Link Aggregation
What Is Link Aggregation?
Link aggregation is a networking technique that combines multiple physical network links between two devices into a single logical connection, increasing total throughput and providing redundancy. By treating several parallel links as one virtual interface, link aggregation allows networks to exceed the bandwidth ceiling of any single cable or port while also ensuring that traffic continues to flow if one or more physical links fail. The technique is standardized under IEEE 802.1AX (formerly IEEE 802.3ad) and is widely implemented in data center, enterprise campus, and carrier networks.
Link aggregation applies to Ethernet links at various speeds, from Fast Ethernet to 10GbE, 25GbE, 40GbE, and beyond. A group of aggregated ports is called a Link Aggregation Group (LAG). The two endpoints of a LAG must negotiate and manage the bundle, either through static configuration or through a dynamic control protocol.
IEEE 802.1AX and the Link Aggregation Control Protocol
The Link Aggregation Control Protocol (LACP) is defined in the IEEE 802.1AX standard and provides automatic discovery, negotiation, and management of link aggregation between two devices. LACP-enabled ports exchange protocol data units (PDUs) called LACPDUs, which carry information about the port's system identifier, port priority, and aggregation state. When two devices detect that their ports share compatible parameters, LACP brings those ports up as a LAG without manual intervention. LACP also handles the graceful removal of a failed link from the bundle and can detect misconfigurations such as a cable connected to the wrong port. Up to eight active links can participate in a single LACP-managed LAG, with additional ports held in standby for fast failover. The alternative to LACP is static link aggregation, sometimes called mode-on or unconditional trunking, which requires matching configuration on both ends without any control protocol exchange.
Load Balancing and Traffic Distribution
A LAG does not send individual packets across all member links simultaneously; instead, it uses a hashing algorithm to assign each flow to a single member link. Common hash inputs include source and destination MAC addresses, source and destination IP addresses, and source and destination TCP or UDP port numbers. The choice of hash inputs affects how evenly traffic is distributed across the bundle. Using only MAC addresses may produce poor distribution if most traffic flows through a single gateway, while including Layer 4 port numbers provides finer granularity. The Cisco documentation on IEEE 802.3ad link bundling details how different hash modes interact with specific traffic patterns, a consideration that directly affects effective throughput in asymmetric environments. Because all packets in a flow traverse the same physical link, per-flow hashing guarantees in-order packet delivery without requiring reassembly at the receiver.
High Availability and Failover
Link aggregation improves network resilience by eliminating the single-link dependency between switches, servers, and storage appliances. When a physical link fails, LACP detects the loss of LACPDUs within seconds and redistributes its flows across the remaining active links. This behavior requires no reconfiguration of routing or switching tables above the LAG layer, making failover transparent to upper-layer protocols. In server environments, link aggregation is implemented through bonding or teaming drivers in the operating system, such as the Linux kernel's bonding driver or Windows Server's NIC Teaming feature. The technical overview at fs.com on link aggregation and LACP provides a practical summary of how these mechanisms behave in common deployment scenarios.
Applications
Link aggregation has applications across a wide range of network infrastructure contexts, including:
- Data center top-of-rack and spine-leaf switch interconnects
- Server-to-switch uplinks for high-throughput virtualization hosts
- Storage area network (SAN) connectivity for NFS and iSCSI traffic
- Campus network distribution and core layer uplinks
- Internet service provider edge connections requiring resilient handoffs