Tài liệu Module 5: Managing the Business Logic Layer pptx

Contents Overview 1 Implementing State in a Shopping Cart 2 Adapting the User’s View of the Site 31 Commerce Server 2000 Personalization... Lab 5: Using Business Desk Students will im

Trang 1

Contents

Overview 1

Implementing State in a Shopping Cart 2

Adapting the User’s View of the Site 31

Commerce Server 2000 Personalization

Trang 2

to represent any real individual, company, product, or event, unless otherwise noted Complying with all applicable copyright laws is the responsibility of the user No part of this document may

be reproduced or transmitted in any form or by any means, electronic or mechanical, for any purpose, without the express written permission of Microsoft Corporation If, however, your only means of access is electronic, permission to print one copy is hereby granted

Microsoft may have patents, patent applications, trademarks, copyrights, or other intellectual property rights covering subject matter in this document Except as expressly provided in any written license agreement from Microsoft, the furnishing of this document does not give you any license to these patents, trademarks, copyrights, or other intellectual property

Microsoft, Active Directory, ActiveX, BizTalk Server, FoxPro, FrontPage, Hotmail, Jscript, Outlook, PowerPoint, Visio, Visual Basic, Visual C#, Visual C++, Visual Studio, Windows, and Windows NT are either registered trademarks or trademarks of Microsoft Corporation in the U.S.A and/or other countries

Other product and company names mentioned herein may be the trademarks of their respective owners

Trang 3

Instructor Notes

This module examines the problem of maintaining state in an e-business application It focuses on several approaches that can be used for saving useful information between each request and response cycle Where appropriate, this module also covers some of the useful features of Microsoft® Commerce

Server 2000, including the Commerce Server Business Desk

After completing this module, students will be able to:

Describe how to manage state for a Web shopping cart

List some of the issues related to state and scalability and explain how they can be addressed

Identify the role of MicrosoftApplication Center 2000 in software scaling

List the mechanisms that can be used for identifying the users of an e-business Web site

Explain how personalization and customization can be used to improve the usability of a site

Explain how to use the tools in Commerce Server 2000 to implement targeted campaigns, promotions, and cross selling in your e-business

Materials and Preparation

This section provides the materials and preparation tasks that you need to teach this module

Required Materials

To teach this module, you need the following materials:

Microsoft PowerPoint® file 2260A_05.ppt

Module 5, “Managing the Business Logic Layer”

Lab 5, “Using Business Desk”

Preparation Tasks

To prepare for this module, you should:

Read all of the materials for this module

Complete the lab

Become familiar with Business Desk

Read about Microsoft’s latest Customer Relationship Management (CRM) initiatives at http://www.microsoft.com/business/crm

Presentation:

100 Minutes

Lab:

70 Minutes

Trang 4

Module Strategy

Use the following strategy to present this module:

Implementing State in a Shopping Cart Ensure that students understand what is meant by state before describing the problem in maintaining state with Hypertext Transfer Protocol (HTTP) After explaining the issues with HTTP, describe the techniques available to

aWeb developer when implementing state on an e-business site Ensure that students understand the connection between state and shopping carts Try not to delve too far into personalization because it is also covered in the next section

State vs Scalability Discuss the problems with scalability in the context of coping with state management Although some of these problems have been covered earlier in this module, this will provide an opportunity to introduce some best

programming techniques that can be used to resolve some of these issues Conclude this section with a discussion on how to support software scalability using Application Center 2000

Membership and Authentication Point out the differences between state management and membership Some advantages of implementing membership are covered in this section However, you can also introduce your own examples to emphasize what membership can offer users

Adapting the User’s View of the Site This builds directly on the concept of membership CRM is covered in one

of the topics Focus on the business needs of customers rather than on the technical implementation of CRM

Commerce Server 2000 Personalization Tools Commerce Server 2000 is extensive in scope and it will not be possible to discuss all of its features in this section Tell the students that only a few of the components—Business Desk, Profiling system, Business Analytics system, and Targeting system—are discussed in this section The discussion must be limited to how these tools help a developer to build relationships with clients by using powerful tools

