This video demonstrates a performance benchmark of two MikroTik CCR 2116 Cloud Core Routers, intended for use in the IP464 cloud router service. The host conducts two distinct tests to evaluate their throughput capabilities. The first test measures raw, unencrypted routing performance between the two devices, which are linked by a 20G bonded connection. This setup achieves nearly 10 Gbps, with CPU utilization remaining under 40%.
The second, more critical test measures performance when routing traffic through an encrypted WireGuard VPN tunnel running between the two routers. The throughput drops dramatically to approximately 1.6 Gbps. The investigation reveals that the bottleneck is a single CPU core reaching 100% utilization, indicating a limitation in the platform's multi-core handling of the WireGuard encryption process. The host concludes that while the encrypted bandwidth is limited, the primary advantages of this hardware for his service are the significantly lower latency and more direct internet peering compared to software-based solutions, which are deemed more valuable than raw throughput.
Benchmarking MikroTik CCR 2116 Routers for a WireGuard VPN Service
00:00:01 - Hardware Introduction: The video focuses on testing two MikroTik CCR 2116 routers, described as high-end, powerful models for routing-intensive tasks.
00:02:36 - Test Environment Setup: A test network is established with a PC initiating an iPerf test that routes through both MikroTik routers before reaching a Minis Forum server. The two routers are interconnected with a 20G bonded link (2x10G).
00:05:38 - Unencrypted Routing Test: The first test measures raw routing performance without any VPN encryption. The goal is to establish a baseline for the hardware's maximum throughput.
00:06:20 - Unencrypted Performance Results: The routers successfully handle just under 10 Gbps of unencrypted traffic. During this test, the CPU load on the routers remains moderate, at around 37-40%, showing they can easily manage high-speed routing.
00:07:30 - WireGuard VPN Test: The test is reconfigured to route all traffic through an encrypted WireGuard tunnel running between the two routers. This simulates the real-world use case for the host's cloud router service.
00:09:09 - Encrypted Performance Results: With WireGuard enabled, the throughput drops significantly to between 1.4 and 1.6 Gbps, a fraction of the unencrypted performance.
00:09:51 - Identifying the Bottleneck: A closer look at the system resources reveals that the performance is limited by a single CPU core on one of the routers hitting 100% utilization. This suggests the WireGuard process is not effectively using all available cores, creating a single-threaded bottleneck.
00:12:00 - Key Takeaway: Latency Over Bandwidth: The host explains that for the intended cloud router application, the primary benefits of this dedicated hardware are low latency and fewer network hops to major internet exchanges. He argues that these factors are more critical for performance than raw bandwidth, and 1.6 Gbps is still sufficient for most users.
00:12:47 - Future Deployment and Service Updates: The tested routers will be deployed in a data center to be offered as a hardware option for the
Error: Couldn't find 4537 IP464 cloud router service. The host also announces upcoming features based on user requests, including support for private IP addresses and fully managed infrastructure.