Leveraging the Power of Asynchrony in ASP.NET [转]
Asynchronous programming has had a lot of attention in the last couple of years and there are two key reasons for this: First it helps in providing a better user experience by not blocking the UI thread which avoids the hanging of the UI screen until the processing is done, and second, it helps in scaling up the system drastically without adding any extra hardware.
But writing asynchronous code and managing the thread gracefully was a tedious task. But as the benefits are tremendous, many new technologies and old technologies have started embracing it. Microsoft has also invested a lot in it since .NET 4.0 and then in .NET 4.5, they made it simpler than ever with the introduction of async and the await keyword.
However, asynchrony has been usable in ASP.NET since the beginning but has never gotten the required attention. And given the way requests are handled by ASP.NET and IIS, asynchrony could be much more beneficial and we can easily scale up our ASP.NET applications drastically. With the introduction of new programming constructs like async and await, it’s high time that we start leveraging the power of asynchronous programming.
In this post, we are going to discuss the way requests are processed by IIS and ASP.NET, then we will see the places where we can introduce asynchrony in ASP.NET and discuss various scenarios where we can gain maximum benefits from it.
How are requests handled?
Every ASP.NET request has to go through IIS then is ultimately handled by ASP.NET handlers. A request is first received by IIS and after initial processing, forwards to ASP.NET which actually handles the request (for an ASP.NET request) and generates the response. This response is then sent back to the client via IIS. IIS has some worker threads that are responsible for taking the request from the queue and executing IIS modules and then forwarding the request to ASP.NET queue. ASP.NET doesn’t create any thread or own any thread pool to handle the request, instead it uses the CLR thread pool and gets the thread from there to handle the requests. The IIS module calls ThreadPool.QueueUserWorkItem which queues the request to CLR worker threads. As we know, the CLR thread pool is managed by CLR and self-tuned (meaning it creates/destroys the threads based on need). Also, we should keep in mind that creating and destroying a thread is always a heavy task and that’s why this pool allows us to reuse the same thread for multiple tasks. So let’s see pictorially the way a request gets processed.

In the above pic, we can see that a request is first received by HTTP.sys and added in the queue of the corresponding application pool at kernel level. One IIS worker thread takes the request from the queue and passes it to the ASP.NET queue after its processing. The request may get returned from IIS itself if it is not an ASP.NET Request. A thread from the CLR thread pool gets assigned to the thread that’s responsible for processing the request.
When should Asynchrony be used in ASP.NET?
Any request can broadly be divided in two types:
1- CPU Bound
2- I/O Bound
A CPU bound request needs CPU time and is executed in same process, while I/O bound requests are blocking in nature and dependent on other modules which do the I/O operations and return the response. Blocking requests are one of the major roadblocks for high scalable applications and in most of our web application we been wasting lots of time while waiting for I/O operations. These are the following scenarios where asynchrony should be used:
1- For I/O Bound requests including
- Database access
- Reading/Writing Files
- Web Service calls
- Accessing Network resources
2- Event driven requests like SignalR
3- Where we need to get the data from multiple sources
Let’s create an example where we would create a simple Synchronous Page and then will convert it to an asynchronous page. For this example, I have put a delay of 1000ms (to mimic some heavy calls like Database/web service calls, etc.) and also downloaded one page using the WebClient as follows:
protected void Page_Load(object sender, EventArgs e)
{
System.Threading.Thread.Sleep();
WebClient client = new WebClient();
string downloadedContent = client.DownloadString("https://msdn.microsoft.com/en-us/library/hh873175%28v=vs.110%29.aspx");
dvcontainer.InnerHtml = downloadedContent;
}
Now we will convert this page to an asynchronous page. There are mainly three steps involved here
- Change it to asynchronous page by adding Async = true in the page directive as follows:
<%@ Page Language="C#" AutoEventWireup="true" CodeBehind="Home.aspx.cs" Inherits="AsyncTest.Home" Async="true" AsyncTimeout="3000" %>
I have also added AsyncTimeout (which is optional) but we can define based on our requirement.
- Convert the method to an Asynchronous one. Here we convert Thread.Sleep and client.DownloadString to the asynchronous method as follows:
private async Task AsyncWork()
{
await Task.Delay(); WebClient client = new WebClient(); string downloadedContent = await client.DownloadStringTaskAsync("https://msdn.microsoft.com/en-us/library/hh873175%28v=vs.110%29.aspx "); dvcontainer.InnerHtml = downloadedContent; }
- Now we can call this method directly at Page_Load and make it asynchronous as follows:
protected async void Page_Load(object sender, EventArgs e)
{
await AsyncWork();
}
But here Page_Load returns the type is async void which should be avoided in almost all cases. As we know the flow of the page life cycle, Page_Load is a part of life cycle and if we set it as async then there could be scenarios and other events where the lifecycle may get executed even if the page load is still running. It is highly recommended we use RegisterAsyncTask which allows us to register the asynchronous method which is executed during the lifecycle while it is most appropriate and avoids any issue. So we should write it as follows:
protected void Page_Load(object sender, EventArgs e)
{
RegisterAsyncTask(new PageAsyncTask(AsyncWork));
}
Now we have converted our page as asynchronous and it won’t be a blocking request.
I deployed both applications on IIS 8.5 and tested them with burst load. For the same machine configuration, the synchronous page was able to take just 1,000 requests in 2-3 seconds while the asynchronous page was able to serve more than 2,200 requests. After that, we started getting Timeout or Server Not Available errors. Although the average request processing time isn’t much different, we were able to serve more than 2X the requests just by making the page asynchronous. If that isn’t proof we should leverage the power of asynchronous programming, I don’t know what is!
