Feature |
Product |
Release introduced |
---|---|---|
MultiLink Trunking (MLT) / Link Aggregation Group (LAG) |
5320 Series |
Fabric Engine 8.6 |
5420 Series |
VOSS 8.4 |
|
5520 Series |
VOSS 8.2.5 |
|
5720 Series |
Fabric Engine 8.7 |
|
7520 Series |
Fabric Engine 8.10 |
|
7720 Series |
Fabric Engine 8.10 |
|
VSP 4900 Series |
VOSS 8.1 |
|
VSP 7400 Series |
VOSS 8.0 |
MultiLink Trunking (MLT) is a point-to-point connection that aggregates multiple ports to logically act like a single port with aggregated bandwidth. Grouping multiple ports into a logical link provides a higher aggregate on a switch-to-switch or switch-to-server application.
To include ports as trunk group members of an MLT, you must statically configure the ports.
You can use a multilink trunk to aggregate bandwidth between two switches. The MLT algorithm ensures that each packet in a flow does not arrive out of sequence, and that a flow always traverses the same link path.
The hashing algorithm uses the following packet fields and the incoming interface (source) port number to calculate the index to outgoing (destination) port number in an MLT:
Traffic type |
Hashing algorithm |
---|---|
IPv4 traffic |
Hash Key = [Destination IP Address (32 bits), Source IP Address (32 bits), Source TCP/UDP Port, Destination TCP/UDP port] |
IPv4 traffic without TCP/UDP header |
Hash Key = [Source IP Address (32 bits), Destination IP address (32 bits)] |
IPv6 traffic |
Hash Key = [Destination IPv6 Address (128 bits), Source IPv6 address (128 bits), Source TCP/UDP Port, Destination TCP/UDP port] |
IPv6 traffic without TCP/UDP header |
Hash Key = [Source IP Address (128 bits), Destination IP address (128 bits)] |
Mac-In-Mac transit traffic |
|
Layer 2 Non-IP traffic |
Hash Key = [Destination MAC Address (48 bits), Source MAC Address(48 bits)] |
To use MLT with the switch, you can have ports running at different speeds. After you use MLT with LACP, LACP dynamically checks for proper speed on all port members. You do not need to have similar physical connection types. After you use autonegotiation with MLT and not LACP, you need to ensure that all ports run at the same speed.
Multilink trunks adhere to the following rules. Unless otherwise stated, these rules also apply to MLT with LACP.
You cannot configure an MLT name that uses all numbers, for example, 222.
Multilink trunk ports support mixed speed links, for example, one link can be 10Gb and another 1Gb. However, no weighting of traffic distribution occurs so if you mix links of different operational speeds, you can overload the lower speed link or underutilize a higher speed link.
This rule applies to multilink trunks only. MLT with LACP does not support different link speeds.
All multilink trunk ports must be in the same Spanning Tree Group (STG) unless the port is tagged. Use tagging so ports can belong to multiple STGs, as well as multiple VLANs.
After the port is made a member of MLT, it inherits the properties of the MLT and hence the STG properties are inherited from the VLAN associated with that MLT. After you remove the port from MLT or after you delete the MLT, the ports are removed from the MLT STG and added into the default STG.
MLT is compatible with Multiple Spanning Tree Protocol (MSTP) (IEEE 802.1s) and Rapid Spanning Tree Protocol (RSTP) (IEEE 802.1w).
Tagging (IEEE 802.1Q) is supported on a multilink trunk.
Multilink trunks have the following general features and requirements:
Supports MLT groups with as many as 8 ports belonging to a single multilink trunk. For more information about the number of MLT groups supported for each hardware platform, see Fabric Engine Release Notes.
Apply filters individually to each port in a multilink trunk.
With MSTP or RSTP enabled, ports in the same multilink trunk operate as follows:
The designated port sends the Bridge Protocol Data Unit (BPDU).
The multilink trunk port ID is the ID of the lowest numbered port.
If identical BPDUs are received on all ports, the multilink trunk mode is forwarding.
If ports do not receive BPDUs on a port or BPDU and port tagging do not match, the individual port is taken offline.
Path cost is inversely proportional to the active multilink trunk bandwidth.
The Link Aggregation Group (LAG) adheres to the following rules:
All LAG ports operate in full-duplex mode.
All LAG ports operate at the same data rate.
All LAG ports must belong to the same set of VLANs.
Link aggregation is compatible with MSTP, and RSTP.
Assign all LAG ports to the same MSTP or RSTP groups.
You can configure a LAG with up to 24 ports, but only a maximum of 8 can be active at a time.
After you configure a multilink trunk with LACP, you cannot add or delete ports or VLANs manually without first disabling LACP.