Zscaler Blog
Get the latest Zscaler blog updates in your inbox
How SASE Vendors Design their Solution - Part 4: The Performance Illusion—Why Your ICMP Ping is Lying to You
The Deception of "Front Door" Latency
When evaluating a SASE solution, performance is paramount. Users expect a fast, seamless experience, and security should never be a bottleneck. However, many legacy SASE providers present performance metrics that are, at best, misleading and, at worst, entirely deceptive.
The most common tactic is to showcase low latency figures based on ICMP pings or MTR traceroutes. A customer will be shown a report indicating a sub-10-millisecond connection time to the vendor's nearest point of presence (POP) and be told this is evidence of a superior network.
This is the performance illusion. ICMP is a network-layer protocol (Layer 3) designed to test basic reachability. When a ping is sent to a SASE POP, it is typically answered by the network interface card of a load balancer or router at the "front door." The request never actually enters the proxy processing engine where the real work of security inspection happens. It's like timing how long it takes to walk to a restaurant's door but not how long it takes to order, cook, and serve the meal. These ICMP Ping MTR limits mean they are measuring the performance of the network, not the performance of the security service.
The Good: Measuring Real-World Throughput with Full Decryption
The only true measure of SASE performance is to analyze real-world Layer 7 application traffic with all security features enabled. This means measuring the complete transaction time for a user accessing an application, from the initial request to the final response, with the SASE service in the middle performing its full suite of security functions.
The most critical of these is full SSL/TLS decryption. Today, over 90% of internet traffic is encrypted. Any security inspection that does not decrypt this traffic is effectively blind. However, decryption is computationally expensive. This is where many alternative SASE architectures, particularly those built on service-chained virtual appliances, falter.
A well-architected, cloud-native SASE platform is designed to handle SSL/TLS decryption at scale. Its SSMA engine decrypts traffic once, performs all inspections in memory, and re-encrypts it, minimizing the performance impact. True testing requires measuring the throughput of the system under this full load. Any performance metrics that are presented with decryption turned off are not reflective of a real-world deployment.
The Ugly: Hiding Behind "Decryption-Off" Test Results
Many first-generation cloud security vendors are acutely aware of the performance limitations of their architectures when it comes to SSL/TLS decryption. As a result, they often hide behind "decryption-off" test results or recommend that customers bypass decryption for trusted applications like Microsoft 365.
This is the "ugly" secret of their performance metrics. They are showcasing the speed of an empty highway, not the reality of rush hour traffic. By advising customers to bypass decryption, they are asking them to make a direct trade-off between security and performance—a compromise that a true SASE architecture does not require.
To get an accurate picture of performance, organizations must move beyond synthetic pings and demand real-world telemetry testing. The best approach is to leverage a Digital Experience Monitoring (DEM) solution that can proactively measure the actual application performance and end-to-end path telemetry for real users. This provides an unvarnished view of the user experience and cuts through the marketing hype of misleading ICMP metrics. When it comes to SASE performance, the only numbers that matter are those that reflect a fully loaded system doing its job.
FAQs
ICMP pings and MTR traceroutes are misleading because they only measure network-layer (Layer 3) reachability to the "front door" of a SASE point of presence (POP). These tests are typically answered immediately by a router or load balancer network interface card, completely bypassing the Layer 7 proxy processing engine where security inspection, policy enforcement, and decryption actually occur. Consequently, they measure basic network transit rather than the true latency added by the security service.
The performance illusion refers to SASE vendors showcasing sub-10-millisecond latency figures based on synthetic Layer 3 ping tests while turning off resource-intensive security features like SSL/TLS decryption and DLP. This creates a false impression of high speed by measuring the performance of an idle network path rather than the real-world processing times of a fully loaded security pipeline handling live application traffic.
With over 90% of enterprise web traffic encrypted, security engines are blind without decryption. However, SSL/TLS decryption is computationally expensive. This makes it the ultimate benchmark because it exposes the architectural efficiency of the data plane. While stitched-together, service-chained architectures experience severe throughput bottlenecks from repeatedly decrypting and re-encrypting traffic, a true Single-Scan Multi-Action (SSMA) engine decrypts the traffic exactly once in memory, processing all policies simultaneously with minimal CPU overhead.
When vendors recommend bypassing decryption for "trusted" applications like Microsoft 365 to preserve network speed, they force organizations into a dangerous compromise between security and performance. Bypassing decryption creates massive blind spots, allowing advanced threats, malware, and data exfiltration techniques to exploit trusted cloud channels completely uninspected by the SASE platform.
Instead of relying on static ping tests, organizations should deploy Digital Experience Monitoring (DEM) solutions to capture active, end-to-end path telemetry directly from the user's endpoint to the target application. DEM tools measure the complete Layer 7 transaction time under full security load (including SSL decryption and DLP), providing an unvarnished, real-world view of user experience and application performance.
Was this post useful?
Disclaimer: This blog post has been created by Zscaler for informational purposes only and is provided "as is" without any guarantees of accuracy, completeness or reliability. Zscaler assumes no responsibility for any errors or omissions or for any actions taken based on the information provided. Any third-party websites or resources linked in this blog post are provided for convenience only, and Zscaler is not responsible for their content or practices. All content is subject to change without notice. By accessing this blog, you agree to these terms and acknowledge your sole responsibility to verify and use the information as appropriate for your needs.
Get the latest Zscaler blog updates in your inbox
By submitting the form, you are agreeing to our privacy policy.


