Skip Headers
Oracle® Database Heterogeneous Connectivity Administrator's Guide
10g Release 2 (10.2)

Part Number B14232-01
Go to Documentation Home
Home
Go to Book List
Book List
Go to Table of Contents
Contents
Go to Index
Index
Go to Master Index
Master Index
Go to Feedback page
Feedback

Go to previous page
Previous
Go to next page
Next
View PDF

2 The Role of the Heterogeneous Services Component

Oracle's synchronous solutions for operating in a heterogeneous environment are Oracle Transparent Gateways and Generic Connectivity. The common component of the Oracle database server for supporting these solutions is Heterogeneous Services. This chapter describes the architecture and functionality of the Heterogeneous Services component and its interaction with Oracle Transparent Gateways and Generic Connectivity.

This chapter contains the following topics:

2.1 Heterogeneous Connectivity Process Architecture

At a high level, Oracle heterogeneous connectivity process architecture is structured as shown in Figure 2-1.

Figure 2-1 Oracle Heterogeneous Connectivity Process Architecture

Description of heter006.gif follows
Description of the illustration heter006.gif

The Heterogeneous Services component in the Oracle database server talks to a Heterogeneous Services agent process which, in turn, talks to the non-Oracle system. We can conceptually divide the code into three parts:

2.2 Heterogeneous Services Agents

A Heterogeneous Services agent is the process through which an Oracle server connects to a non-Oracle system. This agent process that accesses a non-Oracle system is called a gateway. Access to all gateways goes through the Heterogeneous Services component in the Oracle server and all gateways contain the same agent-generic code. Each gateway has a different driver linked in that maps the Heterogeneous Services to the client API of the non-Oracle system.

The agent process consists of two components. These are agent generic code and a non-Oracle system-specific driver. An agent exists primarily to isolate the Oracle database server from third-party code. In order for a process to access the non-Oracle system, the non-Oracle system client libraries have to be linked into it. In the absence of the agent process, these libraries would have to be directly linked into the Oracle database and problems in this code could cause the Oracle server to go down. Having an agent process isolates the Oracle server from any problems in third-party code so that even if a fatal error takes place, only the agent process will end.

An agent can reside in the following places:

Agent processes are started when a user session makes its first non-Oracle system access through a database link. These connections are made using Oracle's remote data access software, Oracle Net Services, which enables both client/server and server-server communication. The agent process continues to run until the user session is disconnected or the database link is explicitly closed.

Multithreaded agents behave differently. They have to be explicitly started and shut down by a database administrator instead of automatically being spawned by Oracle Net Services.

2.3 Types of Heterogeneous Services Agents

There are two types of Heterogeneous Services agents:

2.3.1 Oracle Transparent Gateways

An Oracle Transparent Gateway is a gateway that is designed for accessing a specific non-Oracle system. Oracle provides gateways to access several commercially produced non-Oracle systems; many of these gateways have been ported to several platforms. For example, an Oracle Transparent Gateway for Sybase on Solaris is the Solaris port of a gateway designed to access Sybase database systems.

With Oracle Transparent Gateways, you can use an Oracle database server to access data anywhere in a distributed database system without being required to know either the location of the data or how it is stored. When the results of your queries are returned to you by the Oracle database server, they are presented to you as if the data stores from which they were taken all resided within a remote instance of an Oracle distributed database.

2.3.2 Generic Connectivity Agents

In addition to Oracle Transparent Gateways for various non-Oracle database systems, there is a set of agents that comprise the Oracle Generic Connectivity feature. These agents contain only generic code and the customer is responsible for providing the necessary drivers. Oracle has Generic Connectivity agents for ODBC and OLE DB that enable you to use ODBC and OLE DB drivers to access non-Oracle systems that have an ODBC or an OLE DB interface.

To access a specific non-Oracle system using Generic Connectivity agents, you must configure an ODBC or OLE DB driver to the non-Oracle system. These drivers are not provided by Oracle. However, as long as Oracle supports the ODBC and OLE DB protocols, you can use the Generic Connectivity feature to access any non-Oracle system that can be accessed using an ODBC or OLE DB driver.

Generic Connectivity agents have some limitations. Connections to a Generic Connectivity agent are supported only for the local Oracle database server. Functionality of these agents, especially when compared to the Oracle Transparent Gateways, is limited.

2.4 Heterogeneous Services Components

This section discusses the components of Heterogeneous Services in the Oracle database server. These components are:

2.4.1 Transaction Service

The transaction service component of the Heterogeneous Services component makes it possible for non-Oracle systems to be integrated into Oracle database server transactions and sessions. When you access a non-Oracle system for the first time over a database link within your Oracle user session, you transparently set up an authenticated session in the non-Oracle system. At the end of your Oracle user session, the authenticated session in the non-Oracle database system is also closed.

Additionally, one or more non-Oracle systems can participate in an Oracle distributed transaction. When an application commits a transaction, Oracle's two-phase commit protocol accesses the non-Oracle database system to transparently coordinate the distributed transaction. Even in those cases where the non-Oracle system does not support all aspects of Oracle two-phase commit protocol, Oracle can (with some limitations) support distributed transactions with the non-Oracle system.

2.4.2 SQL Service

The Structured Query Language (SQL) service handles the processing of all SQL-related operations. The work done by the SQL service includes:

  • Mapping Oracle internal SQL-related calls to the Heterogeneous Services driver application programing interface (API); this is in turn mapped by the driver to the client API of the non-Oracle system.

  • Translating SQL statements from Oracle's SQL dialect to the SQL dialect of the non-Oracle system.

  • Translating queries that reference Oracle data dictionary tables to queries that extract the necessary information from the non-Oracle system data dictionary.

  • Converting data from non-Oracle system data types to Oracle data types and back.

  • Making up for missing functionality at the non-Oracle system by issuing multiple queries to get the necessary data and doing post-processing to get the desired results.

2.5 Heterogeneous Services Configuration Information

Heterogeneous Services components consist of generic code and must be configured to work with many different non-Oracle systems. Each gateway has configuration information stored in the driver module. The information is uploaded to the server immediately after the connection to the gateway has been established. The configuration information includes:

2.5.1 Data Dictionary Translation Views

Data dictionary translations are views on non-Oracle data dictionary tables. They help Heterogeneous Services translate references to Oracle data dictionary tables into queries that can retrieve the equivalent information from the non-Oracle data dictionary.

2.5.2 Heterogeneous Services Initialization Parameters

Heterogeneous Services initialization parameters serve two functions:

  • They give you a means of fine-tuning the gateway to optimize performance and memory utilization for the gateway and the Heterogeneous Services component.

  • They enable you to tell the gateway (and, thereby, Heterogeneous Services) how the non-Oracle system has been configured (for example, in what language the non-Oracle system is running). They give information to Heterogeneous Services about the configurable properties of the non-Oracle system.

You can examine the Heterogeneous Services initialization parameters for a session by querying the view V$HS_PARAMETER. Users can set initialization parameters in gateway initialization files.

2.5.3 Capabilities

Capabilities tell Heterogeneous Services about the limitations of the non-Oracle system (such as what types of SQL statements are supported) and how to map Oracle data types and SQL expressions to their non-Oracle system equivalents. In other words, they tell Heterogeneous Services about the non-configurable properties of the non-Oracle system. Capabilities cannot be changed by the user.

2.6 The Heterogeneous Services Data Dictionary

Configuration information is uploaded from an agent to the Heterogeneous Services component immediately after the connection to the agent has been established. The configuration information is stored in Heterogeneous Services data dictionary tables. No further upload takes place until something at the agent changes (for example, if a patch is applied or if the agent is upgraded to a new version).

2.6.1 Classes and Instances

Using Heterogeneous Services, you can access several non-Oracle systems from a single Oracle database. This is illustrated in Figure 2-2, which shows two non-Oracle systems being accessed.

Figure 2-2 Accessing Multiple Non-Oracle Instances

Description of heter002.gif follows
Description of the illustration heter002.gif

Both agents upload configuration information, which is stored as part of the Heterogeneous Services data dictionary information on the Oracle database server.

Although it is possible to store data dictionary information at one level of granularity by having completely separate definitions in the Heterogeneous Services data dictionary for each individual instance, this can lead to an unnecessarily large amount of redundant data dictionary information. To avoid this, Oracle organizes the Heterogeneous Services data dictionary by two levels of granularity, called class and instance.

A class pertains to a specific type of non-Oracle system. For example, you may want to access the class of Sybase database systems with your Oracle database server. An instance defines specializations within a class. For example, you may want to access several separate instances within a Sybase database system. Each class definition (one level of granularity) is shared by all the particular instances (a second level of granularity) under that class. Further, instance information takes precedence over class information, and class information takes precedence over server-supplied defaults.

For example, suppose that the Oracle database server accesses three instances of Sybase and two instances of Ingres II. Sybase and Ingres II each have their own code, requiring separate class definitions for the Oracle database server to access them. The Heterogeneous Services data dictionary therefore would contain two class definitions, one for Sybase and one for Ingres II, with five instance definitions, one for each instance being accessed by the Oracle database server.

Note that instance level capability and data dictionary information are session-specific and hence are not stored in the Heterogeneous Services data dictionary of the Oracle database server. However, instance level initialization parameters can be stored in the database.

2.6.2 Data Dictionary Views

The Heterogeneous Services data dictionary views contain the following kinds of information:

  • Names of instances and classes uploaded into the Oracle data dictionary

  • Capabilities, including SQL translations, defined for each class or instance

  • Data Dictionary translations defined for each class or instance

  • Initialization parameters defined for each class or instance

You can access information from the Oracle data dictionary by using fixed views. The views are categorized into three main types:

2.7 Gateway Process Flow

Figure 2-3 shows a typical gateway process flow. The steps explain the sequence of events that occur when a client application queries the non-Oracle database system database through the gateway.

Figure 2-3 Gateway Process Flow

Description of heter010.gif follows
Description of the illustration heter010.gif

  1. The client application sends a query over Oracle Net to the Oracle database server.

  2. Heterogeneous Services and the gateway converts the Oracle SQL statement into a SQL statement understood by the non-Oracle database system.

  3. The Oracle database server sends the query over to the gateway using Oracle Net.

  4. For the first transaction in a session, the gateway logs into non-Oracle database system using a username and password that is valid in the non-Oracle system.

  5. The gateway retrieves data using non-Oracle database system SQL statements.

  6. The gateway converts retrieved data into a format compatible with the Oracle database server.

  7. The gateway returns query results to the Oracle database server, again using Oracle Net Services.

  8. The Oracle database server passes the query results to the client application using Oracle Net. The database link remains open until the gateway session is finished or the database link is explicitly closed.