Feedback

Author: ManickaP

build/httpclient-scenarios.yml

@@ -11,13 +11,13 @@ parameters:
   type: string
   default: 'true'
 
-- name: getScenarios 
+- name: getScenarios
   type: object
   default:
   - displayName: "GET"
     arguments: --scenario httpclient-kestrel-get $(httpClientJobs) --variable useHttpMessageInvoker=true --property server=kestrel --property client=dotnetinvoker --property method=get --client.options.collectCounters true
 
-- name: postScenarios 
+- name: postScenarios
   type: object
   default:
   - displayName: "POST"
@@ -33,7 +33,9 @@ parameters:
     - displayName: "HTTP/2.0 (mult conn)"
       arguments: --variable httpVersion=2.0 --variable useHttps=true --variable http20EnableMultipleConnections=true --property httpversion=h2multconn
     - displayName: "HTTP/3.0"
-      arguments: --variable httpVersion=3.0 --variable useHttps=true --property httpversion=h3
+      arguments: --variable httpVersion=3.0 --variable useHttps=true --variable http30EnableMultipleConnections=false --property httpversion=h3
+    - displayName: "HTTP/3.0 (mult conn)"
+      arguments: --variable httpVersion=3.0 --variable useHttps=true --variable http30EnableMultipleConnections=true --property httpversion=h3mulconn
 
 - name: clientsxThreads
   type: object

scenarios/README.md

@@ -8,7 +8,7 @@ Continuous benchmarking results are available on [this PowerBI dashboard](https:
 
 ## Requirements
 
-These jobs can be executed using the .NET Crank global tool. 
+These jobs can be executed using the .NET Crank global tool.
 [.NET Core](<http://dot.net>) is required to install the global tool.
 
 Install `crank` with the following command:
@@ -62,7 +62,7 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 
 ### Available scenarios
 
-- `plaintext`: Middleware implementation 
+- `plaintext`: Middleware implementation
 - `https`: Middleware implementation, using HTTPS
 - `endpoint`: Middleware implementation with Endpoint routing
 - `mvc`: Controller implementation
@@ -85,7 +85,7 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 
 ### Available scenarios
 
-- `json`: Middleware implementation 
+- `json`: Middleware implementation
 - `https`: Middleware implementation, using HTTPS
 - `mvc`: Controller implementation
 - `mapaction`: Endpoint routing implementation
@@ -125,8 +125,8 @@ The following scenarios are using ASP.NET CORE MVC
 - `fortunes_ef_mvc_https`
 
 The suffixes represent different database access strategies:
- 
-- No suffix: Raw ADO.NET 
+
+- No suffix: Raw ADO.NET
 - "ef" suffix: Entity Framework Core
 - "dapper" suffix: Dapper
 
@@ -159,7 +159,7 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 
 These scenarios are running several web proxies, including [YARP](https://github.com/microsoft/reverse-proxy).
 
-The downstream service returns a variable size content. By default the result is 10 bytes. 
+The downstream service returns a variable size content. By default the result is 10 bytes.
 
 ### Sample
 
@@ -261,7 +261,7 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 These scenarios measure the performance of different Grpc  server and clients implementations.
 
 - Go
-- Native (C) 
+- Native (C)
 - ASP.NET
 
 ### Sample
@@ -287,13 +287,13 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 
 #### Arguments
 
-- Number of streams: 
+- Number of streams:
   - `--variable streams=1`
   - `--variable streams=70`
-- Number of connections: 
+- Number of connections:
   - `--variable connections=1`
   - `--variable connections=28`
-- Protocol: 
+- Protocol:
   - `--variable protocol=h2`
   - `--variable protocol=h3`
   - `--variable protocol=h2c`
@@ -345,15 +345,15 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 
 #### Arguments
 
-- Scenario: 
+- Scenario:
   - `--variable scenario=broadcast`
   - `--variable scenario=echo`
   - `--variable scenario=echoAll`
-- Transport: 
+- Transport:
   - `--variable transport=websockets`
   - `--variable transport=serversentevents`
   - `--variable transport=longpolling`
-- Protocol: 
+- Protocol:
   - `--variable protocol=json`
   - `--variable protocol=messagepack`
 
@@ -397,7 +397,7 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 
 These scenarios are running various distributed cache benchmarks.
 
-For all the scenarios, the store is initialized with `CacheCount` cache entries. Each request will issue a read or a write based on the `WriteRatio` 
+For all the scenarios, the store is initialized with `CacheCount` cache entries. Each request will issue a read or a write based on the `WriteRatio`
 argument choosing a key randomly. The HTTP response won't contain the cache entry data so that it doesn't impact the raw store perf measurement.
 
