NAS vs SAN: Review and Comparison of NAS and SAN Storage Systems

author avatar

26 Sep, 2021

NAS vs SAN: Review and Comparison of NAS and SAN Storage Systems

How Network Attached Storage and Storage Area Network centralize and simplify Storage Management.

This article is a comparison NAS vs SAN, pointing out these differences, limitations and benefits of both architectures.

It is often said that data is the oil of the 21st century. Enterprises are increasingly becoming data-driven, through collecting, storing, managing, and analysing very large volumes of data. In this direction, they are also investing in IT infrastructures for data storage and management. One of the most important requirements for enterprise data storage is to manage storage space centrally, yet in an extensible way, while also sharing the available storage capacity with multiple servers and hosts. Nowadays, there are two storage technologies that provide effective support for this requirement - namely Network Attached Storage (NAS) and Storage Area Networks (SAN) [1]. Despite their similarities, these NAS vs SAN feature important differences as well. Hence, they are both widely used, yet for different types of applications and in different enterprise contexts.

Network Attached Storage (NAS)

What is NAS?

NAS systems are networked connected storage devices, which enable authenticated users to store and retrieve data from a centralized location [1], [2]. You can think of them as data storage devices such as hard drives connected to a computer network like other network hosts. NAS infrastructures can be flexibly and seamlessly extended with new storage capacity: They scale out based on the addition of more storage devices. Moreover, NAS storage devices can be accessed by different servers and hosts, yet by authenticated users only.

NAS Protocols

In NAS systems hosts can use different protocols to access data and files. A non-exhaustive list of these protocols, along with the organizations that are currently in charge of them, is illustrated in the following table. 

Each one of these protocols provides its own advantages and is destined to serve different purposes. As a prominent example, HTTP is very popular among developers’ communities. Hence, it facilitates application developers to access the data of a NAS. On the other hand, SFT ensures reliable data transfers for streaming data. Likewise, UPnP enables intelligent appliances to access the NAS, based on standards-based connectivity services. All the above protocols operate over the TCP/IP protocol stack, which ensures reliable data transfers. Some of the list protocols (e.g., AFS) are nowadays depreciated or very rarely used.

Benefits of NAS

NAS enables users to access storage capacity over the web, which is very useful for many different applications. NAS infrastructures provide several benefits to their users, including:

  • Storage Management Centralization: NAS provides the means for centralizing storage management activities. For instance, backup, maintenance, and administration functions take place through a central point. This obviates the need for individual users to engage in storage management tasks such as backing up and restoring their files.  This is also how the cost of NAS is justified and amortized: It provides convenience to end-users and saves them time and effort. 

  • Broad Accessibility: NAS storage is accessible through various standards-based protocols. Hence it can be accessed by many different hosts regardless of the operating system, including Windows, MacOS and Unix machines. Hence, it enables heterogeneous hosts to share files. 

  • Plug and Play (PnP) Configuration: NAS is installed and configured with minimal effort. Making a NAS operational is simply a matter of connecting the device to a LAN (Local Area Network), much in the same way other devices and hosts are connected to a LAN network. 

  • Low Administration Costs: Along with its PnP nature, NAS does not require significant administration efforts. Specifically, administering NAS devices is usually much easier than administering high-end servers. 

  • Flexibility in Expansion and Procurement: A NAS infrastructure separates the procurement of storage devices from the procurement of servers and hosts. The storage capacity of a NAS can be expanded independently from investments in other hardware and related administrative actions.

Very good Performance: NAS provides good performance as it is accessed like any other device on the LAN. LAN speeds are typically high and suffice for a wide range of applications.

Limitations of NAS

The downside of NAS stems from the limitations of shared services over a LAN network. Specifically, NAS infrastructures have the following disadvantages:

  • Speed and Performance Limitations: In the scope of a NAS deployment file transfer speeds are constrained by the performance of the LAN network. For example, at times where there is heavy use of the NAS storage, the performance of the LAN will drop, and users will experience performance degradation and poor Quality of Service (QoS). This makes NAS unsuitable for data transfer applications. Note also that NAS protocols operate over TCP/IP, which is not suitable for real-time streaming applications. This limits NAS performance for this class of applications as well.

