Tuesday, January 9, 2018

WCF - Concurrency, InstanceContextMode and SessionMode & request handling behavior

Though WebAPI is penetrating the service world in .Net WCF still has a space in .Net. The main reason is the support for protocols other than http and there are lot of legacy systems out there which has to be maintained till the world fully moved to Cloud native.

WCF is easy to develop but when it comes to production, there are so many internal things to be known than just writing the service. The major area to be known before putting any WCF service is how the service is going to serve requests. Will there be one service object? How many requests will be processed in parallel? In there are more than one requests getting processed how many service objects will be there to process? Is it one to one or one to many? If it is one service object processing  many requests what is means to thread safety? Service object here means the object of class which implements the service contract interface? Hope everyone knows who creates the object of implementation class.

There are so many articles out there which explains how the service process requests. So this post is to consolidate all in one place to avoid googling.

MSDN Article 1 & Article 2 - These are the best article if someone have enough understanding about the concept. If someone is really fresh to WCF this is not a great place to start

Sivaprasad Koirala on Instancing & concurrency - This is good place to start for freshers. He explains with diagrams and there is a good sample which can be used to understand the behavior. But it talks only instancing and concurrency. When instancing combined with Concurrency mode and Session, the behavior will change.

If we are in hurry and have to understand the behavior quickly the best site is given below.


It explains with combinations and tell what happens in each combination. But the unlucky blog post has no comments till now even after 3 years.

It is not ending here. There is throttling behavior which might need to be tweaked. Also the security mode which gets enabled automatically for some bindings though we don't need it and reduce throughput.

Tuesday, January 2, 2018

ReST based Web Sites & ReSTful Navigation


ReST is highly considered as architectural pattern for integrating systems. It mainly uses APIs
point to resource and do operation on those. The URL often follows a cleaner hierarchical path pattern without much key value pairs than conventional key value pair based URL schemes. People often follow the ReST based URL format for APIs but not widely accepted for web sites.

This post aims to investigate on bringing this ReST based URL schemes to web sites similar to APIs.

Why should I do create web sites as ReSTful URLs

The same benefit of ReST based API URLs applies here as well. The URL will be easy to remember. It is easy to have separation of concerns. New features can be totally implemented separately. using their own area / virtual directories. No need to mix with existing screens even it they are related. Simple and short URLs than lengthy story telling URLs.

eg: www.mycompany.com/employeelist.aspx?empId=1 can be easily represented via

Why there are not much web sites following ReSTful pattern

One reason could be the difficulty to follow the pattern. If it is product company the development team will get more freedom to select URLs. Again if the product owners don't know what is ReST and the advantages they may influence the URL pattern. In the other side the consulting industry is heavily driven by client demands. Though regular clients may not ask for particular URLs, semi technical clients may ask.

Some tips for ReSTful web site URLS

Below are some tips to design web site URLs in resource oriented way

No operation oriented screens

The screens should be pointing to resources to display those. For example the below URL displays the employee resource.


Lists and details screens

If the URLs display the resource how to edit them? Which screen to edit those resources? The better way is to use lists and details mechanism.

If we display the resources in a list those can be edited in the same list itself and save using a button. Here the single page application (SPA) concept helps us than navigating to an edit page.

Circular navigation

If the resources have circular relation the navigation may also becomes circular.
For example, the employee page may show the department where he belongs to as hyper link. Clicing that will navigate to dept page where it lists the employees in that dept including the manager. Going to the manager page may display employees under him and clicking on the same employee's link there will end up in same employees page where we started.

Multiple navigation paths

Similarly there would be multiple navigation paths to reach one resource. For example the home page of company may show the departments and various project it is doing currently. Navigating to department as well as project may end up in same employee page.

A powerful search experience to navigate

If a resource is buried under the hierarchy, it would be difficult to find it without multiple clicks. So better to have a search mechanism where the resource can be searched and navigated on clicking the associated URLs.