Introduction
This article describes how to build ASP.NET applications using n-tier architecture. The benefits of having n-tier architecture is that all the modules having dedicated functionality will be independent of each other. Changing one tier will not effect other tiers and there is no single point of failure even if some tier is not working.
Background
In a typical n-tier application there will be 4 Layers. The bottom most layer is the Data layer which contains the tables and stored procedures, scaler function, table values function. This Data layer is typically the database engine itself. We will be using SqlServer
as the data layer in our example.
On top of Data Layer, we have a Data Access Layer (DAL). This layer is responsible for handling Database related tasks i.e. only data access. This Data access layer
is created as a separate solution so that the changes in DAL
only need the recompilation of DAL and not the complete website. The benefit of having this layer as a separate solution is that in case the database engine is changes we only need to change the DAL
and the other areas of the website need not be changed and recompiled. Also the changes in other areas outside this solution will not demand for DAL
recompilation.
On top of DAL, we have our Business Logic Layer(BLL)
. BLL
contains all the calculations and Business Rule validations that are required in the application. It is also in a separate solution for the reason that if the Business rules change or the calculations change we only need to recompile the BLL
and the other layers of the application will remain unaffected.
Finally on top of BLL
we have our Presentation Layer. The Presentation layer for an ASP.NET web forms application is all the Forms (apsx
pages and their code behinds) and the classes contained in the App_Code folder. The Presentation layer is responsible for taking the user input, showing the data to the user and mainly performing input data validation.
Note: Input data filtration and validation is typically done at the Presentation Layer(Both client side and server side). The business Rule validation will be done at the BLL
.
So to visualize the above mentioned architecture:
Note: The Data Access Layer in this article was written using classic ADO.NET
, due to which the amount of code in DAL
is little too much. Nowadays using ORMs like Entity framework
to generate the DAL
is recommended. The DAL
code will be generated by ORM itself.
Using the code
Let us develop a small Toy ASP.NET
application that will use n-tier architecture. We will develop a small Employee Management application for the NorthWind
Database. (For simplicity, I have removed all other tables from the DB and some columns from the Employee table). This application should be able to perform the basic CRUD operations on the DB.
The solution for this application will contain separate projects for DAL
and BLL
. The Data Layer will be SqlServer
. The Presentation Layer is an ASP.NET website running on top of these projects.
The Data Layer
The data layer in this example contain only one table called Employee. The data layer also contains the stored procedures for all the basic operations on the Employee table. So let us look at the table and all the stored Procedures we have in our Data Layer.
Now we will create a set of stored procedures to perform the operations on the Employees Table.
CREATE PROCEDURE dbo.AddNewEmployee
(
@LastName nvarchar(20),
@FirstName nvarchar(10),
@Title nvarchar(30),
@Address nvarchar(60),
@City nvarchar(15),
@Region nvarchar(15),
@PostalCode nvarchar(10),
@Country nvarchar(15),
@Extension nvarchar(4)
)
AS
insert into Employees
(LastName, FirstName, Title, Address, City, Region, PostalCode, Country, Extension)
values
(@LastName, @FirstName, @Title, @Address, @City, @Region, @PostalCode, @Country, @Extension)
RETURN
CREATE PROCEDURE dbo.DeleteEmployee
(
@empId int
)
AS
delete from Employees where EmployeeID = @empId
RETURN
CREATE PROCEDURE dbo.GetEmployeeDetails
(
@empId int
)
AS
Select * from Employees where EmployeeID = @empId
RETURN
CREATE PROCEDURE dbo.GetEmployeeList
AS
Select * from Employees
RETURN
CREATE PROCEDURE dbo.UpdateEmployee
(
@EmployeeID int,
@LastName nvarchar(20),
@FirstName nvarchar(10),
@Title nvarchar(30),
@Address nvarchar(60),
@City nvarchar(15),
@Region nvarchar(15),
@PostalCode nvarchar(10),
@Country nvarchar(15),
@Extension nvarchar(4)
)
AS
update Employees
set
LastName = @LastName,
FirstName = @FirstName,
Title = @Title,
Address = @Address,
City = @City,
Region = @Region,
PostalCode = @PostalCode,
Country = @Country,
Extension = @Extension
where
EmployeeID = @EmployeeID
RETURN
Now we have our Data Layer ready.
The Data Access Layer
Now we will go ahead and create a Data Access Layer for our application. The data access layer will contain 2 main type of classes. A set of classes that will represent the Table entities. And classes to perform the CRUD
operations on the database.
