IP Address Redundancy for Service Function Chain Steering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional IP network forwarding mechanisms are not scalable, dynamic, or efficient for service function chains (SFCs), as they require manual and static configurations, lack centralized control, and cannot recognize higher-level packet modifications, leading to non-optimal SFC architectures and additional performance costs.
Innovation Solution
The method involves using redundant information in IP addresses to encode traffic steering information, allowing edge switches to replace address portions with logical addresses of next service functions in an SFC, enabling flexible and dynamic traffic steering without modifying the underlying transport network, using a centralized or decentralized control plane to install forwarding policies and perform address swapping for reliable steering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional L2/L3 forwarding mechanisms are used for traffic steering in ISP networks, then the network infrastructure remains simple and compatible, but the steering requirements introduced by service function chains cannot be met, leading to non-optimal SFC architectures and additional performance costs
Solution Approach 1:
The patent embeds service function chain metadata directly within the IP address structure itself. The IP address is divided into multiple fields including a service chain identifier field and service function identifier field, allowing SFC steering information to be nested within the existing IP header without requiring separate metadata structures or additional network protocols.
Solution Approach 2:
The patent makes the IP address structure multi-functional by using different fields within the same IP address for different purposes: traditional routing functions use the network and host fields, while service function chain steering uses the service chain identifier and service function identifier fields. This allows a single data structure to serve both traditional IP routing and modern SFC requirements.
2Productivity
If manual and static low-level configurations are used for SFC steering, then implementation is straightforward on traditional networks, but scalability is poor and dynamicity is lacking
Solution Approach 1:
The patent enables dynamic SFC deployment by allowing service function chain identifiers and service function identifiers to be programmed into the forwarding equipment through automated configuration. The system supports dynamic creation, modification, and deletion of SFCs without manual reconfiguration, with identifiers being assigned and updated programmatically to reflect changing service requirements.
Solution Approach 2:
The patent pre-allocates and structures the IP address fields for SFC metadata before traffic flow is established. The service chain identifier field and service function identifier field are prepared in advance within the IP address structure, enabling rapid SFC deployment and dynamic reconfiguration without requiring real-time address structure modifications.
3Loss of energy
If additional SFC metadata is transferred using existing packet headers (DSCP, ToS), then overhead is minimized, but forwarding equipment must be significantly modified to recognize and implement policies based on this metadata
Solution Approach 1:
The patent merges service function chain metadata with the IP address structure itself. Instead of using separate header fields like DSCP or ToS, the service chain identifier and service function identifier are combined into the IP address fields, eliminating the need for additional metadata overhead while requiring minimal forwarding equipment changes.
Solution Approach 2:
The patent uses the existing IP address structure as a template and creates a copy/adaptation of it that includes SFC metadata. The IP address is reinterpreted to include service chain and service function identifier fields within its existing framework, leveraging the already-deployed IP addressing infrastructure without requiring new header formats.
4Loss of information
If additional packet headers are introduced for SFC metadata transfer with tunneling mechanisms, then steering information can be conveyed, but significant modifications of termination points and additional performance costs are required
Solution Approach 1:
The patent extracts the essential SFC steering information (service chain identifier and service function identifier) from complex metadata structures and tunneling protocols, and embeds them directly into the IP address fields. This extraction eliminates the need for additional packet headers and tunneling mechanisms while preserving complete steering information.
Solution Approach 2:
Instead of adding metadata to packets and requiring termination points to interpret it, the patent inverts the approach by encoding the metadata directly into the IP address that already traverses the network. The IP address itself becomes the carrier of SFC steering information, eliminating the need for special termination point processing.
5Measurement precision
If forwarding engines attempt to recognize packet modifications at higher levels (L4-L7), then fine-grained steering decisions can be made, but forwarding engines cannot recognize these modifications neither in-line nor through signaling mechanisms
Solution Approach 1:
The patent introduces service function chain identifiers and service function identifiers as intermediary elements that bridge the gap between high-level application layer modifications and low-level forwarding engine decisions. These identifiers, embedded in the IP address, act as mediators that carry steering information from L4-L7 packet modifications down to the L3 forwarding plane without requiring forwarding engines to understand higher-level protocols.
Data Source
AI summary
A method for routing traffic in a network includes receiving, by an edge switch, a packet belonging to a traffic class, the packet including a source internet protocol (IP) address and a destination IP address, each of the source IP address and the destination IP address including a redundant information portion and a non-redundant information portion, replacing, by the edge switch, the redundant information portion of the source IP address of the packet belonging to the traffic class and/or the destination IP address of the packet belonging to the traffic class with a logical address of a next service function (SF) in a service function chain (SFC) to which the traffic class is mapped so as to provide a modified packet, and steering the modified packet to the next SF in the SFC.