Network Update (LAGG LACP Implementation)

We are excited to announce that the Columbus, OH cluster now has LAGG LACP implemented across the following networks on all nodes:

• Private and Public Ceph Cluster Networks
• Hypervisor-Exposed Network carrying SDN bridges for network data

This implementation operates at Layer 3 + 4, enabling advanced traffic distribution and improved handling of both IP routing and transport-level traffic.

What is LACP, and why is it better than standard networking?

LACP (Link Aggregation Control Protocol) is part of the IEEE 802.3ad standard, which allows multiple physical network links to be combined into a single logical link, also known as a link aggregation group (LAG). Compared to traditional single-link networking, LACP provides several key advantages:

1. Higher Throughput
   • Standard networking uses a single link per connection, limiting bandwidth to that one link’s capacity.
   • LACP splits traffic intelligently across multiple links, effectively multiplying the throughput. In our case, each line now has 4 aggregated links instead of 1.
2. Redundancy and Failover
   • In standard networking, a single link failure can cause downtime.
   • LACP provides automatic failover; if one link fails, traffic seamlessly continues across the remaining active links.
3. Load Balancing
   • Standard networks often rely on simple switch-level mechanisms, which can result in uneven traffic distribution.
   • LACP balances traffic at Layer 3 (IP) and Layer 4 (TCP/UDP ports), ensuring that both routing and transport-level traffic are optimized across all links.
4. Simplified Network Management
   • LACP treats multiple physical connections as a single logical interface, making configuration, monitoring, and maintenance simpler while reducing the complexity of managing multiple standalone links.

What this means for the Columbus cluster

• Private/Public Ceph networks now have significantly higher throughput and resilience, improving storage traffic performance.
• Hypervisor SDN bridges benefit from faster, more reliable network traffic for VMs and containerized workloads.
• Overall, the cluster gains enhanced reliability, higher performance, and better failover protection, supporting current operations and future growth.

This LACP deployment is a critical step in modernizing our network infrastructure and maximizing efficiency across our data center.