready and you have to start living in it, but nobody is in a hurry, nobody needs it - and the system becomes outdated, its functionality ceases to match reality.
- IPv6 support in existing systems ... let's call it “patch”. It seems that you can dodge and use ... through f $$. Then in the browser, the IPv6 address must be entered in [] , then external services should be used. There is no smooth system, as created for IPv4.
- Theoretical productivity growth in practice shows its sharp loss and a significant increase in resource requirements .
- The problems of IPv6 will most affect network operators and much less - end users. But this does not mean that for both sides it is unnoticeable. Operators will have to change, buy, re-tune equipment, train people. But at the same time, this is their professional field - that is, there are few very difficult problems. End users will have to learn quite a bit ... But there are many times more of them themselves, and their qualifications in network affairs are low — they will simply be overwhelmed by a host of simple problems.
- We are waiting for price increases - at least for static IP. Rostelecom , for example
In general, the next bell rang - it's time to really think, if not about a quick transition to a new standard, then at least about its readiness.
Source: https://habr.com/ru/post/116785/
All Articles