Lab 5: Using Business Desk Students will implement state management in an e-commerce application by using Commerce Server 2000 The Commerce Server 2000 Retail Solution Site will be used in the lab The students will implement promotions and cross selling by using the tools supplied with Business Desk

Trang 5

Overview

Implementing State in a Shopping Cart

State vs Scalability

Membership and Authentication

Adapting the User’s View of the Site

Commerce Server 2000 Personalization Tools

*****************************I LLEGAL FOR N ON -T RAINER U SE *****************************

In this module, you will examine the issues involved in maintaining state in an e-business application The module describes how state is important to a Web shopping cart and then explains the techniques that can be used to preserve state It also discusses how state impacts scalability and how to optimize state and scalability in your application At the end of this module, you will look at how state maintenance enables personalization of a site and how to use Microsoft® Commerce Server 2000 to implement personalization in your e-business site

After completing this module, you will be able to:

Describe how to manage state for a Web shopping cart

List some of the issues related to state and scalability and explain how they can be addressed

Identify the role of MicrosoftApplication Center 2000 in software scaling

List the mechanisms that can be used for identifying the users of an e-business site

Explain how personalization and customization can be used to improve the usability of a site

Explain how to use the tools in Commerce Server 2000 to implement targeted campaigns, promotions, and cross selling in your e-business

In this module, you will

examine the issues involved

in maintaining state in an

e-business application

Trang 6

Implementing State in a Shopping Cart

HTTP: A Stateless Protocol

Storing State on the Tiers

Identifying Users with Cookies

Managing State with ASP.NET Sessions

Session and Shopping Cart

The shopping cart is a fundamental requirement of an e-commerce site The shopper needs to be able to add products to a virtual cart, review the cart, remove items from the cart, and ultimately proceed to the checkout where they pay for the goods

Although the operation of a shopping cart appears straightforward, maintaining state is a challenge that needs to be overcome before the shopping cart will work successfully Some of the issues that can arise as a result of bad state management may not become obvious until some time after the site has been running

Topic Objective

To introduce the topics

covered in this section

Lead-in

The shopping cart is a

fundamental requirement of

an e-commerce site

Trang 7

Web server (IIS)

1 TCP/IP connection

2 HTTP request

3 HTTP response

4 Connection closed

Maintaining state is an important functionality of any Web application, particularly in e-commerce applications

What Is State?

State is any information or data that affects the behavior of a Web application For example, if a user browses an e-commerce site and clicks a button to indicate a preference for monetary currency, the site can be designed to reflect that preference every time a price is displayed

Information that constitutes state may be applicable to the user’s current visit only However, if such information can be retained for subsequent visits, it will significantly enhance the user’s experience

State and HTTP

It is important to understand why maintaining state on an e-commerce site can

be a problem State-maintenance problems are caused by the use of Hypertext Transfer Protocol (HTTP), the protocol that delivers information over the Web HTTP is used to transmit requests between Web browsers (clients) and Web servers

How HTTP works

HTTP is a Transmission Control Protocol/Internet Protocol (TCP/IP) application layer protocol that is responsible for the communication that takes place between browser and server and for the delivery of content to the Web browser The data that HTTP delivers in response to a successful Web request

is often written in Hypertext Markup Language (HTML)

HTTP is responsible for establishing the connection between the client and server and establishing the syntax of the request and response cycle As soon as the request for a resource has been received and serviced at the Web server, HTTP breaks the connection This behavior is by design, and breaking the connection allows HTTP to scale and support many clients on a single server

Trang 8

Typical Web browsing activity involves a user requesting a page, reading it, and then clicking a link to another page Users are just as likely to request a page residing on a completely different server as they are to request another page that resides on the current server Therefore, all connections are broken after a request has been serviced

This lack of continuity between individual requests in HTTP means that there is

no awareness on the part of HTTP of an ongoing session The lack of permanent connection between the Web browser and the Web server leads to the problem of maintaining state between successive pages that are accessed by the same user HTTP is, in effect, a stateless protocol