 ### Sample
@@ -512,6 +512,9 @@ crank --config https://raw.githubusercontent.com/aspnet/Benchmarks/main/scenario
 - Enable multiple HTTP/2.0 connections:
   - `--variable http20EnableMultipleConnections=true` (default)
   - `--variable http20EnableMultipleConnections=false`
+- Enable multiple HTTP/3.0 connections (from .NET 9.0):
+  - `--variable http30EnableMultipleConnections=true` (default)
+  - `--variable http30EnableMultipleConnections=false`
 - Whether to use WinHttpHandler instead of SocketsHttpHandler:
   - `--variable useWinHttpHandler=false` (default)
   - `--variable useWinHttpHandler=true` -- *requires Windows*
@@ -538,7 +541,7 @@ These scenarios provide benchmarks to help improve the performance of the .NET G
 ### Sample
 
 ```
-crank --config https://raw.githubusercontent.com/dotnet/performance/main/src/benchmarks/gc/scenarios/CrankConfiguration.yaml --scenario 2gb-pinning --profile aspnet-citrine-win --application.framework net9.0 
+crank --config https://raw.githubusercontent.com/dotnet/performance/main/src/benchmarks/gc/scenarios/CrankConfiguration.yaml --scenario 2gb-pinning --profile aspnet-citrine-win --application.framework net9.0
 ```
 ### Available scenarios
 

scenarios/httpclient.benchmarks.yml

@@ -25,6 +25,7 @@ variables:
     # HttpClient-specific parameters:
     http11MaxConnectionsPerServer: 0 #unlimited
     http20EnableMultipleConnections: true
+    http30EnableMultipleConnections: true
     useWinHttpHandler: false
     useHttpMessageInvoker: false
     useDefaultRequestHeaders: false
@@ -102,6 +103,7 @@ scenarios:
         numberOfHttpClients: 1
         concurrencyPerHttpClient: '{% if cores > 0 %}{{cores}}{% else %}1{% endif %}'
         http20EnableMultipleConnections: false
+        http30EnableMultipleConnections: false
         useHttpMessageInvoker: true
         host: '{{secondaryAddress}}'
 
@@ -168,7 +170,7 @@ jobs:
     isConsoleApp: true
     waitForExit: true
     timeout: 1200 #seconds
-    arguments: '--address {{host}} --port {{serverPort}} --useHttps {{useHttps}} --path /{{scenario}} --scenario {{scenario}} --httpVersion {{httpVersion}} --numberOfHttpClients {{numberOfHttpClients}} --concurrencyPerHttpClient {{concurrencyPerHttpClient}} --http11MaxConnectionsPerServer {{http11MaxConnectionsPerServer}} --http20EnableMultipleConnections {{http20EnableMultipleConnections}} --useWinHttpHandler {{useWinHttpHandler}} --useHttpMessageInvoker {{useHttpMessageInvoker}} --collectRequestTimings {{collectRequestTimings}} --contentSize {{requestContentSize}} --contentWriteSize {{requestContentWriteSize}} --contentFlushAfterWrite {{requestContentFlushAfterWrite}} --contentUnknownLength {{requestContentUnknownLength}} {{headersDictionary[requestHeaders]}} --generatedStaticHeadersCount {{generatedStaticRequestHeadersCount}} --generatedDynamicHeadersCount {{generatedDynamicRequestHeadersCount}} --useDefaultRequestHeaders {{useDefaultRequestHeaders}} --warmup {{warmup}} --duration {{duration}}'
+    arguments: '--address {{host}} --port {{serverPort}} --useHttps {{useHttps}} --path /{{scenario}} --scenario {{scenario}} --httpVersion {{httpVersion}} --numberOfHttpClients {{numberOfHttpClients}} --concurrencyPerHttpClient {{concurrencyPerHttpClient}} --http11MaxConnectionsPerServer {{http11MaxConnectionsPerServer}} --http20EnableMultipleConnections {{http20EnableMultipleConnections}} --http30EnableMultipleConnections {{http30EnableMultipleConnections}} --useWinHttpHandler {{useWinHttpHandler}} --useHttpMessageInvoker {{useHttpMessageInvoker}} --collectRequestTimings {{collectRequestTimings}} --contentSize {{requestContentSize}} --contentWriteSize {{requestContentWriteSize}} --contentFlushAfterWrite {{requestContentFlushAfterWrite}} --contentUnknownLength {{requestContentUnknownLength}} {{headersDictionary[requestHeaders]}} --generatedStaticHeadersCount {{generatedStaticRequestHeadersCount}} --generatedDynamicHeadersCount {{generatedDynamicRequestHeadersCount}} --useDefaultRequestHeaders {{useDefaultRequestHeaders}} --warmup {{warmup}} --duration {{duration}}'
 
   wrk:
     source: