DataMirror Updates High Availability and Data Replication Software

by Alex Woodie

North American IT managers are finally starting to take disaster recovery seriously, according to a new study from IT research firm Gartner. The Connecticut analyst group found a 650 percent increase in the number of IT managers backing up their data in the last year, either to a tape or to another disk through data replication. This is good news for high availability software providers, including DataMirror, which recently unveiled new releases of its iCluster and Transformation Server offerings.

In September 2005, Gartner interviewed 105 IT managers of North American companies about their backup and recovery practices. The analyst group found that 45 percent of them reported backing up their data, either to a local device or to a remote disk. While the flip side of this study is a little disconcerting (over half of you aren't doing anything to back up your data?!), this still represents a significant improvement compared to a similar study Gartner conducted a year earlier, in which it found that only 6 percent of IT managers reported backing up their data.

iSeries shops are also becoming more aware of the need for business continuity strategies, says Amit Agarwal, a product manager with Markham, Ontario-based DataMirror. "In a lot of cases, what we're seeing, especially people who are in areas where a chance of a disaster occurring is higher, such as Florida or Louisiana, they're more alert than they were before these events occurred," he says. "There's a general heightened awareness in all areas. People are aware disasters can occur in all areas."

It's no wonder that DR strategies are on IT managers' minds. One only has to look as far as the Gulf Coast and Hurricane Katrina in August to see the kind of widespread destruction Mother Nature can unleash. With powerful winter storms battering the East Coast and Upper Midwest this fall, and another busy hurricane season forecast for 2006, it would seem to be a good time to implement business continuity strategy.

Luckily for OS/400 shops, we're smack dab in the middle of a bit of a price and feature war among OS/400 high availability providers. One of the old guard looking to solidify its hold on the market is DataMirror, which last week unveiled a new release of its flagship OS/400 high availability product, iCluster.

iCluster 2.2

With iCluster version 2.2, DataMirror is aiming to improve basic monitoring and management of the product, Agarwal says. "Our goal for this product is to make it as user-friendly as you can, so users don't have to spend time" doing basic management tasks, he says. In this regard, DataMirror has made improvements to a graphical console that enables administrators to monitor replication on both the source and the target system, and to start and stop replication jobs.

DataMirror has also enhanced iCluster's object management capabilities by enabling administrators to configure multiple objects to be refreshed as a group, without using journaling. Administrators can also control multiple object sets at the same time using new "global level" commands for starting, stopping, setting the starting position, and setting the switch-over position.

Support for remote and local journaling has also been enhanced with this release. (The company offers users the option of using remote journaling or its proprietary journal-scrape method as a data transportation mechanism.) iCluster supports more types of OS/400 objects, including IFS and Large Objects (LOBs) when remote journaling is being used, and also features new remote journaling cleanup functions and improvements to remote journaling status.

Finally, DataMirror has improved the status-checking procedure during file replications. Specifically, enhancements have been made to the synchronization status checks for triggers, files, IFS objects, and data queues. These improvements are aimed at ensuring that users have the confidence that their source and target systems match, so that a failover, if ever initiated, can proceed as painlessly as possible. While this is the entire point of having high availability middleware keeping two or more iSeries servers in synch, all too often, high availability shops are unsure of their status and unwilling to initiate roll-swaps.

Transformation Server 5.3

DataMirror is also shipping a new release of its multiplatform data replication product, Transformation Server, which is used to pump data from one database to another for the purposes of loading a data warehouse, duplicating data sets for auditing or test purposes, and for migrating to new platforms. While this product shares some capabilities with iCluster, it is really aimed at accomplishing different data processing tasks.

Like it has done with iCluster, DataMirror has concentrated on making it easier to accomplish basic setup and management tasks with Transformation Server 5.3. For example, it has simplified the creation and deletion of integration projects, and made it easier to work with large table sets through a new table-selection interface that uses standard wildcards. There's also a new wizard for configuring audit processes and creating audit trails.

During runtime, administrators can use a new latency monitoring and alerting tool that shows the health of data replication and alerts them when processes the latency, or delay, exceeds a given threshold. If conditions deteriorate, administrators can also stop replication objects from within this monitor; it also allows them to start replication objects.

Other improvements have been made in the detection and resolution of conflicts that occur during bi-directional replication and in the generation of summary data for data consolidation projects, such as data warehousing.

DataMirror is currently discounting 20 percent off the cost of iSeries licenses for both iCluster and Transformation Server, as long as purchases are made before December 31. Detailed pricing information for these products was not released by DataMirror. OS/400 shops can also get discounts from IBM when they purchase new iSeries hardware to run iCluster, which has a value-added enhancement (VAE) status from IBM for the High Availability/Continuous Operations (HA/CO) category.