What is Disaggregated Storage?
Disaggregated storage is a form of scalable storage offering the performance benefits of direct-attached storage with the flexibility of a storage area network (SAN). It features multiple storage devices configured to function as a logical storage pool that can be reconfigured as needed without modifying the physical connections between them.
What does disaggregated mean?
In a traditional infrastructure model, storage devices are physically connected to specific compute servers. But in a disaggregated architecture, organizations decouple the memory, compute, and storage resources of their data center infrastructure, allowing them to be provisioned and maintained separately. Doing so makes the data center environment more adaptable, with the ability to scale resources as-needed, based on application and workload requirements—without interrupting the availability of applications running in the environment.
How does disaggregated storage work?
In this type of composable infrastructure, multiple physical storage devices are connected to one another over a network fabric of virtualized connections to form a software-defined logical data pool. Today’s high-speed network protocols make it possible to pool together resources that are not in the same physical location while providing the high performance, low latency, and fast accessibility needed by many business applications.
What is disaggregated storage infrastructure?
With direct-attached storage (DAS), the storage capacity of a server is limited by the number of PCIe slots it has available. While software-defined storage (SDS) allows for storage pooling, the server’s central processing unit (CPU) has to manage access to the storage pool as well as running applications plus replication, encryption, and other services. This process can create extra complexity, higher CPU load, and greater latency, creating functional limits to how much storage can be added.
Disaggregated storage infrastructures separate the server’s storage and compute resources. Multiple storage devices are combined into a logical storage pool, with storage resources provisioned to specific server instances. The storage devices are connected in a network fabric, such as non-volatile memory express (NVME), emulating a SAN and allowing for the flexible scaling of storage resources available to an application. But while a traditional SAN typically couples workloads with specific storage resources, disaggregated storage can include cloud or other non-local storage. This allows disaggregated storage to retain the per workload design and scalability of DAS while offering the economic and management benefits of non-local storage.
What are the benefits of disaggregated storage?
Many modern applications require the ability to rapidly scale storage resources as increasing volumes of data are generated across the enterprise. Because disaggregated storage can be provisioned separately from other data center infrastructure, new storage can be added as needed without having to purchase unneeded resources. It also helps organizations support a more diverse set of workloads, as disaggregated storage can support multiple applications on a single storage platform.
Performance and resiliency
With storage resources decoupled from compute, data migration is simply a matter of pointing new applications or VMs to the original data location. Applications can therefore be migrated quickly, supporting higher availability. And with storage resources shared across multiple servers, disaggregated storage reduces the number of data silos in a data center environment, increasing resiliency in the event of a hardware failure.
As part of a composable infrastructure, IT administrators can add, upgrade, or replace specific resources in a disaggregated storage model without having to replace entire systems. Provisioning can be automated and streamlined to deliver storage resources to applications and servers as needed, without making changes to the physical network connections.
HPE and disaggregated storage
HPE has combined the hyperconverged experience with the flexibility of disaggregated storage in HPE Nimble Storage dHCI. Built to streamline your infrastructure via data-first modernization, this disaggregated HCI for mixed workloads at scale unlocks IT agility while ensuring apps are always-on and always-fast. It includes business-critical speed and resiliency alongside independent scaling of compute and storage.
Powered with HPE InfoSight, the industry’s most advanced artificial intelligence for infrastructure, HPE Nimble Storage dHCI gives enterprises ultimate simplicity for their virtualized environments with fast application performance, always-on data resilience, and resource efficiency. Eliminate over-provisioning, reduce storage costs, and extend apps and data across your hybrid cloud with bi-directional data mobility with no egress charges.
Plus, this solution is available through the HPE GreenLake edge-to-cloud platform, delivering virtual machines (VMs) as a service to accelerate time to value and simplify IT management. As a cloud consumption experience, you pay monthly for what you use, convert capital to operating expense, and scale compute and storage on-demand.