Shield’s Remote Journal-Based DR Solution Matures at V2R1
December 11, 2007 Alex Woodie
Shield Advanced Solutions last month launched a new version of Receiver Apply Program (RAP), a low-cost disaster recovery (DR) solution for i5/OS that is largely based on the remote journaling technology IBM built into the system software of the System i. With RAP V2R1, the company has introduced new features that should make it easier for customers to get back on their feet following a disaster, as well as making the product available to a broader audience.
Shield first introduced RAP earlier this year to provide an alternative to i5/OS high availability software, which the company feels is too feature rich and too expensive to adequately serve the needs of the majority of small and midsize System i shops. With RAP version 1 (cost: $2,000) installed, System i shops could rely on the automation of remote journaling to move their data to a secondary System i or iSeries server–or a second partition on the same production machine–and use RAP to make sure that any changes to their data (as well as the objects covered by remote journaling) are applied to the back-up instance of their applications.
Since RAP became available this spring, Shield has been pleasantly surprised with the reception of the product, says Chris Hird, president of the Ontario, Canada, company. In just the last couple of months, dozens of companies have started RAP trials, and Shield has sold its first licenses. With all that customer interaction has come a lot of feedback on what features System i shops would like to see added to the product.
One of the most common requests was support for i5/OS V5R3. When RAP version 1 shipped, it only supported the latest release of i5/OS, in large part due to a message monitoring feature that IBM added in V5R4. “We had a lot of interest in this technology but more and more people kept asking for a V5R3 solution so we decided to create a message queue-based solution as well,” Hird says. By adding i5/OS V5R3 support with RAP V2R1, Hird expects to significantly expand the potential user base for the product.
Another common request was to be able to switch the direction of replication. Initially, Hird envisioned RAP being used to replicate data in one direction–away from the disaster to the backup machine. But it turns out that customers wanted it both ways (and as the old saying goes, the customer is always right). “As time went on and we started to get involved in the sales cycle, we found that people wanted to be able to at least switch the direction of the replication.” This was not necessarily a simple thing to do, but with a little help from IBM, Shield was able to make the change, so that any files, objects, or IFS files that are journaled on the source system are automatically created on the target system as well.
RAP V2R1 also brings support for new object types. With version 1, the software supported all of the objects types automatically covered by journaling, including data areas, data queues, and IFS files. However, customers also wanted to be able to replicate their user profiles, so Shield provided customers with a way to do this in V2R1.
Another new (somewhat HA-esque) feature added to V2R1 is an auditing tool to ensure that objects are properly synchronized. While Hird insisted to customers that IBM’s remote journaling technology was responsible for the replication and performed it well, that just wasn’t good enough for them, he says. “People don’t trust the product,” he says. “Even though IBM is doing everything from front to back, they still want a second opinion somewhere.”
The new auditing feature in RAP V2R1 enables customers to see if there are any discrepancies between their two systems, and to choose to have them re-synchronized (by selecting them from a list) if any errors are detected. Once this customer request was addressed, Shield received another customer request to have record-level views of replicated data in the audit tool, and to have any record-level replication errors automatically fixed, which is also new with V2R1.
RAP V2R1 also gains closer integration with Job Genie, the high availability utility developed by Shield for replicating job queue content to the remote system, so users can keep their production systems flowing smoothly during a server disruption. (This function is not a feature of HA products, giving Shield a comfortable niche to work from.) RAP users have access to Job Genie menus with this release, and the integration will grow in subsequent releases, according to Hird. “This release is only the first of many changes we have planned for integrating the two products,” he says.
While some of the new features in V2R1 may sound a bit like HA–the two-way replication, the object synchronization, the support for user files–they are the features that customers are demanding. But as he accommodates requests, Hird is adamant that RAP maintain its roots in DR and not slip upslope to compete with the HA products.
“The one thing that changes constantly is your data. If you can protect that, you have a chance to survive following an outage,” Hird says. “If you want the HA, go spend a hundred grand. We’re not going to try and compete with those guys.”
Pricing has also changed with V2R1. Previously, the whole product could be obtained for a $2,000 license. With this release, a license for the RAP base module costs $5,000, while the Option 1 module, which adds support for object and profile replication and auditing, costs $2,500.
For more information, including a white paper on how to build your own RAP-like DR solution (Shield says you can’t do it affordably), visit the company’s Web site at www.shield.on.ca.