SRv6 opens a new IP era

We know that IP data transmission in current bearer networks is mainly based on MPLS (Multi-Protocol Label Switching) technology. While MPLS improves routing and forwarding efficiency, it also inevitably brings some new problems:

• Multiple configurations: Independent signaling protocols such as LDP (Label Distribution Protocol) and RSVP (Resource Reservation Protocol) need to be used, and labels need to be assigned to each node in the network. The configuration process is relatively complex.
• Low efficiency: Each time data is transmitted, configuration information needs to be sent to all nodes on the end-to-end connection path, resulting in low transmission efficiency.
• Difficult to expand: Each node needs to maintain the status of each connection, making business expansion difficult.
• Cross-domain troubles: Different networks (such as IP backbone network, IP metropolitan area network, mobile backhaul network, etc.) are independent MPLS domains, and the interconnection between networks requires the use of complex cross-domain VPN technology to achieve.

picture

picture

The emergence of SR (Segment Routing) technology based on source routing has changed the idea of ​​forwarding based on destination routing in traditional IP networks. Through SR technology, the first node in the network plans and establishes an end-to-end connection path, and the intermediate nodes only need to forward and no longer need to maintain the connection status, which greatly simplifies the deployment and expansion of the network.

The combination of SR and MPLS (SR-MPLS) helps solve the aforementioned network "chronic diseases" of multiple configurations, low efficiency, and difficult expansion. However, all nodes in the network still need to support MPLS label forwarding technology, which does not fundamentally solve the problem of cross-domain interconnection. In addition, the expansion capability of MPLS labels is limited, making it difficult to better meet the transmission needs of diverse services in the 5G era.

picture

At the same time, in order to solve the problem of shortage of IPv4 addresses and make the transmission of IP network data faster, more reliable and more secure, the Internet Protocol is transitioning from IPv4 (Internet Protocol version 4) to IPv6 (Internet Protocol version 6).

With the large-scale deployment of IPv6, MPLS-based transmission technology has increasingly become a "bottleneck" of the network in key application scenarios of 5G bearer and cloud-network integration.

picture

SRv6, a new generation of IP bearer network data forwarding protocol that integrates the characteristics of SR and IPv6, came into being. SRv6 fully inherits the source routing advantages of SR and the simplicity and easy scalability of IPv6.

picture

• It is no longer necessary to use a separate signaling protocol to assign labels to all nodes, making the network easier to configure and manage.
• It is no longer necessary for all nodes to support MPLS, or even SRv6, making the network more compatible.
• End-to-end deployment of services can be achieved based on IPv6 messages with extended message headers, making the network more concise and efficient.
• The network is programmable, which enables flexible expansion of services. Combined with SDN (Software Defined Network), it can also achieve flexible network scheduling.

picture

The key to SR technology is to assign a Segment Identifier (SID) to each node in the network or each path between nodes as an identifier, and to specify the SID set (Segment List) of nodes and paths that the message needs to pass through at the starting node to guide the forwarding of the message.

SR-MPLS uses a 20-bit MPLS label value as the SID, while SRv6 uses a 128-bit IPv6 address format value as the SID.

Compared with SR-MPLS SID, SRv6 SID is longer and supports carrying more information. It can be used to identify multiple functions or service types such as nodes, links, L2VPN services, L3VPN services, and network services.

picture

The 128-bit SRv6 SID contains three fields: Locator, Function, and Arguments.

picture

With SRv6 SID, SRv6 has the ability to orchestrate paths and services. It can pre-plan the path for packet forwarding and the forwarding behavior of each node on the path, and support the definition of any network function or service.

In addition to SRv6 SID, SRv6 has another "secret weapon".

As mentioned earlier, although the combination of SR and MPLS (SR-MPLS) helps solve the "chronic problems" of multiple configurations, low efficiency, and difficult expansion in the network, all nodes in the network still need to support MPLS label forwarding technology, which still does not fundamentally solve the pain point of cross-domain troubles.

SRv6 makes full use of the scalability of IPv6 and replaces the label forwarding function of MPLS with a new extension header type SRH (Segment Routing Header). This allows the SRv6 network to achieve efficient data forwarding based on native IPv6 technology (Native IPv6) without the need for other technologies, completely solving the cross-domain problem.

Ordinary IPv6 packets can contain zero or more extension headers to implement different business functions. Extension headers are added only when necessary.

