Mark Cree February 18, 2019
Parts of this article were first published on VMblog.com.
The transition to cloud storage has been more complicated and costlier than everyone expected. In response, we’re seeing IT teams implementing hybrid approaches that make the most sense for the business, from on-premises and public cloud to private cloud and the enterprise edge. However, the
more data that is accumulated across an enterprise—growing into billions of files and petabytes of capacity—the harder it is to move, manage and protect due to administrative, latency and performance issues. In 2019, look for hybrid cloud strategies to transform in two significant ways and for IT practitioners to adopt true hybrid cloud storage designs to solve these data gravity challenges.
Cloud storage performance will finally equalize with traditional on-site file storage
An effective hybrid cloud strategy requires information to flow freely between on-premises file storage and on/off-premises cloud storage. However, most of the interactions with file-based storage are a look-up or reference request, up to 90 percent of the time, rather than an actual file transfer. Users and applications constantly make requests for this metadata, or attributes about the file, such as last file access or modification. These metadata requests are very small in terms of size and are very sensitive to
latency; taken together, the time it takes for a cloud, whether public or private, to respond to each of these requests across the enterprise is significant.
In the months ahead, hybrid clouds will get smarter about understanding how massive amounts of files are created, used and shared. By applying deep network intelligence on metadata, IT can automate the response to metadata requests more efficiently: directly from the network instead of retrieving the information from the cloud or through an onsite file system.
Think of it as making a dash to the corner store versus the mega mart when you only need a carton of milk. This network intelligence keeps up with changes to the files and eliminates unnecessary steps or layers that can get in the way of the free-flow of information promised by the cloud. The performance improvement can slash the time to fulfill these requests from milliseconds to microseconds, which means in many cases the storage in the cloud will run just as fast as flash storage on-site. The metadata acceleration also makes legacy NAS systems perform like an all-flash NAS to extend the life and value of your existing IT investments.
True hybrid clouds will provide data mobility without compromise
Data migration and tiering techniques have existed for decades to help organizations make decisions about where to put their actual data. Most cloud data management approaches, which rely on stubs or links, and third-party file systems, struggle with ensuring the timeliness of the data availability and the
consistency of that data across locations. Instead, by putting a copy of metadata in the network, you can automatically move the inactive data to public or private clouds without making any tradeoffs. Files appear as though they’re on local storage even though the vast majority are actually in a cloud. They remain available to the application and business users that need them, even if the content has been inactive for decades. And if the business user has an application that requires a set of files to be actively accessed, they are seamlessly reconstituted on local file storage. Once the files become inactive again,
they are automatically migrated back to cloud by policy.
The major drawbacks to hybrid cloud approaches today is that performance is unpredictable and major workflow changes disrupt IT operations regardless of whether the data required is on a cloud or on-premises storage. As hybrid clouds evolve to offload metadata requests from on-premise file storage and cloud, you’ll be able to manage files and content regardless of location while avoiding changes to existing workflows, cloud repatriation costs and unnecessary management fees.
The benefits of metadata management also extend to your data protection strategy. By reducing the amount of data stored on-premises and offloading metadata requests from data backup applications, you’ll also cut the capacity and time required for data backup. Saving additional time and money while ensuring enterprise-class resiliency and fault tolerance. You’ll finally be able to access application data with consistent performance between different on-premises and multi-cloud environments without limitations. This is what a true hybrid cloud will look like in 2019.