Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure
The follow-up article to our hybrid cloud storage buying guide provides tips, advice and cautions for those seeking to implement a hybrid cloud storage environment.
Some companies are already well on their way to building a sophisticated internal cloud. Others have gotten their feet wet in the public cloud and know the ropes there fairly well. So there is no one-size-fits-all way to go about implementing a hybrid cloud. It all depends on where you are right now.
Hybrid Cloud Storage: Starting Points
Building a hybrid cloud storage deployment could just as easily start from a public or private cloud focus. “Because a hybrid cloud combines public cloud and private data center principles, it's possible to plan a hybrid cloud deployment from either of these starting points. Picking the better starting point, however, will make it easier to address business goals,” said Molly Rector, Chief Marketing Officer, at Data Direct Networks (DDN).
DDN has two product lines that can be bridged together in a private cloud tiering configuration. GRIDScaler File Storage and WOS Object Storage. Data can be transferred from the file storage to the object storage on premise and also shared offsite in a public cloud.
Identify User Type and Workflow
Rector defined a hybrid cloud as a storage environment with requirements of a single data set, requiring storage and access both on-premise and at offsite locations. As such, the data set requires different access, permissions, security and usage guidelines.
For example, DDN typically sees hybrid cloud deployment in large research environments, universities with multiple campuses, government organizations, and within the service provider market. These represent complex environments with very different user needs and varying requirements.
“Clearly identify upfront the user types, bandwidth requirements and data workflow, as well as identifying application vendors that will need to be a part of the solution,” said Rector. “To be effective, a management strategy for hybrid cloud deployment should address configuration management, change control, security, fault management and budgeting.”
Treat the Cloud Like a Hardware Purchase
Greg Schulz, an analyst with StorageIOGroup, calls hybrid clouds a homerun for many environments to bridge what they have existing, while addressing common concerns about public solutions. But he advises storage managers to treat a hybrid cloud project just like they would a traditional storage hardware or software purchase.“The key is doing your due diligence with cloud solutions, technologies, products, providers and vendors, the same way that you would vet any traditional hardware, software or services to avoid surprises,” said Schulz. “This also means documenting what your cloud storage concerns are so that you can then evaluate applicable solutions and services, as well as the technologies that will actually bridge the two worlds or environments.”
Remain it Control
Perhaps one of the hardest aspects of any cloud approach that veers into the public cloud is control. Adam Fore, Director of Cloud Solutions Marketing, NetApp, cautions users who are offloading certain elements of infrastructure and applications to the cloud to make sure they don’t also try to offload responsibility and control of their business data, whether consciously or unconsciously.
“They have undoubtedly worked for years to align their datacenter workloads with the right levels of performance, data protection, and governance,” said Fore. “As they seek to implement hybrid clouds, they will need to consider how to maintain data control while leveraging a variety of different cloud services.”
Easier said than done with some cloud services. It’s all a matter of reading the fine print in contracts and making your own requirements for control and security known upfront – and following through to see they are carried out. Fore noted that small initial compromises when implementing cloud computing are often the first steps in rapidly creating fragmented and chaotic data environments. The results can be difficult and costly to unravel.