精华区文章阅读

发信人: superman (☆风雨无阻☆), 信区: Linux
标  题: INFO of Informix-se for Linux
发信站: 紫丁香 (Thu Sep 10 20:41:10 1998), 转信

INFORMIX-SE

Version 7.2 for UNIX� and Linux

These are the sections outlined in this Technical Brief:

旾NFORMIX-SE Technical Overview
旾NFORMIX-SE 7.2 UNIX Specifications

------------------------------------------------------------------------

INFORMIX-SE is an SQL-based database engine for small- to medium-range
applications. It is the ideal solution for businesses that want the
power of SQL without complex database administration requirements.
With over a decade of proven reliability,
INFORMIX-SE has a strong track record among UNIX database servers.

Organizations don't always have the MIS resources or extensive database
expertise required to administer complex database systems. However, they
still need powerful information management systems梐nd for them INFORMIX-SE
is the low-maintenance,
high-reliability database solution.

------------------------------------------------------------------------

INFORMIX-SE Technical Overview

INFORMIX-SE provides excellent performance, data consistency, client/server
capabilities, and standards adherence with only minimal need for database
administrator (DBA) expertise, making it the ideal multiuser database system
for small corporations or
self-contained departments.

Performance

Database performance is essential for maintaining maximum throughput.
INFORMIX-SE maintains excellent performance levels through cost-based
optimization and flexible indexing options.

Cost-Based Optimizer with Data Distributions

INFORMIX-SE's cost-based optimizer will automatically determine the fastest
way to retrieve data from a database table based on detailed information
about the distribution of that data within the table's columns. The optimizer
can collect and calculate
statistics about specific data distributions during non-peak processing hours
to avoid any potential bottlenecks. It can then pick a return path that has
the least impact on system resources.

Flexible Indexing Options

INFORMIX-SE provides a variety of indexing options to achieve the best
possible performance. Indexes can include from one to eight fields from
the record, and each field can be a different datatype.

INFORMIX-SE supports different B+ tree indexing options that include unique
and clustered indexes. In a unique index, every record must have a different
key value梡reventing duplicate entries. A clustered index organizes the table
in the physical order
of the index, improving performance.

Data Consistency

INFORMIX-SE maintains data consistency via audit trails and transaction
logging, and by establishing and enforcing locking procedures, isolation
levels, and business rules.

Audit Trails and Transaction Logs

In the event of a systemfailure, audit trail facilities can be used to
restore a table and all of the completed transactions from a backup copy
of the table and the audit trail. Similarly, the entire database can be
restored from a backup of the
database and the transaction log file.

Figure 1: Screen Image, INFORMIX-SE Architecture
(31k, GIF)
Figure 1: INFORMIX-SE uses the native OS file system for data storage and
retrieval.

When an operation is unable to complete itself, the partially completed
transaction must be removed from the database to maintain data consistency.
To remove any partially completed transaction, INFORMIX-SE maintains a
historical record of all
transactions in the transaction log and automatically uses these transaction
records as a reference to restore the database to the state prior to the
transaction.

Locking and Process Isolation

Other important features for maintaining data consistency are locking
procedures and process isolation. These security measures prevent other
users from changing data that is currently being read or modified.

Locks

The database server prevents errors by imposing a system of locks. A lock
is a claim or reservation that a program can place on a piece of data. The
database server guarantees that, as long as the data is locked, no other
database server process can
modify it. When another program requests the data being modified, the
database server either makes the program wait or turns it back with an
error.

The throughput of transactions for a specific table can be affected by
the locking strategy used for the table. Applications that use strategies
of exclusive access to data might find that other database server processes
are spending time waiting for
access to the data. For this reason INFORMIX-SE provides several locking
levels. The database server can place a lock on a single row, table, or
database. Row-level locking is the default applied when the table is created.
Table- and database-level
locking are specified in the user's application.

Isolation Levels

The isolation level is the degree to which your read operation is isolated
from concurrent actions of other database server processes梬hat modifications
other processes can make to the records you are reading, and what records
you can read while other
processes are reading or modifying them.

INFORMIX-SE supports different levels of readonly isolation control. For
example, a SQL SELECT request does not lock any records in the database,
allowing full access to the data. This is true except when executing a
SELECT FOR UPDATE request. While
this type of request allows you to continue to access records, it does not
allow other SELECT FOR UPDATE transactions to lock records already involved
in update transactions.

Business Rules

Business rules enforce data consistency at the column level. They specify
permissible data values, column defaults, and column-to-column relationships.

Business rules must be enforced by all applications that access or manipulate
data. However, INFORMIX-SE does not depend on the application to implement
these business rules. Instead, INFORMIX-SE enforces these rules independent
of the user
application. This centralization of responsibilities removes the burden from
the user applications and guarantees adherence to business rules.

INFORMIX-SE supports integrity constraints, stored procedures, and triggers
to enforce business rules.

Integrity Constraints

Informix's implementation of ANSI SQL-compliant integrity constraints ensures
that information is not improperly deleted and that inserted data meets
column specifications. INFORMIX-SE provides two types of integrity
constraints: referential integrity and entity integrity.

Referential Integrity

Referential integrity allows users to define and enforce relationships
between columns. For example, INFORMIX-SE's referential integrity guarantees
that information about an entry in a master table is not deleted when the
corresponding information
still exists in the detail table. This prevents users from deleting a
customer record if an order still exists for that customer in an order
table.

Entity Integrity

Entity integrity enforces acceptable data values for particular columns.
This check allows the DBA to specify a range of permissible values. Default
values allow users to specify a default of any value that is compatible with
the column datatype.

Stored Procedures

Stored procedures are SQL commands and program statements that are stored in
the database as a named procedure to ensure consistent implementation of
commonly used operations. Stored procedures maintain common, optimized
application routines in the
database rather than in the application program.

Stored procedures reduce the amount of network traffic for database
operations because stored procedures can handle multiple tasks-such as
insert, update, and delete-with a single command. After processing the
request, the stored procedure returns only
the results of the request rather than the numerous result sets for the
individual queries.

Figure 2: Screen Image, Transaction with Stored Procedure
(20k, GIF)
Figure 2: Stored procedures reduce the amount of communication between
the client and the database server.

Triggers

A trigger is an alternate method for invoking a stored procedure. Rather
than the application program calling a stored procedure to enforce a
business rule, a trigger can be defined, which will cause the database
server to automatically execute a
stored procedure (or SQL statement) when any attempt is made to INSERT,
DELETE, or UPDATE a field in the table. Since triggers are stored in the
system catalogs, the need for application programs to maintain redundant
code has been removed. Likewise,
consistent integrity constraint enforcement across all transactions is
guaranteed since triggers cannot be bypassed.

Security

INFORMIX-SE provides two levels of access privileges to ensure database
security. Database privileges control access to the database and the
privileges for creating tables and indexes in the database. Table
privileges specify the operations that a user
is allowed to perform against a specified table.

INFORMIX-SE supports alter, insert, and delete security at the table
level while enforcing select and update security at the column level.
Separate privilege statements are used to grant and revoke the appropriate
access level to users. No separate
database login is required since INFORMIX-SE applies security at the user's
login level.

Stored procedures provide an additional security mechanism by establishing
their own permissions, different from the data permissions. The owner of a
stored procedure grants users the right to execute the stored procedure梬hich
allows the user to
perform all the SQL operations in the procedure梑ut restricts other access
against the database. By using stored procedures to prohibit users from
performing operations against the database, except through authorized stored
procedures, DBAs can elevate
security to the procedure level.

Client/Server Ready

With INFORMIX-SE, client/server connectivity is automatically built
in梕nsuring that any Informix application using connectivity libraries
can transparently view and modify databases in either client/server
or local mode.

Figure 3: Screen Image, Client/Server Capability
(16k, GIF)
Figure 3: Informix tools and servers are built with connectivity
libraries and require no additional networking products梐lthough
Informix 5.0 tools in a remote client/server mode require INFORMIX-NET
to establish connection.

INFORMIX-SE System Administration

INFORMIX-SE does not require UNIX kernel modifications for shared memory
or the use of character-special devices-raw devices. Since INFORMIX-SE
takes advantage of the native UNIX operating system and file manager,
backing up databases and tables is as simple as backing up any other
UNIX file梞aintenance requires no special skills and no complicated
utilities must be learned.

INFORMIX-SE supplies easy-to-use utilities that allow users to migrate
INFORMIX-SE data in ASCII format to another INFORMIX-SE database or to
Informix Dynamic Server�, Informix's high-performance, parallel processing
database server for the UNIX
operating system.

Another important aspect of database administration is index verification.
Should INFORMIX-SE ever indicate that an index on a table has been corrupted
by a system failure, the bcheck utility can quickly verify the integrity of
the index file and
rebuild those indexes that have been damaged by the failure.

Standards Compliance

INFORMIX-SE meets the requirements for ANSI SQL-92 Entry Level with a few
exceptions. (Notable exceptions include no support for delimited identifiers,
or transactions that can be serialized.) It also supports numerous
Intermediate and Full SQL-92
features, and some SQL3 features and functions.

INFORMIX-SE passes a subset of the National Institute of Standards and
Technology (NIST) Federal Information Processing Standard (FIPS), number
127-2 test suite for Entry Level SQL, database sizing, and the FIPS Flagger.

NIST certifies conformance with ANSI and FIPS standards. The NIST SQL test
suite is currently the only test suite available for validating ANSI SQL
compliance.

Database Migration

If your information management requirements change in the future, INFORMIX-SE
applications for UNIX can be easily migrated to Informix Dynamic Server.

Informix Dynamic Server is Informix's powerful, multithreaded database server
designed to exploit the capabilities of both symmetric multiprocessor and
uniprocessor architectures to deliver breakthrough database scalability,
manageability, and
performance. Informix Dynamic Server provides superior transaction processing,
high availability, data integrity, mainframe-caliber administration, and
multimedia capabilities梐ll within a single, client/server-ready package.

X/Open Native Language Support

Informix's implementation of native language support (NLS) is based upon
the X/Open� XPG3 specification. INFORMIX-SE 7.2 NLS capability supports
single-byte, eight-bit NLS platforms. This allows INFORMIX-SE to collate
character strings, print dates,
and accept currency input in the rules and formats required by the country
where the products are being used. X/Open-compliant NLS also provides
worldwide support of database applications, so applications can be migrated
to multiple languages while
maintaining the same functionality.

INFORMIX-SE Version 7.2 Specifications

Table Capacity Maximum number of rows per table 1 billion or prior disk
space limitation Maximum length of a row 32,767 bytes Maximum number of
columns per table 32,767 Maximum number of indexes per table unlimited
Maximum number of columns per index 8
Maximum size of any index key 120 bytes
Field Capacity Maximum character field size 32,511 bytes Maximum size of
a quoted string 256 bytes Access Capacity Maximum number of defined
databases per INFORMIX-SE system unlimited Maximum number of defined
tables per INFORMIX-SE system unlimited Maximum number of active users
per INFORMIX-SE system system dependent
Maximum number of open tables per INFORMIX-SE system UNIX open files
limit Maximum number of tables in a join UNIX open files limit Maximum
number of locks per INFORMIX-SE system, database, user system dependent
Maximum number of open databases per user
1 Resource Consumption Memory Requirements Fixed memory for INFORMIX-SE
database server process (re-entrant) 1.2 megabytes User process virtual
address space requirements (average varies by application 180 kilobytes
Disk Requirements Disk space required for INFORMIX-SE 13 megabytes

Note: These specifications were compiled on a Sun� SPARCstation� wit
h SunOS� 4.1.1. These specifications may be different for other machines
and operating system configurations, and should be used only as guidelines.
Dependent on disk space, system
configuration and/or the operating system configuration.

--
  朋友一生一起走  那些日子不再有
  一句话  一辈子  一生情  一杯酒
  朋友不曾孤单过  一声朋友你会懂
  还有伤  还有痛  还要走  还有我

※ 修改:．superman 于 Sep 10 20:55:21 修改本文．[FROM: poster.hit.edu.c]
※ 来源:．紫丁香 bbs.hit.edu.cn．[FROM: aceraltos.hit.ed]

Linux 版 (精华区)