You probably won't be surprised to hear that vendors often use news or statistics as a leverage point to pitch their products.
It's especially true with products such as backup and data-recovery solutions that do not target specific business processes
but rather make recovery possible when something stops or threatens those processes.
For example, after my blog on yet another case of disappearing backup tapes, I got several reminders from vendors telling me how vaulting solution "X" or vaulting solution "Y" prevents that from happening
because, well, no tapes are used.
Tapeless vaulting could be the topic of a whole other column (and perhaps will be), but for now let's focus on another very
controversial and long-debated topic: DR (disaster recovery). The latest pitch on DR comes from XOsoft, a fairly new storage
vendor you may remember from a previous column on XOsoft's intriguing WANSynch hardware-agnostic replication product.
In short, WANSynch provides DR by building seamless remote and local replicas for various applications, including Microsoft
and Oracle databases, and Microsoft Exchange.
The XOsoft portfolio includes WANSynchHA (the HA stands for high availability), which adds automatic fail-over of business
applications in case of disruption. Equally intriguing is XOsoft's Enterprise Rewinder offering; as its name suggests, it
allows a quick recovery from data corruption with a mechanism as easy and as intuitive as a VCR tape rewinder.
Fast-forward to May of this year, when XOsoft launched Assured Recovery. This extension makes possible something that, for
many companies, has been an impossible target: testing DR procedures with minimum cost and disruption.
How does this tie back to those statistics and news I mentioned before? To support the usefulness of its new product, XOsoft
provided some worrisome data points from a recent survey, showing that testing a DR plan is a terribly expensive and somewhat
unreliable exercise in many companies.
Did we really need a survey to learn that? Probably not. Anyone who has been involved with planning and testing contingency
plans knows that even the best-conceived recovery plan becomes quickly obsolete as the company business evolves.
For example, rather trivial matters, such as storage space requirements or needed updates (virus protection, anyone?), can
change dramatically in a few months, creating a significant difference between the live instance and the DR copy of an application.
If you test your DR plan once a year, most of the effort will probably go to syncing those differences, which further increases
cost and possibly downtime. More frequent testing would significantly reduce the differences between the two systems, but
the cost of that approach often outweighs the problem it should solve. In a nutshell, this is the DR testing conundrum many
companies face.
How is Assured Recovery going to improve on that? By creating a brilliant combination of the best features of XOsoft's products.
First of all, Assured Recovery works with DR solutions such as WANSynchHA, which limits its target to the database supported
by that application and to the Windows file system. XOsoft plans to extend future versions to IBM AIX and Sun Solaris, but
for now Assured Recovery is just for Windows and the aforementioned applications.
Here's how Assured Recovery works. Imagine having your primary database at location A, constantly replicated to a copy at
location B by WANSynchHA. When you're ready to test your DR procedure, you push a virtual "DR test" button (the actual trigger
varies according to your environment). That action causes the replication to stop, but your regular application keeps running
at the primary site and stores new transactions to a buffer area.
Does this mean your business isn't down during testing? Indeed it does. Your application is still working and you can focus
on testing your recovery site knowing that your business transactions are still alive and running.
At the recovery site, Assured Recovery will bring up the DR instance of your database, and you can make sure that everything
is working as expected. You can work on solving any problems knowing that your customers are still being served at the primary
location.
When the test is completed, you push another virtual button such as "End Test," which will immediately revert the DR copy
of your database to its pretest state.
After that, the same database will be updated with the transactions from the buffer area and finally the replication will
restart. You're back in business, having verified that your recovery will actually work when needed, all with minimum cost
and disruption.
How much does Assured Recovery cost? XOsoft gives an accurate but still undefined answer: It will add 20 percent to the cost
of its DR application. That's definitely worth considering if Assured Recovery supports your applications.
Join me on The Storage Network blog to discuss this and other topics.