  • User Quotas, yet no Service Guarantees: NAS services operated in a shared environment with many users. Hence, they can hardly provide guarantees for mission-critical applications. Moreover, as NAS is a shared environment, administrators must apply quotas for NAS users. In this way, it is not possible for users to abuse the infrastructure. Nevertheless, quotas guarantee storage space, not latency. 

Nowadays NAS infrastructures are sometimes compared to cloud services. High-quality cloud providers offer greater reliability, improved data protection and an overall better value for money when compared to NAS services.  This is the reason why many companies prefer to access NAS services over a cloud infrastructure (Cloud NAS).

SAN (Storage Area Network)

What is SAN?

SANs have a very prominent position in the networked storage market, as they are one of the prime options for enterprise-scale, mission-critical storage solutions [1], [3]. A SAN is a dedicated network that connects storage devices with servers. SANs are dedicated storage networks [4], which offer gigabit speeds and are not part of the LAN. They are typically geographically distributed and suitable for business-critical applications that require high throughput, low latency, and increased reliability. To ensure their high performance, many SANs use all-flash storage rather than conventional hard disks. Likewise, SANs’ exception reliability stems from that they have no single point of failure i.e., they are designed to sustain device failures.  

 SANs have shared storage infrastructures that enable organizations to centralize storage management operations such as backup, data security, data protection, and disaster recovery. They are based on high-performance architectures that connect servers to logical storage blocks. The latter are provisioned over the storage infrastructure and presented to the user as a unified block of storage (i.e., like a disk drive). Servers are in charge of partitioning and formatting these blocks. In the end, servers view and access distributed logical storage like conventional local disks.  

SAN Protocols

SANs have their own protocols that allow servers to access storage devices. The most popular protocols are illustrated in the following table.

Protocol

Overview

Fibre Channel Protocol (FCP)

High-Speed Data Transfer with Lossless Delivery of Raw Data Blocks, leveraging Fibre Channel Interfaces

Internet Small Computer System Interface (iSCSI)

IP protocol for block-level access to storage devices based on SCSI commands over TCP/IP

Fibre Channel over Ethernet (FCoE)

Encapsulates Fibre Channel packets/frames over the (Gigabit) Ethernet Protocol

Non-Volatile Memory Express over Fibre Channel (FC-NVMe)

High-Performance Protocol (based on Fibre Channel) for accessing flash storage via a PCI (Peripheral Component Interconnect) bus

As evident from the table, SAN protocols emphasize block storage access rather than access to entire files. FCP is dominant in the market, while iSCSI and FCoE hold smaller, yet remarkable shares as well (i.e., in the range of 5%-15%). 

Benefits of SAN

SAN’s prominent position in the market of enterprise-scale storage infrastructures is largely due to its strong security, reliability, scalability, and high performance. Specifically, the main benefits of a SAN infrastructure include:

  • Quality of Service: SANs provide a compelling QoS proposition as they do not suffer from the bandwidth limitations of LAN-based storage infrastructures. They are offered dedicated bandwidth rather than sharing LAN bandwidth. This reduces bottlenecks and leads to improved QoS.

  • Security: SAN is separated from the LAN, which makes it less susceptible to security attacks from adversaries that gain access to the LAN. Furthermore, SAN infrastructures deploy data protection techniques (e.g., encryption techniques) which increase their overall cyber-resilience. 

  • Scalability: SAN networks are very scalable: They can grow in capacity as required by the business requirements at hand. Most importantly, capacity upgrades in SANs are quick, seamless and require a fair administration overhead.

  • Reliability: SAN infrastructures are reliable as they can replicate data across different geographical locations. As such, they enable effective disaster recovery in cases of disastrous situations like major security attacks, terroristic attacks, or even natural disasters. This is what makes SANs an asset when it comes to guaranteeing business continuity.

As already outlined SAN centralizes storage management operations like backup and recovery. As already discussed, this saves time for the end-users and boosts their productivity.

