NETSCALER PERFORMANCE BENCHMARKING VRS IT'S LEADING COMPETITOR - A SUMMARY OF THE RESULTS

NetScaler commissioned the Tolly Group to test NetScaler VPX against its biggest rival. The Tolly Group is a leading global provider of third-party validation services for vendors of IT products, components and services. 

Get the full report click here

Summary

Performance Benchmarking NetScaler VPX vs. F5 BIG-IP Virtual Edition 

• NetScaler VPX outperformed F5 BIG-IP VE in all test scenarios, achieving lower latency, better CPU efficiency, and higher throughput.
• F5 performance declined significantly compared to NetScaler when carrying out multiple ADC tasks.
• NetScaler achieved up to 3x higher throughput for similar CPU usage.
• NetScaler demonstrated up to 64% more efficient use of CPU resources for a given throughput level.
• NetScaler induced up to 89% less latency per request.

 Key Takeaways 

• ADCs play a crucial role in optimizing application delivery and handling security tasks.
• NetScaler's single-pass architecture allows it to handle additional functions without degrading performance.
• Policy functions enable customization of HTTPS headers for improved user experience and security.
• App Sec functions provide real-time scanning for malicious traffic.
• NetScaler VPX consistently outperformed F5 BIG-IP VE in all test scenarios.

 Test Setup & Methodology 

• Tests focused on measuring network latency, data throughput, and CPU usage in real-world traffic environments.
• Comparable virtual appliances from Citrix Systems and F5 BIG-IP VE were tested in an AWS environment.
• Tests were conducted using encrypted sessions with HTTP 2.0 protocol on the front end and HTTP 1.1 on the back end.
• Grafana k6 was used as the traffic generator to measure latency and throughput.

 ADC Throughput for Similar CPU Usage 

• NetScaler VPX showed higher throughput compared to F5 BIG-IP VE across all scenarios.
• NetScaler's throughput was 1.56x higher in core load balancing tests and up to 3.09x higher with additional security functions enabled.
• F5's performance declined as additional functions were enabled, while NetScaler maintained efficiency.

 ADC CPU Usage for Similar Throughput 

• NetScaler demonstrated lower CPU consumption compared to F5 across all tests.
• NetScaler's CPU usage was significantly lower in core load balancing and with additional security functions enabled.
• F5's CPU hit 100% in the LB + Policy + App Sec scenario, leading to a throttled throughput for fair comparison.

 P95 Latency for Similar Throughput 

• NetScaler exhibited lower latency compared to F5 across all scenarios.
• NetScaler's P95 latency was significantly lower in core load balancing and with additional security functions enabled.
• F5's latency increased dramatically with higher user connections, showing a substantial difference compared to NetScaler.

 P95 Latency for Different Numbers of Simultaneous Connections 

• NetScaler's latency remained consistently lower than F5 as the number of simultaneous connections increased.
• At 4,000 users, NetScaler's latency was 96% lower than F5, indicating faster response times and better user experience.