After the SRH is extended, the SRv6 message structure includes three parts: IPv6 message header, SRH extension header and data message.

picture

• IPv6 packet header: used to specify the source address (SA) and destination address (DA) of the packet.
• SRH extension header: used to specify the forwarding path information of the message, including the number of intermediate nodes (Segments Left, SL) and the segment list (Segment List). The segment list is a list of SIDs of all nodes that the message will pass through during transmission, and the number of intermediate nodes refers to the number of nodes passed through.
• Data message: The transmitted business data information remains unchanged during the transmission process.

If you want to simplify the understanding, the structure of the above three parts can be represented by the following figure.

picture

The source address (SA) and destination address (DA) in a common IPv6 message respectively identify the first node and destination node of the message, and remain unchanged during transmission. However, the destination address (DA) in an SRv6 message identifies the next node that the current message passes through, and changes in real time as the data is transmitted.

SRv6 uses the intermediate node number and segment list in the SRH extension header to guide the forwarding of the message. Every time a SRv6 node is passed, the value of the intermediate node number is reduced by 1 and the destination address information is updated. The destination address information is determined by the intermediate node number and the segment list. For example, when SL=n, DA=SID[0]; when SL=0, DA=SID[n].

picture

After saying so much, you may still not quite understand how SRv6 works.

Let's take a look at a specific example to see how SRv6 transmits data in the network.

picture

1. The first node (node ​​A) receives the data transmission demand from the user (source node S) and determines the destination (destination node R). The SRv6 planned transmission path is A→B→C→D. Then in the message sent by node A: SA=S, DA=B, the segment list is <SID D, SID C, SID B>, and the number of intermediate nodes SL=2.
2. After receiving the message, nodes B and C reduce the SL value in the message by 1, change the DA to the next node, and forward the message.
3. After receiving the data, node D finds that the SL value is 0 and determines that it is the tail node. It then forwards the message to the destination user (destination node R) in the normal IPv6 message forwarding mode.

As you can see, when data is transmitted through SRv6, there is no need to distribute labels to each node in the network, nor is there a need to maintain the status of each connection. It does not rely on MPLS tunnels for cross-domain data forwarding, which is very convenient and efficient.

picture

Although MPLS technology is no longer used, SRv6 still supports providing differentiated services for different types of services through different working modes.

SRv6 mainly has two working modes: SRv6 Best Effort (BE) and SRv6 Traffic Engineering Policy (TE Policy).

• The function of SRv6 BE is similar to that of MPLS LDP. LDP uses the shortest path algorithm of IGP (Interior Gateway Protocol) to calculate an optimal path to guide data forwarding. SRv6 BE uses only one service SID (Service SID) to guide the forwarding of packets, which is a best-effort working mode. In this working mode, the SRv6 function only needs to be deployed at the head and tail nodes, which is relatively simple to implement and is suitable for scenarios where some common VPN services need to be quickly opened.
• SRv6 TE Policy uses the characteristics of source routing and encapsulates an ordered Segment List (path information) at the first node to guide how to forward packets in the network. Combined with the programmable characteristics of the Segment List and the introduction of a coloring mechanism (setting the color attribute), SRv6 TE Policy can flexibly specify any forwarding path for packets to implement functions such as traffic engineering, flexible traffic diversion, and load balancing.

picture

We will also release special graphics and articles to provide detailed introductions on the SRv6 BE and SRv6 TE Policy working modes in the future. Stay tuned.

picture

SRv6 not only solves the four major pain points mentioned at the beginning of this article, but also has other unique advantages.

Combining with SDN technology

SRv6 can be combined with SDN technology, using its own flexibility to cooperate with the management and scheduling capabilities of SDN. The entire network is like a programmable software system that can flexibly implement various services. At the same time, network configuration and data transmission become simpler.

As shown in the figure below, the SDN controller collects SRv6 node and path information, plans appropriate paths and services provided by each node according to business needs, and notifies the information to the head node (node ​​A in the figure below). Based on the received information, the head node transmits the business data to the destination node through the SRv6 network.

1. Combined with EVPN (Ethernet Virtual Private Network) technology
   SRv6 can also be combined with EVPN technology to free the IP bearer network from the constraints of numerous complex protocols and simplify it into an Overlay (EVPN) + Underlay (SRv6) networking structure, greatly simplifying the complexity of the network.

SRv6 is the core protocol of the new generation IP bearer network after MPLS. It will not only simplify and unify the architecture of the bearer network, but also continue to promote the development of fixed (fixed network) mobile (mobile network) convergence and cloud (cloud computing) network (communication network) convergence.

I believe that in the near future, there will be more application innovations based on SRv6, which will drive all industries to develop in the direction of intelligence and digitalization.

Technology changes our lives, let us look forward to it together!