Limitation of SAN

The compelling features of SANs come at a relatively high cost. Therefore, businesses must assess whether they need and can afford the establishment of a SAN infrastructure in their data centre. Likewise, SANs scale very well, yet they are not cost-effective when deployed with a few servers only. Hence, the scale of the SAN deployment must be another important consideration for enterprises considering the adoption of a SAN solution. Overall, SANs provide more advantages than limitations, yet these require a considerable investment with a long term outlook.

NAS vs SAN: Understanding the Differences 

Both SAN and SAN provide the means for centralizing storage operations while enabling different hosts to access them. However, they have significant differences that make them suitable for different types of business requirements and applications. Specifically, the main differences between the two technologies are as follows:

  • Protocol Differences: As evident in the previous tables of NAS and SAN protocols, SANs provide a block-level interface to storage, while NAS offers file access interfaces.  Also, NAS operates within a LAN and is based on Ethernet, while SAN infrastructures use either Ethernet or Fibre Channel interfaces.

  • Performance Differences: SAN offers better performance than NAS, as it is not limited by LAN bottlenecks. This makes SAN more appropriate for mission-critical applications that demand low-latency and high performance.

  • Different Target Applications: Based on the above-listed features and comparisons, NAS infrastructures are best suited for file access applications, while SANs are more suitable for data transfer intensive applications. Moreover, SANs are more efficient when it comes to storing very large volumes of data as block storage.

  • Easy of Configuration and Use:  NAS is a “simpler” technology than SAN. It, therefore, offers its users greater simplicity in terms of managing the infrastructure, as well as data access and use.

  • Total Cost of Ownership (TCO) Differences: The two technologies feature different TCOs, with SANs being more expensive than NAS. Nevertheless, the overall value for money depends on the target business requirements. In many cases (e.g., data transfer intensive applications), NAS is simply not enough, and the more expensive SAN is the only viable option.

In the era of Big Data and the Internet of Things (IoT) enterprises need scalable and cost-effective storage infrastructures, which will enable them to seamlessly manage their ever-growing volumes of structured, unstructured, and semi-structured data. In this direction, NAS and SAN provide two very good options that vary in scale, functionality, QoS and overall TCO. While both infrastructures can be deployed as part of on-premise infrastructure, the trend for Big Data and IoT data management is to manage data in the cloud. Nowadays, SANs are integral components of cloud data management in the data centres of Internet Service Providers, Application Service Providers, Storage Service Providers and other enterprises offering enterprise-class cloud services. Thanks to their Gigabit speeds, SANs enable demanding cloud applications such as video editing and live multimedia applications. Coupled with the emerging 5G networking infrastructures, these data centres will also enable ultra-low latency applications such as tactile applications based on Augmented Reality and Mixed Reality. On the other hand, NAS solutions are still attractive when it comes to supporting LAN-based storage at a smaller scale. To resolve the NAS vs. SAN dilemma, enterprises must carefully review their short and medium-term storage requirements, including the nature of applications they must support.


References

  1. “Information Storage and Management: Storing, Managing, and Protecting Digital Information in Classic, Virtualized, and Cloud Environments”, 2nd Edition, EMC Education Services, ISBN: 978-1-118-23696-3 April 2012 (528 Pages).

  2. Yasud Y, Kawarnoto S, et al. Concept and evaluation of X-NAS: a highly scalable NAS system[R]. Mass Storage Systems and Technologies, 2003. (MSST 2003). Proceedings. 20th IEEE/11th NAC Goddard Conference on. 2003, pp.219-227.

  3. I. Redbooks and S. Networking, "IBM Redbooks - Introduction to Storage Area Networks", Redbooks.ibm.com, 2018, [online] Available: http://www.redbooks.ibm.com/abstracts/sg245470.html.

  4. Y. Peng, Y. Zhu and J. Luo, "A network storage framework based on SAN," 2012 7th International Conference on Computer Science & Education (ICCSE), 2012, pp. 818-821, doi: 10.1109/ICCSE.2012.6295196.

More by John Soldatos

John Soldatos holds a PhD in Electrical & Computer Engineering from the National Technical University of Athens (2000) and is currently Honorary Research Fellow at the University of Glasgow, UK (2014-present). He was Associate Professor and Head of the Internet of Things (IoT) Group at the Athens In...

Wevolver 2022