MAC Address Segmentation for Ethernet Bridge Scalability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Ethernet switches face scalability issues in Metropolitan Area Networks (MANs) due to large MAC forwarding tables resulting from numerous client equipment connections, which is not improved by existing techniques like MAC in MAC, and they maintain high complexity in processing client-to-operator network mappings.
Innovation Solution
The method involves dividing the Ethernet network into multiple domains with hierarchical MAC addresses, where each bridge learns and forwards data frames using a designated learning mask to reduce the number of entries in the MAC forwarding matching table, allowing for flexible configuration and reduced table size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional Ethernet switching is used in MAN, then basic data forwarding is achieved, but the MAC forwarding table scale becomes large due to numerous client equipment
Solution Approach 1:
The patent segments the MAC address into two parts: a first MAC address portion (24 bits) that is globally unique and identifies the customer, and a second MAC address portion (24 bits) that is locally unique and identifies the device within the customer's network. This segmentation allows the operator's network to store only the first MAC address portion in the forwarding table, reducing the table size from potentially millions of entries to a manageable number of customer identifiers.
Solution Approach 2:
The patent extracts the customer identification information (first MAC address portion) from the full MAC address and uses it as the key for forwarding decisions in the operator's network. The second MAC address portion is extracted and used only for local routing within the customer's network, separating the functions of customer identification and local device identification.
2Adaptability or versatility
If MAC in MAC technique is applied to reduce forwarding table size, then operator network scalability is improved, but BEB forwarding table size remains large and processing complexity increases
Solution Approach 1:
The patent applies segmentation at the MAC address level, dividing the 48-bit MAC address into two 24-bit portions. This allows the BEB to perform simple length matching against the first MAC address portion without needing to process the complete MAC address or maintain complex mapping tables, thereby reducing processing complexity while enabling scalability.
Solution Approach 2:
The patent changes the parameter used for forwarding table keys from the complete MAC address to only the first MAC address portion. This parameter change reduces the number of entries needed in the forwarding table and simplifies the matching process at the BEB, as only the first 24 bits need to be compared rather than the full 48 bits.
3Measurement precision
If full MAC address is used for learning and forwarding, then accurate device identification is achieved, but the number of forwarding table entries becomes excessively large
Solution Approach 1:
The patent segments the MAC address into a first portion (24 bits) for customer identification and a second portion (24 bits) for device identification within the customer's network. The operator's network uses only the first portion for forwarding decisions, achieving sufficient identification accuracy for routing while dramatically reducing the number of entries needed in the forwarding table.
Solution Approach 2:
The patent applies different levels of specificity to different parts of the MAC address: the first MAC address portion uses global uniqueness for customer identification across the operator's network, while the second MAC address portion uses local uniqueness for device identification within the customer's private network. This local quality approach optimizes both identification accuracy and table size.
Data Source
AI summary
In an Ethernet network divided into multiple domains, a method for Ethernet data frame data learning includes: receiving through a bridge port of a domain of the Ethernet network an Ethernet data frame containing a source Media Access Control (MAC) address; and generating a forwarding matching entry including a destination MAC address, the length of a forwarding mask and a bridge egress-port identifier; wherein the destination MAC address is acquired by matching the source MAC address with the length of the forwarding mask. In a method for Ethernet data frame forwarding, a bridge forwards the Ethernet data frame according to a MAC forwarding matching table including a destination MAC address, the length of a forwarding mask and a bridge egress-port identifier. An Ethernet network and a bridge are also provided. The number of entries included in the MAC forwarding matching table o is reduced effectively.