Now that the cloud has become all-pervasive, this question is relevant for everyone. According to Gartner, by 2020 a corporate no-cloud policy will be as rare as a no-Internet policy is today. “Cloud-first,” says Gartner, “and even cloud-only, is replacing the defensive no-cloud stance that dominated many large providers in recent years. Today, most provider technology innovation is cloud-centric, with the stated intent of retrofitting the technology to on-premises.”
To discuss cloud storage architecture, we first need to unpack the familiar but somewhat hazy meaning of “cloud.” The cloud actually comes in several flavors, and it’s definitely not “one size (or type) fits all.”
What most people mean by “cloud” is more accurately termed the public cloud: IT services hosted by third-party providers like Rackspace, Microsoft Azure, and Amazon AWS for enterprise customers, each of which enjoys secure access to its assigned resources via the internet.
A private cloud is, generally speaking, on-premises infrastructure that offers similar advantages to the public cloud. It’s usually owned, managed, and operated by the organization it serves, but can also be managed by an expert third-party. In addition, a private cloud can be hosted in a data center owned by a third party. Confused yet? Hang on — we’re almost done with definitions.
The third cloud flavor is hybrid cloud, a computing environment which employs a mix of on-premises private cloud and third-party public cloud services, with orchestration between the two platforms.
All of this relates directly to cloud storage architectures. If you’re using a public cloud service, though, you may not be terribly interested in what kind of cloud storage architecture they use. Who cares, really, how they create, aggregate, and deliver massive storage capacity to your users, as long as it reaches you securely 24/7, and scales elastically on demand. For those who do care: In point of fact, there’s only one kind of cloud storage architecture that reliably fulfills all these requirements: object-based storage. When you look under the hood of public storage services, this is almost invariably the cloud storage architecture you’ll find.
Legacy storage platforms like dedicated appliances and NAS don’t fill the bill, and shouldn’t be expected to. After all, they were never designed for the role of cloud storage architecture. Instead, companies seeking the benefits of on-premises cloud-like storage have looked to the cloud storage architecture the big public cloud players use.
The Growing Popularity of Hybrid Clouds
Hybrid clouds are becoming increasingly common. Many companies want the stability, control, and security of on-premises IT together with the ability to expand computing power and storage capacity on demand — which public cloud services provide. The hybrid model, too, has implications for cloud storage architecture.
Meanwhile, in the public cloud, Amazon’s S3 protocol has become the de facto storage standard.
Private Cloud Storage Architecture for Backup and Archiving
Object storage is the ideal cloud storage architecture for all capacity-driven uses cases — video production workflows and origin servers for content distribution, and an array of industry-specific functions in financial services, healthcare, and the government sector. But perhaps the most widespread object storage use case across all industries is data backup.
For decades, tape drives have been the go-to technology for storage backup. If NAS and dedicated appliances are problematic for cloud storage architecture, the role is downright impossible for tape. The medium persists for backup systems, though, because it’s so inexpensive. But the sun appears to be finally setting for tape. Network Computing observes that with price declining for both hard disks and cloud storage, tape’s edge over disks has been replaced by the cloud’s advantages over tape.
This brings us right back to cloud storage architecture. Just as object storage facilitates public cloud services in general, it’s the key enabler for cloud-based backup in particular. And if object storage works for service providers, it works equally well in a private cloud for data backup and archiving right in your own data center.
But hold on: Why not simply use old fashioned on-premises disk-based NAS storage for backup? Because NAS is anything but simple — and also prohibitively expensive — to scale. With the average backup load pushing into petabyte range, such drawbacks can’t be ignored.
Okay, but why not just go with the public cloud for backup, and be done with it? We’re glad you asked! One money-maker for cloud backup storage services is the hefty per-unit retrieval fees they charge, every time you want to take data out. If this only happens once in a blue moon, it might not matter much. But here in the Digital Age, companies have discovered that running historical data through analytic engines yields all kinds of valuable trends, patterns, and business intelligence. This greatly increases the retrieval frequency for archived information, in ever larger chunks. To the point where those cloud retrieval fees start to really add up.
To recap: Whatever type or types of cloud you’re using — or want to use or build in the future — the best choice for cloud storage architecture is object storage technology.