SAN (Storage Area Network) and NAS (Network Attached Storage) are two common types of enterprise-level network storage solutions. While they both provide ways to consolidate storage and share it across multiple servers and users, there are some key differences between the two.
What is SAN?
SAN or Storage Area Network refers to a dedicated high-speed network used for block-level storage. SAN typically consists of:
- Storage devices such as disk arrays, tape libraries, etc.
- SAN switches and directors that provide network connectivity.
- Host bus adapters (HBAs) in servers to connect them to the SAN.
In a SAN environment, the storage devices themselves are not connected directly to the servers. Instead, they are networked over a high-speed infrastructure, usually Fibre Channel or iSCSI. Servers access block-level storage over this network.
SAN is best suited for scenarios that require high performance – such as large databases, virtualized environments, email servers, etc. The dedicated network provides high bandwidth and low latency for block storage traffic.
Key characteristics of SAN
- Block-level storage access – SAN provides servers with block-level access to storage devices. This allows servers to read and write blocks of data to and from the storage.
- Dedicated network – The storage devices and servers are interconnected via a dedicated high-speed network such as Fibre Channel, iSCSI, Infiniband, etc. This ensures high throughput and low latency.
- Scalability – SANs are highly scalable and can support hundreds of servers and storage devices.SAN switches and directors provide the ability to scale out the infrastructure.
- High availability – SAN provides redundancy features such as multipath I/O and clustering that ensure high availability of mission critical data.
- Advanced storage features – Enterprise SAN storage systems provide advanced features like thin provisioning, deduplication, replication, snapshots and backup.
What is NAS?
NAS or Network Attached Storage refers to file-level storage that is connected to a network and can be accessed by multiple clients. NAS storage devices typically consist of the following components:
- NAS head or filer that acts as an interface between the storage and network.
- Hard drives for storage capacity.
- Built-in network interface such as Ethernet to enable network connectivity.
- Operating system and file system to manage the storage and file access.
NAS uses file access protocols like SMB, NFS, FTP or HTTP to allow clients to access files over the network. The storage appears as file shares that can be mapped and accessed like local storage.
NAS systems are typically used for file sharing, backup, archiving and lighter performance needs compared to SAN.
Key characteristics of NAS
- File-level access – NAS allows clients to access data as files over standard file sharing protocols. Does not provide block-level access.
- Less overhead – NAS does not require dedicated cabling or HBAs. It utilizes the existing Ethernet network.
- Ease of use – NAS storage can be easily deployed by connecting it to the LAN and mapping network drives.
- Scalability – NAS scales out by adding multiple NAS heads into the infrastructure. However, a single device has relatively limited capacity.
- Fewer advanced features – NAS may lack some of the advanced storage functionality supported by enterprise SAN storage.
How SAN and NAS access storage
The fundamental difference between SAN and NAS is in how they provide access to storage for clients:
- SAN – Provides block-level access over Fibre Channel, iSCSI, etc. Clients can read/write blocks of data.
- NAS – Provides file-level access over SMB, NFS, FTP. Clients access data in files and folders.
This underlying difference in access methods also impacts performance, scalability, features and usage scenarios supported by the two technologies.
SAN vs NAS Comparison
Here is a detailed feature comparison between SAN and NAS:
SAN | NAS | |
---|---|---|
Access method | Block-level | File-level |
Protocols | Fibre Channel, iSCSI, FCoE, Infiniband | NFS, SMB/CIFS, FTP, HTTP |
Performance | Very high | Moderate |
Latency | Very low | Higher than SAN |
Scalability | Massive (into petabytes) | Limited by single device |
Network | Dedicated (Fibre Channel, Infiniband) | Standard Ethernet |
Advanced features | Very extensive (snapshots, replication, deduplication, encryption, compression, QoS, thin provisioning etc.) | Moderate features |
Redundancy | Highly redundant with multipath I/O, dual controllers etc. | Mostly single controller |
Ideal for | Transactional applications, large databases, virtualized environments | File sharing, backup, archiving |
As seen from the table, SAN has significant advantages in performance, scalability, advanced features while NAS is easier to use for simple file sharing needs.
When to use SAN?
Following are typical use cases where a SAN provides the right level of performance, scalability and availability:
- Transactional databases – Databases like Microsoft SQL Server, Oracle, MySQL, PostgreSQL etc. that have high storage I/O and latency requirements.
- Virtualized servers – Hosting large number of virtual machines that require shared block storage with features like thin provisioning and snapshots.
- Email servers – High performance storage for large email servers and repositories.
- ERP systems – To support demanding Enterprise Resource Planning (ERP) systems like SAP.
- Big data analytics – Providing primary storage for big data deployments like Apache Hadoop.
- Collaboration apps – For collaboration tools like Microsoft Sharepoint that rely on fast shared storage.
In essence, any application that requires very high performance, low latency, scalability and advanced storage features can benefit from a SAN deployment.
When to use NAS?
Here are some typical scenarios where NAS provides a good fit:
- File sharing – Simple and centralized file sharing across an organization.
- Home directories – User home folders in large enterprises.
- Backups – Disk target for backups in place of tapes.
- Archiving – Cost effective archival storage for older low access data.
- Web hosting – Storing website files and content.
- Media streaming – Video surveillance storage, media streaming from NAS.
To summarize, NAS provides good enough performance for scenarios involving file sharing, backups, archiving and lighter workloads. The convenience of file access over NAS makes it suitable for mixed workload consolidation.
SAN and NAS together
While SAN and NAS are different technologies, they can complement each other in larger environments:
- Use SAN for mission critical applications that need performance.
- Use NAS for file sharing, backup, archiving purposes.
- Tier storage between the two for optimal usage of high performance SAN storage.
- Replicate data between SAN and NAS for backup and disaster recovery.
This “best of both worlds” approach allows organizations to benefit from the strengths of both SAN and NAS under a unified storage environment.
Virtual SAN
A virtual SAN or vSAN combines storage from multiple hosts into a single virtual SAN using software. It allows creating a SAN using local disks and flash storage across virtualized servers.
Key characteristics of a virtual SAN:
- Provides block storage like a physical SAN.
- Leverages direct attached storage in virtualized servers.
- Creates a pool of storage across hosts.
- Provides redundancy through mirroring and erasure coding.
- Supports advanced features like snapshots, cloning, encryption, deduplication etc.
- Lower cost compared to physical SAN.
Virtual SAN solutions like VMware vSAN allow smaller organizations to benefit from SAN capabilities through a software-defined approach.
Hyperconverged Infrastructure (HCI)
Hyperconverged infrastructure or HCI takes convergence up a notch by running compute, storage, networking and virtualization in a single software-defined system on commodity hardware. A hyperconverged cluster integrates:
- Compute (virtualized servers)
- Software defined storage (SDS)
- Virtualized networking
- Management software
The SDS layer in HCI may provide either block storage like SAN or file storage like NAS. For example, Nutanix HCI uses a virtual SAN to provide shared block storage for VMs. On the other hand, Scale Computing HC3 uses file based storage for VMs rather than SAN storage.
HCI builds on top of virtual SAN capabilities and makes storage management seamless by integrating it with other data center resources under unified management.
Conclusion
SAN and NAS provide two distinct methods of networked storage access – block versus file. SAN is suited for high performance needs while NAS works well for basic file sharing.
SAN strengths lie in performance, scalability and advanced storage features. NAS is easy to use and deploy for file-based workloads. Organizations typically deploy them together to benefit from both technologies.
Virtual SAN and HCI solutions are blurring the lines between SAN and NAS. They provide the best of both through a software-defined approach on commodity hardware.