Performance Recommendations
The vSZ‑D, SZ100‑D, and SZ144‑D platforms
are designed to minimize latency during user data aggregation and forwarding. Their
architecture enables consistent packet performance with minimal throughput degradation
as
the number of tunnels or clients increases. Fast‑path processing is optimized to
scale with
the underlying NIC capacity, whether 1 GbE, 10 GbE, or higher. These platforms are
engineered to handle large numbers of data tunnels and maintain high forwarding
performance.
The following observations and recommendations apply to vSZ‑D, SZ100‑D, and SZ144‑D performance:
- Use Direct I/O for best throughput. RUCKUS recommends operating vSZ‑D in Direct I/O mode for maximum performance. This recommendation applies to both ESXi and KVM hypervisors.
- vSwitch mode offers flexibility with trade‑offs. vSZ‑D supports vSwitch mode for deployments where the DP is co‑located with other VMs for service chaining on the same hardware. vSwitch mode has lower maximum throughput than Direct I/O due to virtio/vSwitch overhead.
- Encryption introduces expected overhead. Enabling encryption (AES‑128 or AES‑256) on RUCKUS GRE tunnels adds processing overhead for encryption and decryption on both APs and the DP. The software is optimized to minimize latency and leverages Intel hardware crypto modules, but the impact remains bounded by hardware capabilities.
For deployment‑specific recommendations or tuning guidance, contact RUCKUS Support.