In a recent webinar “How to Create A Two-Tier Enterprise With All-Flash and Object Storage”, Storage Switzerland discussed the concept of a two-tier data center. The goal is to reduce the overall number of storage systems in the data center to simplify management and decrease costs. Thanks to all-flash storage and object storage we have the technology and the motivation to make this leap. A critical element of the two-tier data center is for the primary storage system to be able to internally tier data.
What is The Two-Tier Data Center?
The two-tier data center is typically made up of three storage systems. The first is a primary storage system that will hold all the active and near-active data (data that has been accessed in the last two years). Next is a protection storage system that is used to protect the primary storage system and all of its active data. Then there is secondary storage, more than likely an object storage system, that stores all data NOT accessed in two years. Typically 33% of the entire data center’s data set is in the active to near-active tier. That means almost 70% of its data is on that secondary tier.
What is Internal Tiering?
Internally tiering data means that the storage system has the ability, without external software packages, to move data between tiers of storage within the storage system itself. When data is internally tiered, it can move quickly between tiers and isn’t impacted by outside influences. It is the ultimate in seamless data movement.
The rise of cost-effective all-flash arrays may imply that internally tiering data is a feature that is no longer needed. But this point of view ignores the fact that there will continue to be multiple types of flash for varying use cases. For example, today there is NVMe high-performance flash and SAS based high capacity flash. In addition, not all data centers can justify an all-flash investment, these data centers will want the option to use a flash tier and a hard-disk based tier, but again internal to the primary storage system.
Why is Internal Tiering So Critical?
In the two tier-data center design, the primary storage system will hold all data accessed in the last year or two. The secondary storage system, probably object storage, will hold data that has not been accessed in the last two years. The chances of the first set of data being accessed versus the second set of data are night and day. The first set of data could be recalled and accessed almost continuously. As a result, it benefits greatly from having the recalls happen internally instead of across storage systems.
Once data is dormant long enough to make it to the secondary storage system the chances of recall decline significantly and as a result how data is retrieved from that tier is nowhere near as important as how it is retrieved from the primary tier. While most data centers will still want seamless access to this data, it is not required. The goal for the secondary tier is to store this data for a long time and to do so as inexpensively as possible.
Capabilities like internal data movement and even cross-system data movement have been available for some time. But the advent of all-flash systems and object storage now implement those capabilities practical.
The complete details of the two tier data center are laid out in our webinar, “How to Create A Two-Tier Enterprise With All-Flash and Object Storage”, watch it now.