The Employee
class in the above diagram is the Entity that will represent the Employee
table. This class has been created so that the Layers above the DAL
will use this class to perform operations in Employee table and they need not worry about the table schema related details.
public class Employee
{
int employeeID;
string lastName;
string firstName;
string title;
string address;
string city;
string region;
string postalCode;
string country;
string extension;
public int EmployeeID
{
get
{
return employeeID;
}
set
{
employeeID = value;
}
}
public string LastName
{
get
{
return lastName;
}
set
{
lastName = value;
}
}
public string FirstName
{
get
{
return firstName;
}
set
{
firstName = value;
}
}
public string Title
{
get
{
return title;
}
set
{
title = value;
}
}
public string Address
{
get
{
return address;
}
set
{
address = value;
}
}
public string City
{
get
{
return city;
}
set
{
city = value;
}
}
public string Region
{
get
{
return region;
}
set
{
region = value;
}
}
public string PostalCode
{
get
{
return postalCode;
}
set
{
postalCode = value;
}
}
public string Country
{
get
{
return country;
}
set
{
country = value;
}
}
public string Extension
{
get
{
return extension;
}
set
{
extension = value;
}
}
}
The EmployeeDBAccess
class expose the methods to perform the CRUD
operations on the Employee table.
public class EmployeeDBAccess
{
public bool AddNewEmployee(Employee employee)
{
SqlParameter[] parameters = new SqlParameter[]
{
new SqlParameter("@LastName", employee.LastName),
new SqlParameter("@FirstName", employee.FirstName),
new SqlParameter("@Title", employee.Title),
new SqlParameter("@Address", employee.Address),
new SqlParameter("@City", employee.City),
new SqlParameter("@Region", employee.Region),
new SqlParameter("@PostalCode", employee.PostalCode),
new SqlParameter("@Country", employee.Country),
new SqlParameter("@Extension", employee.Extension)
};
return SqlDBHelper.ExecuteNonQuery("AddNewEmployee", CommandType.StoredProcedure, parameters);
}
public bool UpdateEmployee(Employee employee)
{
SqlParameter[] parameters = new SqlParameter[]
{
new SqlParameter("@EmployeeID", employee.EmployeeID),
new SqlParameter("@LastName", employee.LastName),
new SqlParameter("@FirstName", employee.FirstName),
new SqlParameter("@Title", employee.Title),
new SqlParameter("@Address", employee.Address),
new SqlParameter("@City", employee.City),
new SqlParameter("@Region", employee.Region),
new SqlParameter("@PostalCode", employee.PostalCode),
new SqlParameter("@Country", employee.Country),
new SqlParameter("@Extension", employee.Extension)
};
return SqlDBHelper.ExecuteNonQuery("UpdateEmployee", CommandType.StoredProcedure, parameters);
}
public bool DeleteEmployee(int empID)
{
SqlParameter[] parameters = new SqlParameter[]
{
new SqlParameter("@empId", empID)
};
return SqlDBHelper.ExecuteNonQuery("DeleteEmployee", CommandType.StoredProcedure, parameters);
}
public Employee GetEmployeeDetails(int empID)
{
Employee employee = null;
SqlParameter[] parameters = new SqlParameter[]
{
new SqlParameter("@empId", empID)
};
using (DataTable table = SqlDBHelper.ExecuteParamerizedSelectCommand("GetEmployeeDetails", CommandType.StoredProcedure, parameters))
{
if (table.Rows.Count == 1)
{
DataRow row = table.Rows[0];
employee = new Employee();
employee.EmployeeID = Convert.ToInt32(row["EmployeeID"]);
employee.LastName = row["LastName"].ToString();
employee.FirstName = row["FirstName"].ToString();
employee.Title = row["Title"].ToString();
employee.Address = row["Address"].ToString();
employee.City = row["City"].ToString();
employee.Region = row["Region"].ToString();
employee.PostalCode = row["PostalCode"].ToString();
employee.Country = row["Country"].ToString();
employee.Extension = row["Extension"].ToString();
}
}
return employee;
}
public List<employee> GetEmployeeList()
{
List<employee> listEmployees = null;
using (DataTable table = SqlDBHelper.ExecuteSelectCommand("GetEmployeeList", CommandType.StoredProcedure))
{
if (table.Rows.Count > 0)
{
listEmployees = new List<employee>();
foreach (DataRow row in table.Rows)
{
Employee employee = new Employee();
employee.EmployeeID = Convert.ToInt32(row["EmployeeID"]);
employee.LastName = row["LastName"].ToString();
employee.FirstName = row["FirstName"].ToString();
employee.Title = row["Title"].ToString();
employee.Address = row["Address"].ToString();
employee.City = row["City"].ToString();
employee.Region = row["Region"].ToString();
employee.PostalCode = row["PostalCode"].ToString();
employee.Country = row["Country"].ToString();
employee.Extension = row["Extension"].ToString();
listEmployees.Add(employee);
}
}
}
return listEmployees;
}
}
</employee></employee></employee>
The class SqlDbHelper
is a wrapper class for ADO.NET
functions providing a more simpler interface to use by the rest of DAL.
