In this ATC Insight

Cloud storage is a method for provisioning on-demand or permanent data storage capacities via the internet. Enabled by cloud computing — the process of provisioning on-demand access to computing resources — and facilitated by cloud services providers (CSPs), this type of off-site storage allows business leaders to securely offload corporate data to remote sites.  

By opting for cloud storage, you eliminate the capital expenditures (CapEx) and operating expenses (OpEx) of managing on-premises hardware and supporting data infrastructure. The servers where the data is hosted are managed, maintained and secured by a cloud storage provider such as Google, AWS or Azure. Per Accenture, migrations to the public cloud can generate 30 to 40 percent savings in total cost of ownership (TCO).  

You also gain on-demand, elastic access to more resources, as well as low-latency access to data, stored in a cloud database, data warehouse or a data lake. Each option comes with a different degree of flexibility in terms of supported data types and formats, interoperability and cost structure.

Types of cloud storage

By format  

Similar to on-premises databases, cloud storage lets you host data in three formats:  

Object storage: Every piece of data (structured or unstructured) gets "packed" into a block, alongside a unique identifier (metadata). Objects are uncompressed, unencrypted and rapidly accessible via REST APIs, used to request authentication and manage access permissions. These characteristics make object storage a strong contender for cloud-native applications.  

Examples of object-based cloud storage solutions include:  

  • Amazon Simple Storage Service (S3)
  • Google Cloud Storage
  • Azure Blob Storage

Block storage: Single storage columns (i.e., cloud storage notes) are split into individual instances (blocks) featuring data entries of similar size. Frequent element of SAN (storage area network) architectures, block storage is less flexible than object storage, but it allows users to create high-performance databases for long-term data storage.

Examples of cloud-based block storage solutions include:  

  • Amazon Elastic Block Store (EBS)
  • Google Cloud Persistent Disk
  • Azure Disk Storage

File storage: Shared file-storage systems provide users with hotkey access to pre-transformed, hierarchically organized data.  

Examples of cloud file storage apps:

  • Azure NetApp Files
  • Azure Files
  • NetApp Cloud Volumes Services
  • AWS FSx
  • Google Cloud Filestore

By deployment  

In addition to the different types of storage available, cloud storage varies by the deployment approach that is used for cloud network provisioning. The following represent the three options deployed by modern enterprises.

Public cloud storage: Cloud service providers let you provision, pool and share cloud storage resources with other businesses. Effectively, you rent a partitioned storage instance on a virtualized server that is shared with other clients.  

Private cloud storage: Instead of storing data on a CSP-operated server, you can either virtualize an on-premises server or commission a dedicated private cloud from the provider. In each case, you have a private connection to the cloud storage, but a more limited scaling capacity.  

Hybrid cloud storage: The hybrid model combines these two storage options, allowing your company to choose where different data is stored. Sensitive financial data can be hosted locally to stay in compliance with archiving and replication requirements, for example, whereas other insights can be hosted in an analytics-ready state by a third-party cloud storage provider.  

Cloud storage vs. storage virtualization: The main differences

Cloud storage and storage virtualization are frequently confused because virtualization is a core part of cloud computing. There are several underlying differences, however, including:

Storage virtualization assumes the usage of an on-premises hardware system to create a simulated storage environment, primarily accessible via a local connection or intranet.  

Cloud storage, in turn, implies the use of a pooled storage resource, shared over the internet. It's a multi-layer environment, featuring resource management, administration and automation layers, among others, on top of virtualization that is performed on bare-metal hardware.  

Though these descriptions should highlight the key differences, it should be noted that you can transform virtualized storage into a cloud environment. In general, this is done by adding abstraction, resource pooling and management layers, as well as self-service capabilities.  

That said, most prefer to use virtualized storage for local data hosting needs. This is especially true for users looking to provide, for example, extra disk space to on-premises apps or set up a low-latency production environment.

Advantages of cloud computing storage

As with most options in the marketplace, cloud computing storage definitely has its advantage. For most users, the top advantages include:

  • Just-in-time capacity: accommodate expanding data volumes (i.e., from edge or IoT devices).
  • Global agility: cloud data repositories are accessible from any device, at any location.
  • Durability: replicate critical data across regions to prevent accidental losses and reduce latency.
  • Business continuity: data residing in the cloud is not prone to on-site outages, breaches or service disruptions. Most CSPs guarantee up to 99.99% uptime.
  • Lower TCO: the storage provider absorbs the hardware management and maintenance costs, whereas you pay for on-demand, usage-based access to storage. Key considerations for cloud computing storage

The cloud assumes off-site data storage (under most scenarios). This increases the scope of obligations on both ends (yours and the provider) when it comes to:

  • Security: CSPs are responsible for physically securing the infrastructure, whereas as a user, you also need to implement the recommended cloud security best practices pertaining to user identity management, application authentication and network endpoint protection, among other aspects.
  • Regulatory compliance: Certain industries are bound by strict compliance rules when it comes to data storage, archival, replication and usage. From geographic limitations on allowed storage regions to state-specific data privacy regulations, there are often mandatory, non-negotiable rules to take into account.

Doing cloud storage right

Cloud storage offers unbeatable convenience for business users, data scientists and developers alike.  

As with other types of cloud services, utmost diligence is required when it comes to cloud storage architecture mapping, configuration and provisioning.

The right storage solution is pivotal to all businesses and can differ depending on workload or application. In order to understand which cloud storage solution is best for your organization, you must first understand the requirements of the application associated with the storage. There are certain items to consider: performance requirements, capacity limits and requirements, required connectivity type, accessibility of your data, ingress/egress costs and performance when accessing your data, resiliency in the architecture, latency and time needed to recover, and final cost.  

Below are a few items from WWT to help you with your cloud storage solution: