What does "high availability" mean for database management in your organization? In a typical 9-to-5 IS shop, high availability might simply mean the database must be up and running at least 12 hours a day on normal business days. But for a large retail credit card company--which might be processing thousands of transactions per hour, 24 hours a day from customers around the globe--the very survival of the organization can depend on a truly high-availability database architecture. In this type of mission-critical application, high availability means near 24-hour-a-day functionality with maintenance or upgrade-related down time limited to a few short hours in the middle of the night.

To achieve this level of availability, database managers must employ one of several available methods of archiving or storing important data and then restoring that information in the event of a system failure. As databases grow in size and complexity, especially in mission-critical applications, the ability to store and recover becomes vitally important.

Some organizations have met this challenge by investing in sophisticated hardware technologies such as redundant array of inexpensive disks (RAID). These arrangements utilize firmware to allocate data among an array of disks, at least one of which is totally redundant, and a check sum technique to track data totals on the active disks. In the event of a disk failure, the check sum is used to identify the lost data, and an emergency swap is launched to recover to a replacement drive. While this and other hardware schemes deliver reliable availability, they are expensive to erect and maintain.

To balance database reliability and costs, many IS managers are now turning to a simpler, less expensive software pathway to higher availability. If you need optimum availability in a mission-critical application, you might consider these proven software-driven solutions.

Software Mirroring. Log and database mirroring provide an effective means of recovering data in the event of a media failure, without taking the database server off-line. Informix software mirroring is an ideal way to protect critical data such as root dbspace and logical and physical log files. Under this arrangement, Informix automatically and simultaneously writes selected key data to both a primary disk and to designated "mirror" disks. In the event of a disk failure, Informix instantly shifts to the mirror disks with no interruption in system performance. Informix servers also support hardware mirroring. But whereas hardware mirroring takes place at the disk level and must replicate the entire disk, software mirroring gives administrators the flexibility to choose which portions of the database should be mirrored.

Data Replication. Informix provides two powerful and flexible methods of replicating vital enterprise data. Using a full-database approach, Informix supports replication of the entire database to a secondary server that is configured with hardware and software to precisely match the primary machine. Informix continually replicates to this secondary server on a real-time, instance-to-instance basis, creating, in effect, a non-stop recovery mode that ensures uninterrupted availability.

Informix also supports the replication of a portion of the database to one or many secondary servers in an arrangement that provides important multiple-copy read/write flexibility. This unique one-to-many capability is especially useful in geographically diverse organizations, such as retailers who need to archive, protect and update store-specific data. Informix provides a recoverable queue feature that, in the event of a server or connection failure, automatically collects transactions and then forwards those transactions to the appropriate database once the system is restored.

Backup/Restore. Informix offers a number of on-line maintenance and administration features to support high database availability. Those capabilities include on-line backup and restore, partition-level backup and restore, and incremental backup. These on-line features allow 24x7 operations to backup mission-critical data on the fly and to bring the database back on-line quickly in the event of a failure.

Point-In-Time Recovery. Informix also provides a point-in-time (PIT) recovery feature that lets data management administrators restore an entire database up to a specific point in time. This feature is especially useful in the all-too-common event of an application logic error that corrupts a table or database object. Using the PIT recovery, it is possible to restore the database up to the point of the corrupting event, exposing the organization only to the loss of data compiled between the event and the discovery of the event.

Ensuring the availability of a database--with virtually no planned or unplanned down time--is often the single most important part of a database manager's job. Many must now achieve levels of availability approaching fault tolerance.

In my next column, I will discuss how some organizations are creating "near fault tolerant" systems at dramatically lower costs. If it's your job to make sure the data is there when people need it, I think you will find it to be a very interesting read.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access