Share this post

November 05, 2015

Achieving the pinnacle of disaster recovery

Priyadarshi Prasad - Atlantis


For all the complexities and costs involved, the one thing that traditional big-iron storage got right was their high availability and disaster recovery solutions. Developed over a decade or so, those solutions evolved and pushed the boundaries of DR driven by increasingly complex customer requirements. Influenced by an “always-on” business model of the 21st century, coupled with calamitous events such as power outages that affect millions of customers and businesses, the traditional storage solutions offer a wide range of HA & DR solutions for such eventualities. However, these traditional storage solutions come with a price — and not just the additional license costs by nickel & diming every feature and capability — they also add inhumane amounts of complexity in setting them up and managing them. Traditional storage vendors often push their HA & DR solutions because the complexities involved forces customers to go for expensive professional and consulting services that are very lucrative for the legacy players.

The good news is that (some of) the hyper-converged players have made huge strides in addressing the HA & DR gaps in their portfolios. And more often than not, these solutions leverage the inherent simplicity of their new architectures allowing customers to set up their own HA & DR configurations with little outside help. The figure below shows the continuum of HA & DR solutions that Atlantis offers to its customers today:

The Atlantis Advantage brought to HA & DR

By the look of it, the solutions above appear me-too of traditional storage. Nothing could be further from truth though. A few highlights of the above Atlantis solutions:

  • Simple

    All of the above solutions can be set up by customers on their own, with about five clicks. There are no expensive professional services required to either set up or to manage them.

  • Native

    The capabilities are also native to the Atlantis software stack, not requiring any additional hardware appliances to provide these capabilities (subtle reference to RecoverPoint or VPLEX appliances☺).

  • Efficient

    Each of the above solutions differ from typical traditional solutions in a very important aspect — they leverage the built-in inline deduplication of Atlantis architecture. This makes the solutions not only efficient in terms of storage space and bandwidth required but also more performant since duplicate writes require only metadata updates to the backend storage.

  • All-inclusive license

    The entire HA & DR feature set is included with a single license of Atlantis USX and Atlantis HyperScale appliance. No nickel & diming.

Hyper-convergence without Compromises

One of the last bastions of traditional HA & DR solution has been the ability to support a 3-site disaster recovery solution. This is the capability of protecting data at a local or metro level, and then protecting the same data at a geographic level. The use case is obvious I suppose – protecting data at a local or metro level can ensure that businesses can recover from local issues quickly with no or very little data loss. However, should there be a metro-wide disaster, data stored at the far site comes handy. This can help businesses get on their feet quickly (within a few minutes or hours depending on their DR preparedness) albeit with some amount of data loss that their solutions are designed to withstand.

I am happy to report that Atlantis now offers a 3-site disaster recovery solution to customers looking to simplify and reduce the cost of their IT infrastructure by moving to hyper-converged solutions. We combine our metro-wide stretched cluster solution and our asynchronous replication solution to provide this capability as shown in the figure below:

As depicted above, data belonging to an Atlantis volume is mirrored to another volume using our stretched cluster solution. It is important to note that the mirror volume may be at a second site about 100 miles apart or it can even be within the same data center as the first volume. The stretched cluster solution ensures an automatic (no administrator action required) and transparent (to applications) failover of storage resources in an event of a disaster. Data also gets copied to the volume at the far site, thousands of miles away, using Atlantis asynchronous replication at a minimum resync interval of an hour for now.

I believe Atlantis is the only hyper-converged solution to offer this advanced DR capability to customers. Looking at the pace of innovation in the hyper-converged world, I am sure others will catch up. The good news from a customer standpoint is that the reasons to keep and manage their legacy storage infrastructure are getting fewer and fewer. While the business justification for going hyper-converged has always been strong, the comprehensive HA & DR solutions now provide a strong technical justification as well.

What’s your take on adopting hyper-converged solutions? Join us to discuss more at

@AtlantisSDS, or @Priyadarshi_Pd

Current rating: 5 (4 ratings)