The term “cloud first” comes up often in IT circles these days. It’s a concept that calls for each new application an organization is bringing online, IT should consider the “cloud first” as a place to host that application. It sounds like a good idea, but most organizations simply are not ready for that type of commitment to the cloud. Most have “today” problems to solve, often related to storage. But that does not mean organizations should ignore the potential of the cloud all together. Instead they should look for a storage system that can solves today’s problems while building a path to the cloud.
Solving Today’s Storage Problems
Today’s storage problems typically come down to some combination of the need for more performance, the need for more capacity or the need to control costs. Both structured and unstructured workloads are pushing current storage systems to the edge. IT professionals want the performance simplicity of an all-flash architecture. Structured data (Databases) are an obvious match to all-flash systems, but now the value of placing unstructured data on all-flash systems is becoming understood, as is the value of using a file-system to host virtualized images (VMware, Hyper-V, KVM), as well as container technology like Docker. As a result, data centers are looking into all-flash NAS systems with increasing frequency.
The capacity problem is universal but is more acute in the unstructured data use case. Part of the exacerbation of the problem is the well documented explosive growth in unstructured data, but part of it is the workload variety. The reality is all-flash NAS could run workloads ranging from user home directories to Hadoop or Splunk to virtualized or container workloads.
Handling workload variety requires the all-flash NAS be scale-out, where each server (node) in the cluster contributes its resources to the unified pool. Scale-out is ideal for a multi-workload data center since it generates a lot of parallel IO traffic. Scale-out allows the organization to start small, and grow the environment as the organization grows.
The problem is most NAS file-systems are deeply intertwined with the hardware on which they run. While organizations like the turnkey simplicity of a NAS already integrated onto a server with storage, they also understand if the vendor has a software focus it allows them to adopt newer technology faster and use more commodity components that keep prices down. Being software focused is also critical as organizations look to move to the cloud. A software-based NAS can be altered to run in the cloud enabling application to run unchanged.
From Cloud First to Cloud When Ready
Despite to need to solve today’s problems, IT professionals are under pressure to develop a cloud strategy. And that pressure is well founded. When used correctly (as a tool) the cloud can add significant value to the offering that IT provides to the organization. Initially, that may be using the cloud as a low cost storage tier, then using the cloud as a disaster recovery location and then using the cloud to host or handle application overflow.
To make all these cloud options work requires more than just a cloud connection. A cloud connection typically means data can flow to the cloud, most often as a tier of storage and nothing more. If the organization wants to operate on that cloud based data it has to pull the data back on-premises.
Instead, what organizations need is a NAS that creates a cross-cloud data fabric. In this use case the cloud simply represents another location where nodes within the scale-out storage cluster can reside. Elastifile is a good example of this technology. It provides an all-flash scale-out NAS that can reside on-premises or in several public clouds. The organization can then, by policy, move workloads or data sets to and from a public cloud provider as the organization sees fit. Once in the cloud, the organization can start-up compute resources to operate on that data, then spin down those resources to minimize cost when no longer needed.
Elastifile reports that many of their current customers start off with an on-premises implementation only. Essentially, they are buying an on-site all-flash NAS, something that Elastifile does well but is not necessarily the problem that the company set out to solve. What these organizations are doing is buying into the vision. They’ve bought an all-flash NAS with a cloud future.
When the organization is ready, IT can spin up an Elastifile instance in the public cloud. They can play it safe and have the initial use of the cloud be as a tier. Just a cheap place to store data that does not need to be on an all-flash array anymore. Then they could look to replicating all the data on-premises to the cloud, and because Elastifile supports the file-centric processing of in-cloud data, they now have an on-premises all-flash NAS with integrated DRaaS capabilities. Finally, when the organization is ready and when it identifies the right applications, it can leverage the Elastifile File System to host the application in the cloud.
Most importantly, IT planners when pressed for a cloud strategy can point to the Elastifile solution and say, “There it is.”
Cloud first and multi-cloud are two sets of buzzwords that almost every IT professional is hearing. The problem is they have real problems to solve right now. Even if the cloud could help them, they don’t really have a path to the cloud with their current solutions. Companies like Elastifile provide a unique opportunity to solve the immediate press problems of today while creating a foundation that gives them a path to the cloud in the future.