What is cloud storage?

How cloud storage works

In the past, enterprise storage was defined by physical constraints. It meant managing hard disks on servers. You had to predict the space you needed in advance and replace hardware before it wore out. If you ran out of room for new servers or drives, you risked seeing your systems grind to a halt. But the cloud changed this model.

When you connect to the cloud, it looks like a single disk. But storage services are made up of networked servers running in data centers. Software takes these servers and makes them look like that one disk.

Cloud storage architecture

Cloud storage divides data into three layers:

1. Storage nodes

This is the hardware the storage sits on. If a client needs high-speed storage, they get fast solid-state disks. For applications that need space more than speed, they have slower hard disks.

2. Storage approaches

This layer controls how applications interact with the storage.

• Objects: Data is individual units with unique names. Objects contain raw data, metadata, and a unique identifier. The identifier combines the storage instance, namespace information, and the object name.

• Blocks: Arranges data in raw, fixed-sized volumes of blocks. Each block has an address but, like a physical disk, no metadata. Block storage acts like a disk attached to a computer.

• Files: Contents look like a directory tree on a computer. It has folders and files, with features that allow more than one user to read and write to the files at a time.

3. Metadata services

This layer divides a file's info, like name, attributes, and permissions, from its contents. The system stores it in fast, distributed databases so clients have access to it as quickly as the files.

Access protocols

Each storage type uses a different approach for talking to clients:

• Object storage uses Representational State Transfer (REST), a protocol for managing data on a server. The API for Amazon's S3 file service serves as a standard for these services.

• File storage uses standard operating system protocols. For Linux, there's the Network File System (NFS). For Windows and Mac, there's Server Message Block (SMB), the standard for Windows servers.

• Block storage uses the protocols required for connecting servers directly to disk hardware. They allow the systems to talk to the cloud storage with hardware protocols.

These are three different ways to use cloud storage, but they share the underlying system. So, the data they manage has a common life cycle:

• Login and access: Every message from the client has a cryptographic signature. It identifies the sender so the service knows what they are allowed to do.

• Ingestion and chunking: Requests to a cloud service can be large. So, the gateway removes the extra header information and splits the data into pieces.

• Encryption and striping: Most cloud storage systems encrypt data before writing it to disks. Then it distributes it across different systems to spread the risk.

• Acknowledgment: First, the system writes the data to disk and verifies that it's sound. Then it sends the okay to the client.

Durability and redundancy

Cloud storage is reliable because providers use sophisticated distribution and failure-recovery mechanisms.

• Multi-zone replication: Cloud storage duplicates writes between physical facilities. These “availability zones” have their own power, HVAC, and networking resources. So, if a center fails, data remains safe, online, and accessible.

• Object versioning: Cloud storage uses versions to protect clients from deletion and ransomware. It creates read-only versions for every change a client makes to files.

What are the advantages of cloud storage?

Cloud storage offers significant advantages over on-premises storage solutions.

1. Scalability

On-premises storage requires capacity planning. If you overestimate, you end up with costly hardware sitting idle. If you underestimate, you risk problems down the road.

Cloud storage removes this risk. You can scale disk sizes up and down on demand. So, you can let your storage grow organically or adjust it on the fly.

2. Cost shift: CAPEX to OPEX

3. High availability and built-in disaster recovery

Maintaining a system means buying and building everything at least twice for redundancy. And you need to maintain complex failover setups and processes. Cloud storage builds this into its product.

4. Global access and collaboration

Large companies need the ability to access files from anywhere. Cloud storage makes this possible: it's what "cloud" means. With local storage, you need to do it yourself. This means accepting the security risks and overhead.

Cloud storage security, compliance, and data governance

Cloud storage comes with many benefits, but at the same time, it brings a new set of security concerns. You can access its data from nearly everywhere. So, it relies on strong encryption and access control. It also needs to meet your company's regulatory requirements.

Core data protection controls

Cloud storage platforms use several tools to secure data:

1. Encryption in transit: Keeping data secure starts with encrypting it on the network. The cloud service encrypts data between it and the client, as well as inside its system.

2. Encryption at rest: The system also encrypts the data before writing it to disk. Providers typically support three ways to manage the encryption keys:

   • Provider managed: the provider handles key generation, rotation, and storage.

   • Customer managed: the customer controls the key with the provider's management system.

   • Customer supplied: the customer generates the keys offsite. This means the provider can't read the file. But responsibility for maintaining keys is on the client.

3. Network isolation: By restricting access to networks that belong to the client, cloud providers keep data safe from attacks over the public internet.

Regulatory compliance and data sovereignty

Legal and regulatory mandates are part of modern business. Cloud providers need the right tools for these processes.

• Data location and laws: Cloud applications work in more than one region. They can also allow clients to restrict where data resides when required by law, such as with GDPR in the EU.

• Read-only storage: In some sectors, clients must follow data retention periods. This data must be kept intact and available, which calls for read-only storage.

• Audit logging: Cloud storage services log API transactions. These logs are audit trails. They identify data and time, IP address, and the actions of the request.

Identity and access management

Identity is the primary security factor in cloud storage.

• Role-based access control (RBAC): This assigns permissions based on a user's role. This is often a better approach than using identity.

• Attribute-based access control (ABAC): This uses data about the request, like department, time, and client location.

• Granular resource policies: Controls applied directly to buckets or volumes. They enforce the conditions governing access to the contents.

Cloud storage is critical infrastructure

Cloud storage has become a foundational component of modern IT strategy, enabling organizations to scale efficiently, adapt quickly, and support growing data demands without the limitations of traditional infrastructure. By reducing the need for large upfront hardware investments and simplifying storage management, it helps improve operational resilience, business continuity, and overall agility.

As organizations continue to modernize their environments and support distributed users and applications, cloud storage provides the flexibility and reliability needed to keep systems accessible, secure, and ready for future growth.

Common questions about cloud storage