Need for preservation of state

Implementing HTTP as a stateless protocol is ideal for many simple Web sites, thereby offering the advantage of scalability mentioned above However, more comprehensive Web applications will require state to be preserved The lack of state preservation means that each request for a Web page is treated as a unique transaction HTTP does not remember or consider earlier requests from the same client/user Without preservation of state, there is no sense of history, even a history that is only a few minutes old You need to use extra mechanisms

to offer a user the ability to put products in a Web shopping cart and see those same products again in subsequent page requests

HTTP version 1.1 does contain the concept of an ongoing session between client and server in which the connection is not broken after each HTTP request However, dedicating network resources to a client in this way can impact site scalability The issues related to sessions and scalability are discussed later in this module

Note

Trang 9

Storing State on the Tiers

State on the Client

Reduces server load

State on the Web server

Need not rely on the client’s ability to store state

State by Using the Shared Property Manager

COM+ feature for sharing state across multiple objects

State in a Database

Persistent storage

Sharing state across multiple Web servers is useful in load-balancing scenarios

A Web application can store state in various places The location you choose will depend on the type of application and the length of time for which state needs to be stored

Storing State on the Client

Storing state on the client reduces server load, but it requires the client to be capable of storing state Storing state on the client also may require the user to explicitly accept storage of state data, for example, by agreeing to accept a cookie or by agreeing to download and install custom software In practical terms, this method allows the user to disconnect from the site and yet recover the state of the site when reconnected Cookies are described in detail in the following section

Storing State on the Web Server

By storing state on the server, you do not have to rely on the client’s ability or willingness to store state data in order to maintain state A simple

implementation of state management might make use of the ASP.NET Session

object This object enables the server to identify the client and maintain state on the server

ASP.NET introduces several improvements to the implementation of session management These will be discussed later in this section

Storing State by Using the Shared Property Manager

COM+ provides the Shared Property Manager, which makes sharing state in a multi-user environment as easy as it is in a single-user environment

The Shared Property Manager provides a mechanism to share and access global information among server processes You can use the Shared Property Manager for a range of tasks, including workflow, performance monitoring and

management, and process serialization through global locks

Topic Objective

To describe the possible

locations for storing state

Lead-in

A Web application can store

state in various locations

Note

Trang 10

The Shared Property Manager eliminates the housekeeping tasks, which are needed to share information among applications The Shared Property Manager takes care of memory management, thereby freeing developers to concentrate

on program logic

To use the Shared Property Manager to store and share state between different objects and applications, the objects must be running in the same server process as the Shared Property Manager

Storing State in a Database

State data that is already stored on a Web server can be also stored on a database This enables the server to retrieve a client’s current state from the database Having a central storage point for state, such as a database, allows multiple Web servers to share state data, thereby removing the affinity between

a particular user and a single server in a multiple Web server scenario The ability to store state data in this manner in the back end (that is, in a database) is very important for implementing load-balancing solutions

However, there are also disadvantages to storing state in the back end The disadvantages are:

The database resources required in storing state information

The computing cost that is incurred by frequently moving data from the client to the database

Note

Trang 11

Identifying Users with Cookies

Cookie is a name-value pair that is stored on the client

Issues with using cookies include:

Maintaining privacy of user data

Maintaining users’ personalization configurations individually

Other methods for identifying users include:

Dynamically rewriting the URLs in pages

Adding hidden fields to forms

Click to purchase</a>

(View mod5ex1.asp and mod5ex2.asp from Sampapps)

(View mod5ex3.asp and mod5ex4.asp from Sampapps)

As discussed in the previous topic, cookies can be used to store state on the client side

Cookies can also be used as part of the overall solution when state is being stored on the server A prerequisite for maintaining state on the server is the ability to identify users Cookies provide a very simple mechanism for enabling user identification

Using Cookies to Identify Users

In Web terminology, a cookie is a name-value pair that is stored on the client The value of a cookie is simply a text string When a user visits a Web site, the server can send down one or more cookies as part of the HTTP response header The browser stores the cookies in a cache or file The cookies are then sent to the server when the user requests subsequent pages from the same domain Technically it is possible for a cookie to contain all of the information about an ongoing transaction, such as products in the shopping cart, the user’s name and address, and even credit card details Storing all information in cookies is a very insecure practice and is discouraged Most e-commerce sites will store a unique identifier in the cookie and keep all of the shopping cart details on the server

