Most organizations now live in a multi-cloud world. Workloads can run on-premises one day, on cloud the next, and return to on-premises soon after. A challenge for organizations looking for this type of workload mobility is the differences between the infrastructures. Current on-premises solutions, despite attempts to make them more cloud-like, haven’t worked. Most vendors forget that the public cloud is more than just a business model it is an actual way of doing business. On-premise solutions should mirror the flexibility – both in technology and economic – of the public cloud and best approach to achieve this is by leveraging composable storage.
What is Composable Storage?
Most on-premises storage systems, even scale-out systems, are hard wired. They have a given amount of storage, compute, IO and capacity resources that are shared across all workloads. This sharing of all resources leads to storage silos which leads to data center inefficiency and budget overages.
Composable storage enables organizations to specifically define what components of a storage system will be assigned to a specific workload or group of workloads. For example, a performance sensitive, Oracle environment may need extra controller power in order to keep pace with a high transaction application as well as a high number of flash drives. A VMware environment may only need two controllers and, thanks to deduplication, a very finite number of drives. Lastly, an analytics process may need more bandwidth than random IO performance. It may also only need two controllers but require additional network connections.
The needs of each of these workloads may also change from day-to-day or week-to-week. To achieve high levels of efficiency, IT may want to reallocate resources dynamically between workloads as the need arises, instead of having expensive resources sitting idle.
All of the movement of these resources is controlled through the storage software. IT administration can rapidly define a new virtual storage system out of the discrete physical components.
The challenge for IT is that few solutions provide composable storage infrastructure. Most that do are delivered as a turnkey hardware/software solution. This means the customer pays a premium for the hardware with the software licenses locked to the specific hardware. If the customer adds an additional node, they must purchase the storage software again instead of transferring the license from the old node.
Composability 2.0 – Software Only
The next generation of composable storage is solutions where the hardware and software are disaggregated. While they may still come bundled with certified hardware solutions, software and hardware are independent of one-another. Software independence means organizations can transfer licenses and quickly move to new hardware platforms when new technology justifies it.
Software independence also enables a true consumption or subscription model. While the customer purchase the hardware upfront, the software is only activated as the organization uses that portion of the hardware. Most consumption models built on legacy storage solutions — or even composable storage 1.0 solutions — require a minimum purchase by the customer. Meanwhile, upgrades take place as large chunks instead of small bites.
The second generation of composable infrastructure is more compatible with the cloud. Although composable storage 2.0 is more flexible than cloud, it shares similar properties in terms of seamless scale, upgrades and business models. Organizations looking to maintain mission-critical applications on-premises, while having the option to move to and from various cloud providers, will find composable storage 2.0 an ideal strategic solution.