You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I asked about the `-k` parameter for the `ab` usages to ensure I understand how HTTP Keep Alive was being used in the RPS measurements.
@@ -10,10 +11,9 @@ I asked about the `-k` parameter for the `ab` usages to ensure I understand how
10
11
11
12
> Static file handling is a fundamental task for any web server. This test measures how efficiently each server serves a simple HTML page under concurrent requests. A web server optimized for static content should deliver high requests per second (RPS) with minimal latency and resource usage. This test is crucial for scenarios where websites serve mostly cached, pre-generated pages, such as blogs, documentation sites, and content delivery networks (CDNs).
12
13
13
-
14
14
### Command
15
-
The test was conducted using Apache Benchmark (`ab`) with the following command:
16
15
16
+
The test was conducted using Apache Benchmark (`ab`) with the following command:
17
17
18
18
```sh
19
19
ab -n 100000 -c 100 http://localhost:8080/
@@ -23,10 +23,10 @@ ab -n 100000 -c 100 http://localhost:8080/
23
23
- -c 50 → Number of concurrent users (50)
24
24
-http://localhost:8080/ → URL of the static test page
25
25
26
-
27
26
### Results
28
27
29
28
Test run on:
29
+
30
30
- MacBook Pro 16-inch 2021, Apple M1 Max, 64 GB
31
31
32
32
| Server | Requests per second (RPS) | Latency (ms) | Normalized Performance (%) |
@@ -53,6 +53,7 @@ For fun, here is the same test using the `-k` parameter. In practice this may be
53
53
> Serving large files efficiently is critical for websites that deliver downloads, streaming media, or large assets such as high-resolution images or software packages. This test evaluates how well each web server handles the transfer of a 10MB file under concurrent requests. A well-optimized server should maintain high transfer rates while keeping CPU and memory usage minimal.
54
54
55
55
### Command
56
+
56
57
The test was conducted using Apache Benchmark (`ab`) with the following command:
57
58
58
59
```sh
@@ -66,6 +67,7 @@ ab -n 500 -c 10 http://localhost:8080/file?size=10485760
66
67
### Results
67
68
68
69
Test run on:
70
+
69
71
- MacBook Pro 16-inch 2021, Apple M1 Max, 64 GB
70
72
71
73
| Server | Requests per second (RPS) | Latency (ms) | Transfer Rate (MB/sec) | Normalized Performance (%) |
@@ -77,7 +79,6 @@ Note to calculate the Transfer Rate in MB/sec, I am taking the `ab` result of `K
77
79
78
80
Same test with `-k` which assumes we are using an HTTP proxy with Keep Alive.
79
81
80
-
81
82
| Server | Requests per second (RPS) | Latency (ms) | Transfer Rate (MB/sec) | Normalized Performance (%) |
@@ -92,6 +93,7 @@ It is unexpected that Apache Tomcat would perform worse with `-k` enabled. It is
92
93
> Web servers must efficiently handle high traffic volumes, especially during peak loads. This test measures how well each server performs when faced with 1,000 simultaneous users making requests to a simple HTML page. A well-optimized server should maintain a high request rate with minimal latency and avoid excessive CPU and memory consumption. This test is crucial for sites experiencing traffic spikes, such as e-commerce platforms, news websites, and online services.
93
94
94
95
### Command
96
+
95
97
The test was conducted using Apache Benchmark (`ab`) with the following command:
0 commit comments