Issues with using cookies

You must be aware of the following issues involved with using cookies:

Maintaining privacy of user data Cookies are very easy to impersonate and are not always stored securely on the client’s computer—an issue of particular concern on computers shared

by several users For this reason, any information stored as a cookie must be restricted For example, although it may be acceptable to retrieve the contents of a shopping cart by sending a cookie to a Web application, it should not be possible to use a cookie to retrieve a shopper’s credit card details from an earlier purchase

Topic Objective

To describe how cookies

can be used in identifying

users

Lead-in

Cookies can be used to

store information on the

client or to identify the user

Trang 12

Maintaining users' personalization configurations individually

As discussed later in this module, cookies can also be used to provide personalization features, such as personal home pages on a site In certain environments, multiple users may share the same computer, browser, and cookie cache In this case, you do not want one user to inherit the personal preferences of another

Some users disable cookies in their browsers because they are concerned about these issues Therefore, you cannot rely on cookies as the basis for maintaining user or session state

Identifying Users through Other Techniques

There are some common alternatives to cookies

Dynamically rewriting the URLs in a page that is sent to the browser

By dynamically rewriting the Active Server Pages (ASP) on a site, you can append extra parameters at the end of each Uniform Resource Locator (URL) One of these parameters could be used to identify a shopper When a user requests a page, the site can retrieve the parameter that you have designated as the shopper identifier and can dynamically append it to every subsequent link

on the page This technique can be used to identify a user on every page accessed on the site

Let us look at how you can implement this concept First examine the following URL:

http://localhost/shoppingcart.asp?shopid=12432354 The shopid parameter that has been included in the preceding URL is used for identifying a particular shopper Before a link is built on the shoppingcart.asp page, the shopid parameter can be requested from the URL and included in the hyperlinks Links on the shoppingcart.asp page will look similar to the

following:

Adding hidden fields to the forms in which the user is entering data

Many sessions involve users providing information about themselves and their desired purchases in a form When forms are being used, hidden fields can contain session-tracking data Every time users submit a form, they also submit

a hidden field that is used by the server to identify them An example of a hidden field that would be included within the FORM tags is as follows:

<input type="hidden" name="shopid"

Sampapps folder These

files pass information in the

URL that identifies the

shopper

Delivery Tip

Open the files mod5ex3.asp

and mod5ex4.asp in the

\InetPub\WWWRoot\2260\

Sampapps folder These

files pass information in the

FORM tag that identifies a

hopper

Trang 13

Problems with using these techniques

Although these techniques are suitable for session tracking, unlike a cookie, the information will not persist over multiple visits On closing the browser, the session state data will be destroyed On a subsequent visit, a new session identifier needs to be created These techniques are not useful for a site that seeks to implement personalization because the information that is gathered to implement a personalized Web site is gathered over multiple visits To benefit from these techniques and to implement personalization, the users would be required to log on to the site and have their session details retrieved from a database before passing them from page to page in a URL or passing them through a hidden field on a form

Trang 14

Managing State with ASP.NET Sessions

Enhanced Session Object with Built-In State Functionality

Automatically manages requests from a single client into

a session on the server

Raises appropriate session lifetime management events

Automatically releases session information

ASP.NET Session Characteristics

Session State Storage

Sessions can trigger events in global.asax

Public sub Session_OnStart() Session("FavColor")= "Blue"

End Sub

</script>

End Sub

</script>

ASP.NET ships with an enhanced Session object containing built-in session

state functionality that provides you with an easy, robust, and scalable way to implement state management Sessions are created in the server-side code in an ASP.NET page By using session state, a shopping cart can easily be

maintained on the server

ASP.NET sessions also offer the following advantages:

Automatically identify and classify requests from a single client browser into a session on the server

If ASP.NET session support has been configured, each time a user accesses the application, ASP.NET identifies that the connection is new and

automatically configures a session that can be used to store shopping cart items

Raise appropriate session lifetime management events (Session_OnStart and Session_OnEnd) that can be handled in an application code

