Exceptions Vs Result Pattern

Exercise to investigate K6 Load testing tool and BenchmarkDotNet NuGet package for Load and performance testing doing a comparation between Exceptions and a Result Pattern.

The results of this tests does not imply Exceptions should never being used neither Result Pattern has to be implemented always. Each technique should be used properly attached to the use case. To read moare about that discussion:

• https://martinfowler.com/articles/replaceThrowWithNotification.html
• https://www.reddit.com/r/csharp/comments/1aoz4b0/result_pattern_vs_exceptions_pros_cons/
• https://www.reddit.com/r/dotnet/comments/11751gx/exceptions_vs_result_pattern/

Description

Having two endpoints that do the same thing, one using exceptions and the other using a Result Pattern, we will compare the performance of both.

Load Test Scenarios

1. We simulate ramp-up of traffic from 1 to 20 users over 10 seconds.
2. Stay at 20 users for 50 seconds

How To Execute them

1. Assure you have k6 installed. If not, you can download it from here
2. Set Exceptions.Vs.ResultPattern.Api as startup project
3. Run it without debugging (Ctrl + F5)
4. Open a terminal and navigate to the root of the project
5. Run the following command to execute the load test scenario you want to test. Assure you pass the correct file name.

   k6 run api-test-add-book-v1.js

6. Check the results in the terminal

Results

Get unexisting book using Exceptions

Around 4228 request per second

Try add a book with invalid data using Exceptions

Around 3893 request per second

---

Get unexisting book using Result Pattern

Around 30296 request per second. 7 times faster than Exceptions scenario

Try add a book with invalid data using Result Pattern

Around 23469 request per second. 6 times faster than Exceptions scenario

---

Performance Test Scenario

How To Execute them

1. Set Exceptions.Vs.ResultPattern.PerformanceTests as startup project
2. Build the solution in Release mode
3. Run it without debugging (Ctrl + F5)
4. Check the results in the terminal. Also the results are being exported in HTML and Markdown on \Benchmarking.Exceptions.Vs.ResultPattern\bin\Release\net8.0\BenchmarkDotNet.Artifacts\results folder

Results

BenchmarkDotNet v0.14.0, Windows 11 (10.0.22631.4112/23H2/2023Update/SunValley3) 12th Gen Intel Core i5-1240P, 1 CPU, 16 logical and 12 physical cores .NET SDK 8.0.304 [Host] : .NET 8.0.8 (8.0.824.36612), X64 RyuJIT AVX2 DefaultJob : .NET 8.0.8 (8.0.824.36612), X64 RyuJIT AVX2

Method	Mean	Error	StdDev	StdErr	Min	Q1	Median	Q3	Max	Op/s	Gen0	Exceptions	Completed Work Items	Lock Contentions	Allocated
AddBookOrThrowBenchmark	501.399 μs	11.9045 μs	34.7260 μs	3.5079 μs	432.779 μs	477.070 μs	493.169 μs	523.111 μs	586.209 μs	1,994.4	6.8359	100.0000	-	-	42.97 KB
AddBookOrFailBenchmark	13.057 μs	0.2603 μs	0.5137 μs	0.0742 μs	11.850 μs	12.688 μs	13.016 μs	13.378 μs	14.353 μs	76,588.3	7.9041	-	-	-	48.44 KB
GetBookOrThrowBenchmark	425.688 μs	8.3973 μs	10.3127 μs	2.1987 μs	409.043 μs	416.912 μs	426.451 μs	432.276 μs	441.762 μs	2,349.1	6.8359	100.0000	-	-	43.75 KB
GetBookOrFailBenchmark	4.889 μs	0.0918 μs	0.1832 μs	0.0262 μs	4.497 μs	4.765 μs	4.871 μs	4.971 μs	5.317 μs	204,528.6	3.1815	-	-	-	19.53 KB

AddBookOrFailBenchmark using Result Pattern is 38 times faster than AddBookOrThrowBenchmark using Exceptions.

GetBookOrFailBenchmark using Result Pattern is 87 times faster than GetBookOrThrowBenchmark using Exceptions.