class SqlDBHelper
{
const string CONNECTION_STRING = @"Data Source=.\SQLEXPRESS;AttachDbFilename=|DataDirectory|\NORTHWND.MDF;Integrated Security=True;User Instance=True";
internal static DataTable ExecuteSelectCommand(string CommandName, CommandType cmdType)
{
DataTable table = null;
using (SqlConnection con = new SqlConnection(CONNECTION_STRING))
{
using (SqlCommand cmd = con.CreateCommand())
{
cmd.CommandType = cmdType;
cmd.CommandText = CommandName;
try
{
if (con.State != ConnectionState.Open)
{
con.Open();
}
using (SqlDataAdapter da = new SqlDataAdapter(cmd))
{
table = new DataTable();
da.Fill(table);
}
}
catch
{
throw;
}
}
}
return table;
}
internal static DataTable ExecuteParamerizedSelectCommand(string CommandName, CommandType cmdType, SqlParameter[] param)
{
DataTable table = new DataTable();
using (SqlConnection con = new SqlConnection(CONNECTION_STRING))
{
using (SqlCommand cmd = con.CreateCommand())
{
cmd.CommandType = cmdType;
cmd.CommandText = CommandName;
cmd.Parameters.AddRange(param);
try
{
if (con.State != ConnectionState.Open)
{
con.Open();
}
using (SqlDataAdapter da = new SqlDataAdapter(cmd))
{
da.Fill(table);
}
}
catch
{
throw;
}
}
}
return table;
}
internal static bool ExecuteNonQuery(string CommandName, CommandType cmdType, SqlParameter[] pars)
{
int result = 0;
using (SqlConnection con = new SqlConnection(CONNECTION_STRING))
{
using (SqlCommand cmd = con.CreateCommand())
{
cmd.CommandType = cmdType;
cmd.CommandText = CommandName;
cmd.Parameters.AddRange(pars);
try
{
if (con.State != ConnectionState.Open)
{
con.Open();
}
result = cmd.ExecuteNonQuery();
}
catch
{
throw;
}
}
}
return (result > 0);
}
}
Note: If we use any ORM
(Object Relation Mapper) then DAL need not be written. The ORM
will generate all the DAL code. Entity framework
is one of the best ORMs
available. This DAL can simply be replaced with a class library containing the Entity Framework
generated Entities and Contexts.
The Business Logic Layer
The business logic layer will have a reference to the DAL and will mainly perform Business rule validation and business logic specific calculations. In out example, I will write a simple BLL
that will govern the IO between the DAL
and Presentation layer. In real applications the BLL
will contain more logic and code.
public class EmployeeHandler
{
EmployeeDBAccess employeeDb = null;
public EmployeeHandler()
{
employeeDb = new EmployeeDBAccess();
}
public List<employee> GetEmployeeList()
{
return employeeDb.GetEmployeeList();
}
public bool UpdateEmployee(Employee employee)
{
return employeeDb.UpdateEmployee(employee);
}
public Employee GetEmployeeDetails(int empID)
{
return employeeDb.GetEmployeeDetails(empID);
}
public bool DeleteEmployee(int empID)
{
return employeeDb.DeleteEmployee(empID);
}
public bool AddNewEmployee(Employee employee)
{
return employeeDb.AddNewEmployee(employee);
}
}
The Presentation Layer
The presentation layer now contains only a set of pages and code behinds and it will use the BLL
and the the Employee class to perform all the operations. The add Operation can be seen as an example how the BLL
is being used to perform an operation.
Employee emp = new Employee();
emp.LastName = txtLName.Text;
emp.FirstName = txtFName.Text;
emp.Address = txtAddress.Text;
emp.City = txtCity.Text;
emp.Country = txtCountry.Text;
emp.Region = txtRegion.Text;
emp.PostalCode = txtCode.Text;
emp.Extension = txtExtension.Text;
emp.Title = txtTitle.Text;
EmployeeHandler empHandler = new EmployeeHandler();
if (empHandler.AddNewEmployee(emp) == true)
{
Response.Redirect("Default.aspx");
}
Note: All the CRUD operations have been implemented. Please refer tio the sample code for all the details. When we run the application we can see all the EDIT/UPDATE, DELETE and ADD operations in action.
Point of Interest
I created this small application to demonstrate application development using n-tier architecture. The demo application has been created to show the basic idea behind the 3-tier architecture. There are many things that are still missing from this sample from the completion perspective. Client side validation and server side validation in presentation layer, Business rule validation and calculations in BLL are some missing things.
Since the idea here was to talk about how to put n-tier architecture in actual code, I think this article might have provided some useful information on that. I hope this has been informative.
[UPDATE] Note: In this article I am reusing the Employee
model in the presentation layer. This model is defined in Data Access Layer. Due to this the presentation layer has to refer to the data access layer. This is not ideal in the real world scenarios(as pointed out in many of the comments below). Ideal solution for this would be to have two different models for Employee
. the current model which is defined in the data access layer can be called as the data model and the business logic layer can create a model for employee which will be called as domain model. The business logic layer will then have to contain the code for mapping the data model to the domain model and vice versa. This mapping can be done either manually or a tool like AutoMapper
can also be used to perform such mapping. With this change the presentation layer need not refer to the data access layer but it can refer to the business logic layer and use the Employee domain model from that.
In this article the n-tier architecture is specifically a data centric n-tier and not a domain centric one. If we need to design the application in a domain centric n-tier architecture then we need to follow a different way of organizing our layers. But perhaps that is a topic which deserves a separate discussion altogether but I wanted to point out the possibility of a domain centric n-tier architecture in this article.
History
- 12 November 2013: Added explanatory text to explain the design flaw.
- 13 August 2012: First version