ASP.NET supports a global.asax file, which is similar to the global.asa file

in ASP This file can contain code that is run when a new session is started and when an existing session ends

Automatically release session information if the browser does not revisit an application after a specified time-out period

By default, the session time-out value is 20 minutes This time-out value can

be increased by editing the config.web file

enhanced Session object

that can be used to store

state on the server

Trang 15

ASP.NET Session Characteristics

ASP.NET offers a session state facility that is familiar to ASP developers, while offering the following improvements:

A reliable session that can be run in its own process space and is capable of surviving Internet Information Services (IIS) failures

ASP.NET session state is, by default, stored externally from the worker process in a dedicated Microsoft NET state server process End-user code cannot be run in the state server process, thereby reducing the chances of faulty code crashing the service In the event of a failure or system restart of IIS, state for a user is preserved

A scalable session state facility that can be used in both Web farm computer) and Web garden (multi-processor) scenarios

(multi-Because the ASP.NET session is stored externally from the worker process, the state data can be accessed by another worker process These worker processes could be running on another processor on the same computer or could potentially be running on an entirely different computer

A session state facility that works with browsers that do not support the use

of HTTP cookies for privacy or legal reasons

ASP.NET supports embedding a user identifier in a URL This allows the session Identifier (ID) to be automatically retrieved by IIS prior to invoking

the Session object An example of an URL with an embedded session ID is

as follows:

http://localhost/litware/SESSION$0123456789ABCDEF/page1.aspx

In the above URL, the session ID has been embedded between the forward slash following the application name (LitWare) and the forward slash immediately before the filename (page1.aspx)

Session State Storage

Session state can be stored external to the ASP.NET process Session state is

stored, by default, in an object called the Session State Store The process that

manages this object is called the state server process It is this reliable process that can persist session state even if the ASP.NET process fails

Alternatively, the contents of the ASP.NET Session object can be written to a

table in a Microsoft SQL Server™ database This feature provides another layer

of protection to the contents of the ASP.NET Session object—specifically

guarding against a complete computer failure, such as the case of a complete power outage on the Web server In this case, the contents of the ASP.NET

Session object can be reloaded after the computer has recovered from the

failure

Trang 16

Creating an ASP.NET Session

A new ASP.NET session is created each time a user without a session interacts with an aspx page in a Web application The session can trigger events in the global.asax file The following code creates a session-scoped variable with a particular value in the global.asax file:

Trang 17

Session and Shopping Cart

Preserving shopping cart contents and state information

Contents of the cart

A unique product identifier—SKU or ISBN number

The number of items required

The unit cost

A user-friendly description of the product

Save cart contents in a database to avoid losing information because of Session objects time-outs

After a session is created between the browser and server, you need to determine exactly what information you want to maintain and store as part of the user’s session In an e-commerce store, there is a clear relationship between state and a user's shopping cart Along with the products that users put in their carts, you may want to keep additional state information, such as the preferred monetary currency of a user

Contents of the Cart

The information contained in a shopping cart is specific to the application that uses the cart However, it generally consists of multiple items, each of which consists of one or more of the following:

A unique product identifier, for example, a Stock Keeping Unit (SKU) or an International Standard Book Number (ISBN)

The number of items required

The unit cost

A user-friendly description of the product You must decide what information must be stored in the cart and what information must be left in the database until it is needed For example, you must decide whether to hold the unit price of an item in the cart or retrieve this directly from the database when the user proceeds to the checkout

Also, a decision needs to be made on whether to write the cart information to a

database and, if so, how long it must be kept in the database Also, the Session

object will time out if it is not used in a specified period; by default this period

is set to 20 minutes Therefore, writing cart information to a database is a

common practice In fact, the ASP.NET Session object provides a built in

mechanism for storing session data in a SQL Server database

Topic Objective

To explain the relationship

between a session and a

Web site shopping cart

Trang 18

Unfortunately, the session ID is unique only among all the current sessions It is not a globally unique identifier This means that if the cart information is stored

in the database for a long period of time, it is possible that at a later time, another shopper will receive the same session ID and retrieve a partially filled cart in error

