IPv4 addresses are gradually
running out , and in the future they may not be enough for all Internet of Things devices to work -
it is
expected that by 2030 the number of connected gadgets will reach 125 billion. To solve the problem, the
International Telecommunication Union , a specialized UN agency, presented
a migration plan for IPv6 .
However, the IT community has seriously criticized it.
Further, we tell why.
/ Flickr / Ryan Hyde / CCWhat is this plan
The plan was developed by a study group (
Study Group 20 ) of the International Telecommunication Union. It is a kind of "role model" when migrating from IPv4. The draft document describes the structure of IPv6 addresses, provides recommendations for working with global routing prefixes, and proposes a migration model by “mapping” IPv4 addresses to corresponding IPv6 addresses (on
page 15 of the document ).
The essence of the latter is as follows. The first hexadecimal digit of the subnet identifier (A) (4 bits) is used to determine the location. The second hexadecimal digit (B) (4 bits) is needed to classify subnets, for example, DMZ, LAN, or IoT. The third (C) and fourth (D) hexadecimal digits (also 4 bits each) are used to identify specific subnets. When mapping an IPv4 address to IPv6, the digits A and D are filled with zeros.
The developers of the plan note that the use of the proposed model will be beneficial to the end users of the network, since it is “easy to implement and adjust to fit your needs”. In their opinion, this will be useful when working on IoT projects for smart cities, government agencies and private companies.
What does the community think about it
However, the plan proposed by the UN division was criticized by engineers and online registrars. According to
the editors of The Register resource, a representative of the Internet registrar RIPE called the model “inexpedient and useless”.
Benedikt Stockebrand, IPv6 expert, author of
numerous publications on the topic (including the
book “IPv6 in practice: a guide to the next generation Internet”), who also maintains a
video blog about IT, in his address to RIPE Mailing Lists
highlighted the following disadvantages of the proposed model:
1. Inapplicability to real networksNetworks are divided into subnets that differ in properties: behavior and delays, security, QoS, and others. However, the document ignores these categories. All this can lead to the appearance of weakly protected networks or it can lead to their fragmentation. This will seriously complicate the construction of large networks (for example, in corporations).
2. The model transfers old problems to a new shell.The document proposes to make the transition to IPv6 by reassigning IPv4-addresses to the addresses of the new protocol. This will take all the problems from the “era” of IPv4 into the era of IPv6 and lead to the growth of the network topology.
3. We'll have to abandon the basic practices for ensuring information security in networksFor example, microsegmentation and multi-level security zones are used to protect the perimeter of the data center. In the proposed model, their implementation will lead to inflation of the routing tables.
4. “Lifetime” of IPv6 will decreaseStockbrand has calculated that the model will shorten the life of the protocol by 25%, that is, more than 42 years (at the current rate of development of the Internet). He attributes this to the fact that the proposed model uses two octets of data for purposes not related to routing.
In his letter, Stockbrand concluded that the UN plan had not yet been fully considered, and in its current form, its implementation would jeopardize the future of the Internet. However, the proposal of the International Telecommunication Union is still only a draft and it can be expected that in the future it will be finalized taking into account the proposals of the community.
/ Flickr / Robert / CCIPv6 deployment is slow
6 years have passed since the
global launch of IPv6, but the protocol has not yet been universally distributed. Jeff Houston, chief scientist at APNIC Regional Internet Registrar (Geoff Huston), notes that a statistical analysis of the past four months
shows a significant slowdown in IPv6 deployment.
There are
several reasons for this . The first is the human factor. There are not so many companies in the world where technical specialists can convince management to start switching to IPv6 (as they did on Google and Facebook). And the second is technical problems.
For example, in Yandex, we
encountered difficulties in implementing spam defenses. Anti-spam algorithms of Yandex. Mails combine not only statistical and heuristic methods, machine learning, but also a decision-making mechanism based on these factors. One of the methods checks the IP addresses of the computers involved and stores the reputation of their IPv6 addresses, but their total number greatly exceeded the RAM size of all Yandex servers.
Thus, we can conclude that the introduction of IPv6 is still “in limbo”. Perhaps a new proposal by the International Telecommunication Union in the future will help rectify the situation or pave the way for other effective initiatives. But while the published document is only a draft. And it will take some time to finalize it.
Additional blog reading on the VAS Experts website:
VAS Experts specialization - the introduction of services for managing and analyzing Internet traffic:
• SCAT DPI | IS SORM | BRAS | DLP systems | Our cases | Events | News •