As the global shift toward remote and asynchronous work continues to increase, so do our data storage needs. So it’s no surprise that storage has become fundamental to cloud providers, and offers a relatively easy way for companies to get started with cloud.
AWS vs Azure vs GCP: Cloud provider comparisons
It’s not all apples to apples in the world of cloud. Here are some other cloud comparisons guides.
- AWS vs Azure vs GCP cloud comparison: Databases
- IAM: Comparing AWS, Azure, and Google Cloud IAM services
- Serverless: AWS Lambda vs Azure Functions vs Google Cloud Functions
- Virtual Machines: AWS EC2 vs Azure Virtual Machines vs Google Compute Engine
- AKS vs EKS vs GKE: Managed Kubernetes services compared
What is data storage?
Data can be stored on paper, disk, magnetic tape, optical disc … even in the DNA of bacteria. But warehousing all this data takes up space and costs money, and some data storage media isn’t really designed for the long term. Enter the cloud.
Cloud providers have figured out how to create a variety of storage methods at scale, resulting in high reliability for low cost. You pay for what you use, and for the performance and features that fit your needs.
The most common forms of data storage are object storage, archive storage, file storage, and block storage, and each type is fit for different purposes
Object storage compared: Amazon S3, Azure Blob Storage, Google Cloud Storage
When it comes to object storage, think of an object as a file. Your impressive collection of dad joke memes are objects, and they can be stored in the cloud on object storage services.
AWS calls their object storage service Amazon Simple Storage Service, or S3 for short.
GCP has the aptly named Google Cloud Storage.
And Azure ups the geek quota by calling their service Azure Blob Storage (BLOB means “binary large object”). Both AWS and GCP call the things you put your data into a “bucket,” while Azure calls it a “container.”
These storage services have plenty in common. They all offer:
- Encryption at rest
- Fine-grained security, including options to make an object publicly accessible or completely private
S3, Cloud Storage, and Blob Storage all offer storage class tiers — the more performant and redundant the storage class, the more you pay. There are also options to reduce costs for less frequently accessed data. These storage classes are called:
- Amazon S3 Infrequent Access (also note the new S3 Glacier Instant Retrieval)
- Azure Blob Storage Cool Tier
- GCP Nearline Storage and Coldline Storage
Object storage archiving
Many organizations are required to hold onto data for a certain amount of time for regulatory or legal reasons. This can become expensive. Thankfully, the cloud providers have their own low-cost archiving solutions.
Obligated to delete data after a required retention period has collapsed? Fortunately, the cloud providers also have storage lifecycle management built right into these solutions, and can automatically delete expired objects without you having to do this manually. Magic!
File storage compared: Amazon EFS, Azure Files, Google Cloud Filestore
If you’re looking for high-performance storage, you’ll want to choose a solution that has a high number of input/output operations per second, or IOPS. This is where file storage can come in. Think of it a bit like the shared drive on your network. Common file storage services will provide Network File System (NFS) or Server Message Block (SMB) access, allowing other systems to attach over the network and use that storage as if it were a local volume.
- AWS calls their file storage service Amazon Elastic File System (Amazon EFS).
- Azure calls theirs Files.
- And GCP has Filestore.
As with object storage, these services are pretty much the same, with some slight differences in protocol, maximum volume size, and performance.
AWS does outdo the other providers with two additional file storage options, though. Amazon FSX for Windows is optimized for use in Windows server environments, and Amazon FSx for Lustre for high-performance computing use cases.
Block storage compared: Amazon EBS, Azure Managed Disks, Google Persistent Disks and Solid State Disks
If you have the need — the need for speed — block storage is where it’s at. Think of block storage as a raw hard drive. It’s going to provide the highest IOPS.
All three cloud providers have block storage options, because they’re pretty much required to run virtual machines. Block storage is what lets our virtual machines boot into their operating system.
- AWS calls their block storage service Amazon Elastic Block Store (Amazon EBS).
- Azure has managed disks.
- GCP has persistent disks and solid state disks.
All offer similar features, allowing you to provision a very fast SSD device, or save some money with more traditional magnetic hard drives. Scheduled and ad hoc snapshots allow you to pack up, restore, and duplicate your block storage devices. And encryption of block storage is mouse-click easy across all providers.
All this cloud talk is well and good, but how do you actually get all your data into the cloud in the first place?
Well, the most basic way is to simply upload your data using the web UI or command-line interface (CLI) from your cloud provider of choice. AWS and Azure also offer downloadable virtual appliances that can be installed in your own data center to act as transfer gateways, syncing data to the cloud in the background.
But what if you have petabytes of data stashed in your data center? It would take years to transfer all that data, even with the fastest internet connection. And that, my friends, is why you can never underestimate the bandwidth of a delivery truck.
All the major cloud providers let you ship your data via post using hard drives.
- Azure Data Box Disk is basically a box full of SSD drives with a USB 3.0 port. Fill up the disks with up to 40 TB of data, ship it back to Microsoft, and your data is loaded into your storage account. Azure also has a larger version capable of up to 1 PB, called Azure Databox Heavy, which kind of resembles one of the roadie cases lugged around on an Iron Maiden tour.
- GCP has a similar hardware device, called Transfer Appliance, that can also accommodate up to 1 PB in a round trip. But the current king of data import is AWS with their Snow Family. The AWS Snowcone and AWS Snowball devices both offer data transport and include compute hardware, which allows them to perform like small AWS data centers.
- Then there’s AWS Snowmobile, described as a “45-foot long ruggedized shipping container, pulled by a semi-trailer truck”. It will drive to your data center, plug in, and ingest up to 100 PB of data at 240 gigabits per second!
Which is better: AWS, Azure, or GCP?
We’re going to be neutral here and say the storage options across AWS, Azure, and GCP are all pretty similar. And that’s to be expected. Storage has largely become a commodity, with providers focusing on improving performance and reducing costs.