This Node is a personal computing device, such as a PC, supporting a commercial browser (e.g. Netscape Navigator or Internet Explorer). The level of the browser is expected to support SSL and some level of DHTML. Most online buying implementations will send a "Cookie" to the browser on this node in order to maintain the shopping session. The cookie will contain a session id, which can be used to reconnect with a partially filled shopping basket or order and to re-establish the conversation for each interaction.

Firewalls provide services that can be used to control access from a less trusted network to a more trusted network. Traditional implementations of firewall services include:

• Screening routers (the protocol firewall in this design)
• Application gateways (the domain firewall)

The two firewall nodes provide increasing levels of protection at the expense of increasing computing resource requirements. The protocol firewall is typically implemented as an IP router, while the domain firewall is a dedicated server node.

Additional Resources

• (in English) ESS

The load balancer, or dispatcher, node provides horizontal scalability by dispatching HTTP requests among several, identically configured Web servers.

Additional Resources

• (in English) ESS

This node provides the infrastructure for application logic and can be part of a commerce server node. It is capable of running both presentation and business logic.

Firewalls provide services that can be used to control access from a less trusted network to a more trusted network. Traditional implementations of firewall services include:

• Screening routers (the protocol firewall in this design)
• Application gateways (the domain firewall)

The two firewall nodes provide increasing levels of protection at the expense of increasing computing resource requirements. The protocol firewall is typically implemented as an IP router, while the domain firewall is a dedicated server node.

Additional Resources

• (in English) ESS

The purpose of this node is to interface between any front end access channel, such as the web, a call center, or a client/server ("fat client") PC, and whatever back-end application system is needed (including applications from other companies). It will perform the following kinds of services:

• Convert protocols from the front end to match whatever the back-end systems understand
• Decompose a single message from the front end (such as a web server) into several back-end messages (or transactions), and then re-compose the replies
• Navigate from the front end to whatever back-end system needs to be accessed
• In more complex cases, control the process or unit of work for a number of back-end interactions based on a request from the front end

The intent is to relieve each front end from having to handle the complexity of interfacing with potentially multiple back-end systems, which may be in different companies. The front end (such as the web server should just need to send a message to the integration server and have it look after the interface.

A second purpose for locating these interface services on the Integration server concerns security. There is a firewall between the web server and the integration server and the web server need have no knowledge of all the back-end addresses. Many location do not want a server located in the DMZ to have access directly to sensitive data and systems. In this case the web server can only send messages to the integration server, nowhere else.

This node is a logical representation of the functions required to manage all the nodes and components in the system, including the management of problems, changes, performance, configuration of assets, and others.

There are usually two aspects to systems management:

• A managing aspect (with server components running on one or more systems management servers)
• A managed aspect (with systems management client components running on every node in the system)

For example, there should be client components running on every node, which are able to accept and install changes sent from a change management server. The changes may be either pulled from the client on demand, or pushed from the server with centralized control. On critical nodes, there may also be problem management components that send a "heartbeat" back to a central monitoring site. If a heartbeat is missed, the managing site raises an alert.

It is important to remember that there is not only a server side but also a client side to systems management. The client side needs to reside on all nodes in the system. The one exception would be unmanaged workstations of the general public or of other business partners. In these cases, a particular organization would have no control and limited ability to provide systems management.

This node is a logical representation of the functions needed to manage the security of a system. It works in conjunction with the Directory Node. Think of the directory as the repository that holds:

• Data about security such as user IDs and associated passwords, or digital certificates (used to authenticate a user)
• Lists of services that a user is authorized to perform (authorization or access control)

Think of the security node as holding the set of components that define the decisions to be made. The node may perform the actual security processing; for example, verify certificates. The authentication in most current designs validates the access to the Web Application Server, but it can also authenticate the access to the Database Server.

The security domain describes the components needed to implement the technical dimension of security and how these components interact to implement the technical aspects of a security policy.

The components that implement security are distributed throughout the network. It's unlikely there's a node in the system that does not include some components implementing some aspect of security. The Security Node represents centralized services that support security on other nodes.

The treatment of security combines network design for security with a particular emphasis on achieving a secure implementation of Internet and Intranet network access. Security is built by use of the following security services:

• Confidentiality: providing privacy by protecting sensitive information from access unless authorized
• Identification and Authentication: identifying entities, verifying their identities, and assuring individual accountability
• Access Control: providing mechanisms for granting access to authorized and authenticated users
• Data Integrity: providing detection of the unauthorized modification of data
• Non Repudiation: assuring that any transaction that takes place can subsequently be proved to have taken place (also called accountability)
• Isolation: providing protection by isolating a resource and therefore restricting potential access to it
• Audit: monitoring and review of security-relevant events

Together these services must provide end-to-end security, integrating security facilities across heterogeneous environments.

