The difference between NAS, San, and Das

Currently on the disk storage market, the storage classification (table one below) is divided according to the server type: Closed system storage and open systems storage, closed system mainly refers to mainframe, AS400 and other servers, open systems are based on servers including Windows, UNIX, Linux and other operating systems Open System storage is divided into: built-in storage and external storage, open system plug-in storage is divided into: direct-attached storage (direct-attached Storage, abbreviated DAS) and networked storage (fabric-attached Storage, (FAS); networked storage of open systems is also divided into: Network access storage (network-attached Storage, short NAS) and Storage Area network (Storage, or SAN) according to the transport protocol. As the majority of users are using open systems, the external storage occupies more than 70% of the current disk storage market, so this article mainly for open system plug-in storage to discuss the description.

Table I:

Today's storage solutions are mainly: direct-attached storage (DAS), storage Area Network (SAN), Network access Storage (NAS). As shown in Table II below:

Direct-attached storage for open systems (direct-attached Storage, or DAS) has been used for nearly 40 years, and as user data continues to grow, especially over hundreds of GB, its backup, recovery, scaling, and disaster preparedness issues are becoming increasingly troubling to system administrators.

The main issues and deficiencies are:

Direct-attached storage relies on the server host operating system for data IO Read and write and storage maintenance management, data backup and recovery requires server host resources (including CPU, system IO, etc.), the data flow needs to return the host to the server attached to the tape drive (library), data backup usually occupies the server host resources 20-30 , so many enterprise users ' daily data backups often take place late at night or when the business system is not busy, so as not to affect the operation of normal business systems. The larger the amount of data that is stored directly, the longer the backup and recovery time, and the greater the dependency and impact on the server hardware.

The connection channel between direct-attached storage and server host usually uses SCSI connection, bandwidth is 10mb/s, 20mb/s, 40mb/s, 80mb/s, etc., as the server CPU processing power is more and more strong, the storage hard disk space is more and more, the number of hard disks in the array is more and more. SCSI channel will become an IO bottleneck, the server host SCSI ID resource is limited, can establish a limited number of SCSI channel connections.

Whether direct-attached or server-host extensions, a cluster (Cluster) that extends from one server to multiple servers, or an expansion of storage array capacity, can lead to business system downtime, resulting in economic losses for businesses and critical business systems for 7x24-hour services in banking, telecommunications, media and other industries , it is unacceptable. and direct-attached storage or server host upgrade extension, can only be provided by the original equipment manufacturers, often by the original equipment manufacturers Limited.

Storage Area networks (Storage Region network, or SAN) use Fibre Channel (Fibre Channels) technology to connect storage arrays and server hosts through Fibre Channel switches and to establish regional networks dedicated to data storage. San has matured over a more than 10-year history to become the industry's de facto standard (but each vendor's fiber-switching technology is not exactly the same, and its server and SAN storage have compatibility requirements). SAN storage uses the bandwidth from 100mb/s, 200mb/s, to the current 1Gbps, 2Gbps.

Network access Storage (network-attached Storage, referred to as NAS) uses network (TCP/IP, ATM, FDDI) technology to connect storage systems and server hosts through network switches, and to establish a storage private network dedicated to data storage. With the development of IP network technology, network access Storage (NAS) technology has undergone a qualitative leap. The 10Mbps bandwidth of the late 80 to the early 90, network access storage as a file server storage, performance is affected by bandwidth, and later the emergence of Fast Ethernet (100Mbps), VLAN virtual network, Trunk (Ethernet Channel) Ethernet channel, The read and write performance of the network access storage is improved, and the advent and commercialization of Gigabit Ethernet (1000Mbps) in 1998 has brought about a qualitative change in network access storage (NAS) and wide market recognition. Because the network access storage uses the TCP/IP network for data exchange, TCP/IP is the standard protocol of the IT industry, and the products of different vendors (servers, switches, NAS storage) can realize interconnection and interoperability without the requirement of compatibility as long as they meet the protocol standards. , and the Advent and commercialization of 2002 Gigabit Ethernet (10000Mbps), storage network bandwidth will greatly improve the performance of NAS storage. The strong demand for NAS has become a reality. First, the NAS almost inherits all the advantages of the disk array, and can connect the device through a standard network topology, freeing itself from the shackles of servers and isomerization architectures.

Second, in the rapid expansion of enterprise data volume, sans, large-scale tape libraries, disk cabinets and other products are good storage solutions, but their noble identity and complex operation is the financial and technical strength of small and medium-sized enterprises can not accept anyway. NAS is the product that meets this demand, providing a high price/performance ratio while addressing enough storage and expansion space. As a result, Nas is naturally the best choice for most businesses, especially large and medium-sized enterprises, both in terms of suitability and TCO.

Analysis and comparison of NAS and SAN

For I/O is an inefficient bottleneck of the entire network system, experts have proposed a number of solutions. One of the most effective ways to seize the crux of the problem and practice testing is to simplify storage management by separating the data from the general application server.

Problem:

Figure 1

Figure 1 shows the original problem: each new application server must have its own storage. This results in complex data processing, and as the application server continues to increase, the efficiency of the network system will drop sharply.

Workaround:

Figure 2

As you can see from Figure 2, the storage is separated from the application server for centralized management. This is what is called the Storage Network (Storage Networks).

Benefits of using the storage network:

Unity: The form of scattered God is not scattered, in the logic is completely integrated.

Centralize data management because they are the true lifeblood of your business.

Easy to expand, that is, very strong shrinkage.

Fault-tolerant, the entire network without a single point of failure.

Experts have taken two different approaches to this approach, namely, NAS (Network attached Storage) networked access storage and SAN (Storage area Networks) storage Zone networks.

NAS: Users access data through the TCP/IP protocol, using industry-standard file sharing protocols such as: NFS, HTTP, cifs for sharing.

SAN: Access Data via a dedicated Fibre Channel switch with SCSI, Fc-al interface.

What are the fundamental differences between NAS and San?

The most essential difference between Nas and SAN is where the file management system is.

Figure 3

As can be seen from Figure 3, the File Management system (FS) in the San architecture is also on each application server, while the NAS is the same file management system used by each application server through a network sharing protocol (for example, NFS, CIFS). In other words: the difference between NAS and SAN storage systems is that NAS has its own file system management.

NAS is focused on applications, users and files, and the data they share. Sans are a reliable infrastructure for focusing on disks, tapes, and joining them. In the future, a comprehensive solution from desktop to data centralized management to storage devices will be NAS plus SAN.

The integration of NAS and SAN is also a trend for storage devices, such as EMC's new product VNX series. The diagram on the right shows the logical structure of a NAS: a double dashed box represents a NAS. It obtains storage space from the back-end San via the Fibre Channel, creates the file system, and then shares it to the server via Ethernet. The SAN provides storage units that are LUNs and are at the block level. After the NAS has been created as a file system, it becomes a file-level.

If the logical diagram above is not clear enough, you can look at the following physical connection. The NAS is connected to the San via FC switch and the application server is connected to the NAS via the Ethernet switch. The San also provides block-level storage directly to the application server.

The difference between NAS, San, and Das