Trang 19

State vs Scalability

How State Impacts Scalability

Resource Recycling

Software Scaling

Using Application Center 2000

So far, you have seen how state can be maintained by using several techniques However, state maintenance can sometimes lead to scalability issues Let us examine how a balance can be achieved between state and scalability

Topic Objective

To introduce the topics in

this section

Lead-in

So far, you have seen how

state can be maintained by

using several techniques

However, state maintenance

can sometimes lead to

scalability issues Let’s

examine how a balance can

be achieved between state

and scalability

Trang 20

How State Impacts Scalability

Limitations of Increasing Resources

Session and state management use valuable side resources

server- These resources are idle most of the time

Holding Server-Side Resources Results in Scalability Problems

The Stateless Paradigm

1. Acquire the required resources

2. Use the resources to complete the task

3. Release the resource as quickly as possible

E-commerce applications must be able to scale appropriately when faced with increased user demand Holding state (or maintaining a session) for a user of a Web application consumes valuable server-side resources This approach can cause scalability problems because additional users can quickly outnumber limited resources

Limitations of Increasing Resources

The server-side component of a user's session—for example, an based Web application—must use system resources to accomplish tasks for the user The resources acquired by a user's session form part of that user's state Although it is true that you can increase the number of users that the site can simultaneously process by adding more resources, this is not always practical or beneficial Most of the resources dedicated to a user remain idle for most of the time Typically, both the time taken by the user to decide on their next action and the communication time between client and server far outweigh the time spent using the server resources To service more clients without increasing resources, your existing resources must be shared very efficiently

ASP.NET-The Stateless Paradigm

As we have seen, because HTTP is a stateless protocol, it does not maintain an open connection between page accesses, which makes it ideal for scalability Each Web site has a maximum number of network resources For example, there are a limited number of TCP/IP sockets and they can be used by only one client at a time By using these resources for only as long as they are needed, HTTP makes these resources immediately available to other users However, you must use cookies or dynamic URL rewriting to create virtual sessions

Topic Objective

To explain the problems of

maintaining state in a

solution that needs to be

scaled and how to address

these problems at a higher

Trang 21

While the mechanisms used to create virtual sessions solve the problem at the network level, there are other resources that may be constrained on the server, including:

Memory

Database connections

Onward connections to other servers Such resources are acquired by the user's session on the server as the user performs certain tasks If the server leaves the user's session open for a relatively long time, all of the resources it continues to consume are denied to other sessions and users Taking into account the stateless paradigm, every individual task performed on behalf of a user must:

1 Acquire the resources it needs

2 Use the resources to complete the task

3 Release the resources as quickly as possible

The above sequence is similar to how HTTP uses the underlying socket when transporting a Web page Recycling of resources is key to creating scalable solutions

Trang 22

Resource Recycling

In an infinite resources scenario, a resource can be held

on to for as long as it is needed

In a single resource scenario, the application must follow the acquire-late-and-release-early principle

In a real-world scenario, apply the release-early principle wisely

acquire-late-and- If reconstructing state is very time consuming, this can lead to poor performance of the application

Resource recycling is one of the most important strategies that you can adopt to improve site scalability For example, consider a typical application that uses connections to a database Database connections are typical resources that are limited in a Web application In this example, assume that each client session requires a single database connection

To understand how resource recycling can improve scalability, we will compare the following two scenarios:

Infinite resources are available to your application

A single resource is available to your application

Infinite Resources Scenario

Let us examine a case in which you have an unlimited number of database connections available to your application Every client that accesses the application is assigned its own database connection, and no client is ever denied access because of a lack of resources The client can hold on to that resource for

as long as it wants to because any new clients can be assigned another connection from the infinite resource pool

A Single Resource Scenario

Next, let us examine a case in which your site has only a single database connection that must be shared between all clients In such a situation, there is a real danger that clients will be denied access if that resource is being used by someone else Because of this, it is vital that each client uses the resource for as short a time as possible Reducing this use time reduces the possibility that the resource will be in use when another client requests it

Trang 23

Acquire late and release early

To understand how a site can be designed so that a resource is used for the shortest time possible and as little as necessary, let us examine the exact points

in time that a resource is allocated and released Many developers write code such that all the needed resources are acquired at the start of a program or method These resources are retained until the program or method ends

