Q&A: Brandon Tanner on the Hybrid Cloud

Brandon Tanner
We recently sponsored a Disaster Recovery Journal (DRJ) webinar, during which Brandon Tanner, our senior manager, discussed the evolution of hybrid cloud disaster recovery as a service (DRaaS) and the challenges addressed during its development. (If you weren’t able to attend the webinar, you can listen to it here.) During the Q&A session at the end of the webinar, attendees wanted to know more about how hybrid cloud DRaaS fits into their work environment. We've highlighted a few of their questions below.

Q: How does a managed service in the cloud differ from one our IT team manages, and who is responsible for what?
A: It varies depending on who the managed service provider (MSP) is, but if the MSP offers a hybrid solution, they typically handle both environments. So, for example, instead of your IT team handling a particular on-site infrastructure and solution, the MSP handles both the on-site and off-site component, whether it's a public or private space.

That service provider is tied to service level agreements that give you remediation both for local and off-site solutions, so it's a seamless end-to-end solution. With an in-house solution, you're on the hook for managing it yourself.

Q: What specific workloads are best suited for the hybrid cloud?
A: It varies depending on your business. For example, data analytics and seasonal demands are some of the workloads the public cloud does a good job of.

Dedicated workloads specific to the organization may have certain sets of data, parameters, types of software or uses associated with them. These workloads might need to be managed locally to ensure connectivity, minimize bandwidth requirements and keep costs down. It depends on how an application is built and how users access systems and data. So you have to understand what apps people are accessing and what speed those apps require. You also need to know whether or not they need to run independently if, for instance, the outside network is unavailable.

Q: What are your strategies for providing DRaaS to customers who have a mixed environment of VMs and physical servers?
A: The solution needs to address how you handle both physical and virtual environments and how they fit into your data management strategy, whether it's data replication or recovery. You may have hardware that's replicated to other hardware, and you may have your virtual environment that's replicated to a virtual environment. Or you may have an on-site solution that's backing up both physical data and virtual environments locally. Your recovery strategy then becomes a matter of asking yourself, "Do I need to dropship equipment in, do I need to keep spares on-site, or do I want to replicate that data off-site, where there's spare hardware that can be used?"

In our experience, from a recovery standpoint, we take physical infrastructure and recover it into a virtual environment, and oftentimes, once we've done that, the client stays in the virtual environment. The only exception is when the client uses equipment with a specific use. We've also seen a lot of testing that has moved the physical world into a virtual world. But you can't virtualize everything, so you have to account for that hardware component as part of your solution, both in a private infrastructure and in a public cloud infrastructure.  

Q: What are some of the gotchas to be aware of with hybrid cloud and DRaaS offerings?
A: Number one is connectivity and communications, both WAN and LAN. You could say it costs you a penny a gig to store things up in the cloud. But you still have to be able to access it. Connectivity could be a major gotcha, depending on the architecture of your solution. If you put everything up in a public cloud, and you're running the users in the private cloud, all the data has to move back from that cloud environment. You're moving a lot of data back and forth, so architecture related to your applications and systems is critical.

The other thing is cost containment. With these hybrid models, it's easy for a private cloud provider to give you a fixed cost or a model with some variability. If you have a hybrid model with stuff in the public cloud and you need to recover something or need help with an issue, a lot of those costs are a la carte. They're advertised as storage costs, cost of server instances, those kinds of things. That all comes with the hybrid cloud solution, so you need to make sure that either you or your provider has the knowledge to account for some of those additional variable costs.

For more cloud Q&As, check out this post