There are some other places where we can introduce asynchrony in ASP.NET too:
1- By writing asynchronous modules
2- By writing Asynchronous HTTP Handlers using IHttpAsyncHandler or HttpTaskAsyncHandler
3- Using web sockets or SignalR
Conclusion
In this post, we discussed asynchronous programming and saw that with the help of new async and await keywords, writing asynchronous code is very easy. We covered the topic of request processing by IIS and ASP.NET and discussed the scenarios where asynchrony can be more fruitful, and we also created a simple example and discussed the benefits of asynchronous pages. Last we explored some places where asynchrony can be leveraged in ASP.NET.
Thanks for reading!
http://news.oneapm.com/advantage/
Leveraging the Power of Asynchrony in ASP.NET [转]的更多相关文章
- (翻译)《Hands-on Node.js》—— Introduction
今天开始会和大熊君{{bb}}一起着手翻译node的系列外文书籍,大熊负责翻译<Node.js IN ACTION>一书,而我暂时负责翻译这本<Hands-on Node.js> ...
- Announcing Microsoft Research Open Data – Datasets by Microsoft Research now available in the cloud
The Microsoft Research Outreach team has worked extensively with the external research community to ...
- WM-N-BM-09 WM-N-BM-14
USI Delivers WICED Module to Gain Great Success Customers Broadcom’s Wireless Internet Connectivity ...
- Hadoop vs Elasticsearch – Which one is More Useful
Hadoop vs Elasticsearch – Which one is More Useful Difference Between Hadoop and Elasticsearch H ...
- 通过利用immutability的能力编写更安全和更整洁的代码
通过利用immutability的能力编写更安全和更整洁的代码 原文:Write safer and cleaner code by leveraging the power of "Imm ...
- 【内推】微软北京深圳招聘多名Cloud Solution Architect
Azure is the most comprehensive, innovative and flexible cloud platform today and Microsoft is hirin ...
- [ABP] ASP.NET Zero 5.6.0 之 ASP.NET Zero Power Tools 上手日志
之前破解了这个工具后,却没有使用它. 现在使用这个小工具,帮我完成创建Entity类,Dto类,AppService类,View视图等DDD相关工作以及Entity Framework Migrati ...
- [ABP] ASP.NET Zero 5.6.0 之 ASP.NET Zero Power Tools 破解日志
两个要破解Patch的关键dll名称:AspNetZeroRadToolVisualStudioExtension.dll和AspNetZeroRadTool.dll AspNetZeroRadToo ...
- 【C#】转一篇MSDN杂志文:ASP.NET Pipeline: Use Threads and Build Asynchronous Handlers in Your Server-Side Web Code
序:这是一篇发表在2003年6月刊的MSDN Magazine的文章,现在已经不能在线阅读,只提供chm下载.讲的是异步请求处理那些事,正是我上一篇博文涉及的东西(BTW,事实上这篇杂志阐述了那么搞然 ...
随机推荐
- Javascript 数组自定义排序,并获取排序后的保存原索引的同序数组(堆排序实现)
比如数组A: [ 0: 5, 1: 2, 2: 4, 3: 3, 4: 1 ] 排序后的结果为:[1, 2, 3, 4, 5],但是有时候会有需求想要保留排序前的位置到一个同位数组里,如前例则为:[4 ...
- angularJS--神奇的$scope
我们在使用angularJS时,$scope对于angularJS是非常重要的,它是angularJS的基础,但$scope到底是什么呢?下面进行一些介绍. 1.$scope是一个普通的js对象 2. ...
- java线程中生产者与消费者的问题
一.概念 生产者与消费者问题是一个金典的多线程协作的问题.生产者负责生产产品,并将产品存放到仓库:消费者从仓库中获取产品并消费.当仓库满时,生产者必须停止生产,直到仓库有位置存放产品:当仓库空时,消费 ...
- Codevs 1198 国王游戏 2012年NOIP全国联赛提高组
1198 国王游戏 2012年NOIP全国联赛提高组 时间限制: 1 s 空间限制: 128000 KB 题目等级 : 钻石 Diamond 题目描述 Description 恰逢 H 国国庆,国王邀 ...
- Java Lambda简明教程(一)
Lambda表达式背景 许多热门的编程语言如今都有一个叫做lambda或者闭包的语言特性,包括比较经典的函数式编程语言Lisp,Scheme,也有稍微年轻的语言比如JavaScript,Python, ...
- 选择第n小的元素之python实现源码
def partition(A, p, r): j = p+1 for i in range(p+1, r+1): if(A[i] < A[p]): tmp = A[i] A[i] = A[j] ...
- Name control
static: (Page 406) In both C and C++ the keyword static has two basic meanings, which unfortunately ...
- Hibernate的检索策略
hibernate 的中的session依照load()和get()按照参数的制定OID(ObjctID)去加载一个持久化对象.另外Query.list()方法则按照HQL语句去加载持久化的对象. 以 ...
- yum 安装 依赖报错
今天使用yum安装的时候 报错: Error: Multilib version problems found. This often means that the root cause 应该是yum ...
- thinkphp给图片打水印不清晰
项目中打印条形码的函数,从thinkphp自带的water函数修改而来的. 贴上代码: /** * water2 * 改写thinkphp的water函数更强健的函数,增加了写入位置参数 去掉了alp ...