However, if there are only two short sections of code that use the resource in the overall program or method, it makes sense to acquire the resource at the latest possible point in time (that is, immediately before possessing the resource becomes absolutely necessary) Similarly, after you no longer have an

immediate need for a resource, release it at once so that it can be ready for use

by another client

Scalability vs efficiency

The acquire-late-release-early strategy may appear to some developers to be inefficient However, the key issue here is not efficiency but scalability, and scalability relies on rapid resource recycling Also, remember that the use of connection pools will reduce the overhead of acquiring a connection, so this strategy is not as inefficient as it may at first appear It is precisely the acquire-late-release-early strategy that COM+ applies to achieve component scalability through Just-In-Time (JIT) activation and As-Soon-As-Possible (ASAP) destruction

The Real-World Scenario

In the real-world scenario, of course, you will have neither a single resource nor infinite resources However, applying the lessons of how to maximize usage of

a single resource will help you achieve optimum scalability Also, remember that in the real-world scenario, scalability may be just one of many

nonfunctional requirements of your application

If it takes a long time to prepare the state data that you need to acquire for performing a particular task, the performance drawbacks of continually discarding and reacquiring state may be prohibitive The acquire-late-release-early principle must not be applied to such cases

Tip

Trang 24

Lack of cluster monitoring tools

A scalable Web application requires developers to have a good understanding

of how their design will impact the site’s ability to scale Resource recycling is

a good example of a best practice that a developer must be aware of

Application Center 2000, which is part of the Microsoft NET Enterprise Server family, provides a powerful tool for scaling Web applications Many Web applications implement software scaling—adding additional servers to the overall solution and building a Web farm Software scaling introduces several benefits However, it also introduces a number of issues that Application Center 2000 will help you to resolve

Benefits

The two key benefits of software scaling are:

Increased capacity Software scaling allows companies to increase capacity by:

• Replicating the existing application onto a new server

• Adding the server to an existing cluster of Web servers

• Updating the load-balancing solution to make use of this new server

High availability Because each server is running identical applications, individual servers can fail without interrupting the users of the application An intelligent load-balancing solution can redirect requests to the remaining servers in the cluster

Topic Objective

To introduce the benefits

and drawbacks in scaling

applications

Lead-in

Adding servers to a Web

application is known as

software scaling Although

very popular, there are

some drawbacks to this

approach

Trang 25

Drawbacks

Some of the drawbacks in software scaling are:

Synchronizing applications is labor intensive

Manually moving applications from one server to another can prove labor intensive This is compounded by subsequent updates requiring distribution

to all servers that are participating in the cluster

Predicting required resources for peak loads is difficult

Software scaling answers the question of how to scale However, an important issue that remains is the degree of scalability that is required Seasonal peaks are common in e-business stores The store must be able to cope with the peaks through the implementation of additional resources However, the cost of leaving those resources underused when the peak recedes could be significant

Managing and monitoring clusters is difficult

The more computers that are added to the cluster, the more difficult it becomes to monitor and manage them Ideally, there should be a single interface to manage all of the computers in the cluster

Trang 26

Using Application Center 2000

Features

Simplified cluster operations

- Network load balancing

- Component load balancing

- Request forwarding

Simplified application management

On-demand scalability

Application Center 2000 offers solutions for many of the issues that can be encountered with software scaling Application Center offers a solution that includes Internet Protocol (IP) and component load balancing, simplifies manageability, and offers on-demand scalability Application Center 2000 can

be used throughout the Web application infrastructure in the presentation, business logic, and data tiers

Features

Some of the key features that can be used in an e-business environment include:

Simplified cluster operations Application Center includes a New Cluster Wizard that simplifies the configuration of load-balancing options The load-balancing options include:

• Network load balancing Network load balancing (NLB) is a clustering technology that is offered

as part of Microsoft Windows® 2000 Advanced Server and Datacenter Server NLB allows for load balancing of incoming TCP/IP requests among the servers in the cluster Application Center does not alter NLB, but it does offer a centralized management interface and simplified wizard-based configuration

