Big box storage consolidation aims to eliminate all of an organization’s storage systems. A big box consolidator is a single system with both flash and hard disk drive media. These systems also leverage DRAM as an acceleration tier. A single system with all of the organization’s data should reduce storage administration time and, in theory, should reduce long-term costs through better resource efficiency.
The Challenges to Big Box Consolidation Success
The first major challenge to big box consolidators is upfront cost. All at once, the organization must purchase a system that can meet the needs of all enterprise workloads. While technically possible, there can be a significant cost attached to such a system.
To fulfill its long-term goal, the consolidated system should provide sufficient capacity to accept data migrations from other systems. It also needs enough DRAM and flash to meet all of an organization’s performance expectations. However, the storage systems that meet those requirements already exist within the data center. Not only is the big box consolidation an all-at-once purchase, it can replace and eliminate existing storage resources.
Once purchased, the big box consolidation system creates a second major challenge: migration. IT now needs to migrate data from all the various storage systems in the environment to the consolidated system. They also need to complete the migration quickly since big box consolidation starts to generate an ROI once data migration is complete.
Once migration is complete, a big box consolidation system creates a third challenge that IT needs to address: dealing with noisy neighbors. Surprisingly few consolidated storage systems provide specific storage quality of services (QoS) capabilities. QoS can help ensure critical applications get their required storage performance and capacity resources. Lacking QoS means the storage system needs to count on excess flash, DRAM and storage compute resources to meet the needs of all work loads.
If QoS or excess resources can’t deliver its performance promise for every workload, then a fourth challenge emerges: one-off storage systems. Since the workloads running in the data center at the time of the consolidated storage system purchase are a known variable, the system, initially, meets its goal of a single consolidated storage system.
The problem arises when a new, unexpected workload emerges. New workloads may not support shared storage, require a different unsupported protocol, or stretch performance requirements to new levels. At that point, the user or IT requests a “one-off” storage system. Over time, the number of one-offs increases and the once-consolidated storage architecture fragments once again.
The continuously changing and evolving data center creates the fifth challenge for big box consolidation. The chances of accurately predicting the next big application or operating environment as well as its impact on storage makes consolidating to a single system very difficult. While not impossible, it requires a very flexible system where components can change with the data center.
Alternatively, the organization can embrace change. The cost of a big box storage system places it at a disadvantage, but it can make up for the expense by simplifying management. However, instead of taking a chance on big box consolidation, IT can embrace storage system diversity but layer in a storage monitoring solution that centralizes storage management across multiple systems.
The quest for a cost-effective storage consolidation method continues. If selecting a big box consolidation option is not desirable, then software-defined storage may provide a compelling alternative. In our next blog we examine software-defined storage to see if it can help solve the storage sprawl problem.