Directory Services provides an integrated LDAP server that easily integrates with other LDAP servers to provide for the directory needs of an e-business solution.

Existing applications are run and maintained on nodes that are installed in the internal network. These applications provide for business logic that uses data maintained in the internal network. The number and topology of these existing application and data nodes is dependent on the particular configuration used by these legacy systems.

The database server node's function is to provide persistent data storage and retrieval in support of the user-to-business transactional interaction. The data stored is relevant to the specific business interaction, for example, bank balance, insurance information, current purchase by the user, and so forth.

Note that the mode of database access is perhaps the most important factor determining the performance of this Web application, in all but the simplest cases. The recommended approach is to collapse the database accesses into a single call or very few calls. This can be achieved through coding and invoking stored procedure calls on the database.

This node represents the functionality supporting the creation of the data that resides on the Database Server and Commerce Server Nodes. It also represents the function to manage and stage that data into production on the servers. The functionality of this node is quite broad, and might be thought of as encompassing an entire subsystem.

The timely synchronization of several Web Servers is sometimes achieved by using a Shared File System as the content storage, capitalizing on the replication capability of this technology.

Lightweight Directory Access Protocol (LDAP) refers to the protocol that is used to communicate from a calling program (running on a node such as a Commerce Server) and a Directory Node. Information is kept on the LDAP-based directory node about such topics as people and/or services.

For example, the directory could be used to store information needed to identify registered shoppers (referred to as authentication). It could also be used to store information about what functions those shoppers are allowed to perform after being identified (referred to as authorization).

Secure Sockets Layer (SSL) refers to encryption technology which is commonly used between the browser on a users PC and a Web Server Node. It is used to protect the data in messages from being viewed in an un-authorized fashion while travelling over a TCP/IP network. It can also continue to protect the messages as they flow over internal TCP/IP networks between nodes after passing the web server.

A digital certificate is an electronic credential issued by a trustworthy organization such as a bank, credit union, or large company. The digital certificate vouches for an individual's, business's, or organization's identity and authority to conduct any transaction over the Internet. The issuing organization is called a Certificate Authority, or CA. VeriSign is an example of an existing CA.

Digital certificates address the issue of ensuring that the owner of a public key is really who he claims to be .This technology provides a mechanism to distribute public keys in a special format called a certificate. In addition to the key itself, the certificate contains information about the sender. The whole package is then signed by the issuing CA using its private key. The receiver of the certificate can then verify the CA's digital signature using the CA's public key. Digital Certificates are made available to the public through on-line directories based on X.500 standards.

Digital certificates are defined by the X.509 standard. A X.509 certificate is typically a small file that contains:

• Subject's distinguished name
• Issuer's distinguished name
• Subject's public key
• Issuer's digital signature
• Validity period
• Certificate's serial number

IBM Global Services refers to that organization within IBM that provides professional IT Services and Consulting.

IBM Universal Server Farm refers to a set of outsourcing sites run by IBM. Server technology from multiple different vendors is operated by IBM on behalf of organizations which would rather not manage their own server operations. Frequently in Online Buying sites, the nodes in the DMZ front end and the Data Base Server Node are outsourced. These are then networked to the company's own datacenter for backend order processing.

Enterprise Solutions Structure (ESS) is a major IBM initiative to establish a standard architectural framework to support creation, reuse, and maintenance of architecture and design assets. These intellectual capital assets are used by IBM services practitioners for developing and delivering enterprise solutions. ESS draws on experiences with building customer solutions to distill "best practice" structures, models, and sample deliverables. The framework provides a rich set of architectural building blocks for solution architects and provides guidance on when and how to use this content to advantage. Specifically, this architecture provides a common, consistent approach for understanding and documenting business requirements via a business model, designing a logical architecture of key components and services, and finally, implementing a physical architecture based on actual products, platforms, and services.

The term "Reference Architecture" is used to refer to the collection of assets which as a whole describe how to implement a given type of business solution. For example, there is a reference architecture which shows how to implement a call center. There is another one which shows how to implement an online buying application. This site which provides Patterns for e-business is based to a large extent upon the ESS reference architecture assets. The intent is to share a summary of those reference architectures with you in this way.

See Also

• Thin Client Transactional Pattern

Additional Resources

• (in English) "Enterprise Solutions Structure," IBM Systems Journal 38, No. 1 (1999)
• (in English) "Technical Reference Architectures," IBM Systems Journal 38, No. 1 (1999)