• Component load balancing Component Load Balancing (CLB) offers the distribution of COM+ requests in a way similar to how NLB offers the distribution of TCP/IP requests CLB uses a list of potential servers in conjunction with a response-time/round-robin hybrid distribution scheme to determine where to route the request This allows the COM+ layer to be scaled independently of the application’s other tiers

Topic Objective

To describe how Application

Center 2000 resolves the

issues in software scaling

Lead-in

As part of the NET server

family, Application

Center 2000 offers solutions

to the scalability and

manageability issues that

have been earlier

discussed

Trang 27

• Request forwarding Request forwarding (RF) tackles the challenge of integrating NLB and session state in a Web application If a session is established and subsequent requests are redirected to a different server by NLB, RF can intercept those requests and redirect to the original server before the request is handled by the Web application

Simplified application management Application Center offers three configuration options for all of its services: Microsoft Management Console (MMC), Web browser, and command line Application Center 2000 has the ability to group the servers in a cluster and manage them as if they were a single resource Management also allows for the deployment and synchronization of content and allows for the

distribution of COM+ components and IIS application programming interface (API) filters

On-demand scalability

A feature that will be very useful to e-business developers is the ability to scale on demand When monitoring indicates that a seasonal increase in traffic has started, the server wizard can be used to introduce additional servers to the cluster When the traffic reduces, these servers can return to their original roles

Trang 28

Membership and Authentication

Need for Membership

Authenticating Users

Using Microsoft Passport

We have already examined several techniques for identifying a user These

include cookies, the Session object in ASP and ASP.NET, and posting hidden

fields in forms These techniques introduce state and allow a Web site to track users and maintain an individual shopping cart on the server for them

This section focuses on the concept of membership and authentication These techniques go beyond the concept of constructing state for a transaction and move closer to building relationships between the site and the intended users

Topic Objective

To introduce the topics

covered in this section

Lead-in

Sessions allow us to track a

user and build information

They even allow us to store

that information persistently

As site requirements

increase, we move toward

the concept of membership

and authentication

Trang 29

Need for Membership

Access to information or services not generally available

Frequently required information can be stored by the site to avoid the requirement for re-keying

Personal preferences can be gathered either explicitly

or implicitly

The contents of shopping carts can be maintained over

a longer period of time

Special offers and prices can be offered to a particular member, groups of members, or all members

Many sites gather information from users to help provide those same users with better service, such as personalized home pages on the site This requires that a site be able to uniquely identify a user, usually by relying on a logon process After the user is identified, the personalized information previously gathered can be used to enhance the user's experience of the site This association of user identity and persistent state is referred to as membership

Benefits of Membership

By becoming a member of a site, a user can gain various benefits:

Access to information or services not normally available

Premium content can often be used to encourage a user to sign up for membership On an e-commerce site, premium content might involve product reviews from peers or site experts

Frequently required information can be stored by the site to avoid the inconvenience to the user of re-keying this information on each visit

Examples of such information are the billing address, shipping address, and preferred monetary currency for transactions

Personal preferences can be gathered either explicitly or implicitly

For example, a news or portal site may provide a dynamic home page for each user, based on previously gathered preferences

The contents of shopping carts can be maintained over a longer period

An example of this can be a standard shopping cart that the user modifies before each purchase In the case of an online grocery store, the user could use the membership facility to store their shopping list online Prior to their weekly purchase, they can add and remove items, but a large percentage of the goods that they require are already selected

Let’s look at how

membership can be used to

improve a user’s shopping

experience

Trang 30

Special offers and prices can be offered to a particular member, groups of members, or all members

As part of a loyalty scheme or marketing campaign, membership-based systems often allow for much more accurate targeting of offers Also, in business-to-business (B2B) solutions, different companies may receive differing discount levels depending on the amount of business they do

Tiêu đề	Managing the Business Logic Layer
Trường học	Microsoft Corporation
Chuyên ngành	Information Technology / E-Business
Thể loại	Giáo trình hướng dẫn
Năm xuất bản	2001

Định dạng
Số trang	60
Dung lượng	0,93 MB