IPv6 is coming, what should we do with SDN?

IPv6 is coming, what should we do with SDN?

IPv6 has been called for so many years, and finally the "wolf is coming". Since the end of 2017, when the General Office of the CPC Central Committee and the General Office of the State Council simultaneously issued the "Action Plan for Promoting Large-Scale Deployment of IPv6", IPv6 has come in waves. All network operators must actually deploy IPv6 and accept inspections. The reason for the urgent need to transition to IPv6 is that IPv6 is a key basic technology for the evolution of intelligent networks. For a time, IPv6 seemed to interrupt the network reconstruction process of SDN. Major network operators were busy with IPv6 transformation. At least in the network, dual-stack services must be run, and the access requirements of both terminals must be met. All network terminal devices must support IPv4 and IPv6 dual-stacks, and SDN network reconstruction seems to have stopped.

[[241285]]

The introduction of IPv6 has made the reconstruction process of SDN difficult. SDN technology is still facing IPv4 networks. Now IPv6 has come out. How SDN should adapt to it and whether it can fully support it has become a problem. IPv6 has posed a big problem to SDN.

The arrival of IPv6 has indeed brought an impact on SDN, making SDN deployment difficult. However, IPv6 is the pillar of future network development and is indispensable. It is only a matter of time before it replaces IPv4. How could SDN not expect this? In fact, ONF (Open Network Forum), which formulates the SDN standard, has specified support for basic IPv6 protocols in OpenFlow v1.2, including adding support for matching fields such as IPv6 source address and destination address, IP protocol number, service type, ICMPv6 type and encoding, IPv6 neighbor discovery and IPv6 flow label. Support for IPv6 extension headers is specified in v1.3. The latest SDN standard can fully support IPv6.

What is the relationship between IPv6 and SDN?

In fact, they are two completely different technologies. SDN can be likened to a platform technology. IPv6 is a specific protocol and protocol stack running on the platform. It is a relationship of support and being supported. They are fundamentally two things in two spaces and are not contradictory. In the past, SDN ran on the IPv4 protocol, and SDN controllers communicated and visited each other through IPv4 addresses. The flow table sent and the control interaction protocol Openflow were all based on the IPv4 protocol. Now, not only IPv4 but also IPv6 must be supported. The controller address can be IPv6, and the flow table sent and Openflow can both be IPv6. This implementation is not complicated. It just replaces the IPv4 address with IPv6 to make it work.

The core functionality of SDN can only be realized with tools such as IPv6 segment routing and service chaining to truly leverage the true benefits of a software-first network architecture. When you overlay powerful SDN orchestration on a simplified IPv6 kernel, you will find that the efficiency, performance, and reliability that SDN can provide will be truly enhanced.

Originally, IPv6 and SDN had no essential connection and were completely different technologies, but they were "put together" by chance. The development of the next generation of the Internet is inseparable from the promotion of new technologies and new applications. IPv6 and SDN are both the pillars of the development of the next generation of the Internet and neither is dispensable.

SDN separates the control functions of layer 2 and 3 network devices from the devices themselves, and manages the separated control functions in a centralized and unified manner, making the devices simpler and more convenient. It then forwards data through Openflow, making the resources flowing out of the underlying devices easier to control and manage. The increase in the number of IPv6 address terminals means larger routing tables and more complex queries. The SDN concept of separation of control and forwarding and centralized management can play an important role in IPV6.

Based on the SDN architecture, the NAT64 device at the egress is only responsible for data forwarding, which can greatly improve the network forwarding capability, and the corresponding control is placed on the cloud computing resource management platform. The programmability and flexibility of SDN can realize the continuous optimization of the network architecture. Only by upgrading the relevant control software can more applications and functions be upgraded, so that the network infrastructure can support IPv6 applications without worrying about resource waste. Open SDN technology will help the transition from IPv4 to IPv6.

On the contrary, IPv6 also has a positive role in promoting SDN. Only by using tools such as IPv6 segment routing and service chaining can the core functions of SDN truly exert the true advantages of software-first network architecture. Latif Ladid, chairman of the Global IPv6 Forum, emphasized that IPv6 has achieved tremendous growth in recent years, and global deployment is in full swing. The development of new technologies such as 5G and IoT, and even SDN/NFV and cloud are powerful driving forces for IPv6 deployment and application. IPv6 has become the foundation of emerging technologies such as 5G, IoT, SDN/NFV, cloud computing and edge computing.

However, it is impossible for existing IPv4 networks to be completely replaced by IPv6 at once. This requires a long process. IPv4 and IPv6 will coexist for quite a long time in a network or even on the same device. This requires SDN to be adaptable to this hybrid networking.

SDN can use the standard control interface provided by the control plane to write application software for IPv4 and IPv6 interconnection, and then add it to the SDN controller in the form of a plug-in. When the controller receives the communication data packet between the IPv4 host and the IPv6 host, it will hand the data packet over to the application module for IPv4 and IPv6 interconnection for processing, and then set the corresponding forwarding rules in the switch based on the processing results.

***, the switch completes the transmission and exchange of data packets according to the forwarding rules issued by the controller. This process has not been verified and is limited to theoretical implementation, so there will definitely be many problems in the specific implementation, which need to be constantly corrected and improved. Some experimental tests have been carried out on SDN based on pure IPv6 networks.

Therefore, from a technical point of view, it should not be difficult for SDN to support IPv6, but it still requires a lot of practice and correction. Since it is a dual-stack network, it is unknown what problems SDN will encounter if it adapts to the dual-stack network. Now, IPv6 has come all of a sudden, and the entire network must enable IPv6 and run dual stacks, which forces SDN to move forward. It is not enough to just support standards. It requires a lot of experiments and practice. At this stage, there is still a lot to do for SDN to support IPv6.

IPv6 is to solve the current shortage of IPv4, and SDN is to meet our needs for dynamic resource migration and business changes in cloud computing scenarios. Both are indispensable and necessary for future networks. We must do both and be strong. In 2016 and 2017, SDN was implemented in large quantities, and SDN technology deployment was crazy. It can be said that SDN software is a value-added service that promotes the sales of network equipment and increases the price of equipment. In 2018, IPv6 stole the limelight, and the dual-stack protocol was launched across the network to prepare for the switch from IPv4 to IPv6. At this time, SDN seemed to calm down all of a sudden, and the SDN construction boom slowly cooled down, and it was in a state of trial and see. However, the wheel of history will not go backwards. Even if these technologies have various limitations, the direction of development and improvement will not change. I believe that in 2019, IPv6 and SDN will go hand in hand and will become hot spots for network transformation, allowing the network to move towards integration and ushering in a new era belonging to the next generation of the Internet.

<<:  IEEE technical white paper officially released, lossless network research enters a new stage

>>:  Why is it considered a fantasy for Internet giants to operate “a nationwide LoRa network”?

Recommend

Akamai Launches Prolexic Network Cloud Firewall

April 25, 2023 – Akamai Technologies, Inc. (Akama...

Global spectrum auctions valued at $27.5 billion in 2020

On February 8, according to data released by GSA,...

In the 5G era, how are operators doing in the government and enterprise market?

4G has just entered a stable development period, ...

Will the convergence of 5G and IoT be a game changer for the industry?

The convergence of 5G and IoT has long been discu...

Learn the network protocol stack in ARM-uboot from scratch

[[401440]] Network protocol stack in uboot The ne...

SD-WAN’s reputation

1 Introduction This article reviews ETSI GS MEC 0...

...

Interviewer: What are the ways of communication between components in React?

[[409233]] This article is reprinted from the WeC...