There are two primary use cases for NAS. The first demands high performance when using NAS as primary storage to host application data and virtual machines. Modern NAS systems address this use case.
The second, and more common, use is as a storage vault for all of the organization’s unstructured data. In terms of capacity, the unstructured data vault is the larger use case but, unfortunately, NAS is the least effective for this use case.
The Problem with NAS and the Unstructured Data Vault
The primary problem with using NAS for storing vast amounts of unstructured data is that NAS is too expensive for the task. Most NAS software vendors use either proprietary hardware or off-the-shelf hardware which the customers must buy directly from them, meaning the customer pays a premium for that hardware. These systems are also feature-rich and the vendors factor the cost of the development of all of these features into the price of the hardware, increasing the price premium. Ironically, most of these features are not needed in the vault use case.
These same NAS systems lack features that are required for an unstructured data vault. Features like integrated search, geographic control over data placement, and long-term data verification are all M.I.A. from most NAS software offerings.
Is Object Storage the Answer?
Most object storage systems deliver many of the requirements of an unstructured data vault. Many can scale on the hardware of the organization’s choosing, have almost no capacity and file count limitations, and have the data preservation features to enable organizations to safely and cost-effectively, store data for decades.
The problem is that object storage systems are not NAS systems. They don’t support, as a default, the SMB and NFS protocols that users and applications use to access NAS storage across a network. They also don’t support the hierarchical folder structure of a NAS, making it hard for users to organize data.
The typical solution has been to leverage a gateway that converts SMB and NFS to object storage, while also emulating the hierarchical structure that users are familiar with. The problem with this approach is that gateways introduce a number of additional challenges. The biggest is that the gateway often comes from a vendor which is different than the object storage vendor. Involving another vendor means additional cost and finger-pointing if there is a problem. Often these gateways will also add additional features like high-speed caching or data distribution that the customer may not need, which adds additional cost and complexity.
Finally, the NAS turns the object storage system into a NAS. The portion of object storage allocated to the NAS often becomes a silo and loses many of the capabilities and benefits of the object storage methodology, like unlimited file count. While there is some gain in platform consolidation, the organization is once again limited to the capabilities of a NAS file system.
Object storage vendors have been pining for the unstructured data vault use case of NAS for almost as long as there has been object storage. The object storage advantages are indeed compelling, but the bridge to object storage, in the form of the gateway, has been a single-lane rickety bridge instead of a seamless connection to the interstate. It’s time for object storage vendors to create a smooth transition to object storage.
To learn more about NAS vs. Object Storage, watch our on demand webinar “Solving the NAS to Object Storage Challenge”.