Transfer of Information (TOI) : Understanding the three primary storage technologies DAS, NAS and SAN and their

architecture in two part series. Objective : By the end of this knowledge transfer, audience will understand the three primary storage options, but also be able to appreciate the differences among the three primary technologies.

Duration : 1 1 hrs. Audience : Anyone who is interested.

Pre-requisite : No prior technical knowledge required, an average IT person with basic computer and storage information is good enough for this TOI.

Prepared by : Ashwin Pawar

Email [email protected] , www.ashwinpawar.blogspot.com

Part 1: Introduction Data is unquestionably the lifeline in today's digital organization. Storage solutions remain a top priority in IT budgets precisely because the integrity, availability and protection of data are vital to business productivity and success. Several storage options exist today to support various needs of the small, medium and enterprise level businesses. The three primary options are :

Direct Attached Storage (DAS) Network Attached Storage (NAS) Storage Area Networks (SAN)

Reader is encouraged to research these areas further and point to the author for any technical or typo error you come across in this TOI.

Direct Attached Storage (DAS) : Direct Attached Storage (DAS), as the name suggests, involves the direct connection of servers to storage. This can either be with the use of an internal server disk controller with either internal or external drives, or with a completely external storage subsystem that includes the controller and disks. The main characteristic of DAS is that the storage connection from the server to the storage is hard-connected. This connection is typically through a direct SCSI cabled connection (Figures 1 and 2).

Figure 1:

Figure 2:

Advantage: DAS systems are recognized for their ease of management, generally lower costs, and overall simplicity.

Disadvantage: They are often characterized as creating "islands of information," because each DAS device is isolated from all others. In a small business environment, that isolation can be only a minor problem, but in a larger business, data accessibility by diverse groups of users becomes a business critical activity.

Network Attached Storage (NAS) : As the name suggests, its the storage attached to the network, by network it means the common Ethernet network. A NAS is essentially a file server storage that often integrates an optimized operating system dedicated to file sharing. By, optimized OS means, It is designed to serve files to multi-protocol and multi platforms. All the processing is done locally on the NAS whereas client only demands it. Its basically a cooked system, something that can be used out of the box. One major difference with NAS is that it provides file-level I/O via traditional CIFS and NFS network file shares, while DAS and SAN provide block-level I/O. This is important to consider for performance since there can be performance penalties associated with the use of file-level I/O with a high-performance storage solution. NAS devices are fairly easy to implement and offer storage consolidation and file sharing of data over a standard Ethernet network.

Figure 3: Depicts a typical NAS configuration:

Advantage: One unique feature of a NAS is that it solves the CIFS/NFS interoperability problem by allowing either Windows or UNIX clients to access its file shares without the need for additional software.

Disadvantage: One major difference with NAS is that it provides file-level I/O via traditional CIFS and NFS network file shares, while DAS and SAN provide block-level I/O. This is important to consider for performance since there can be performance penalties associated with the use of file-level I/O with a high-performance storage solution.

Vendors selling NAS solutions : There are many vendors in the market selling NAS solutions, like HP, Netapp , EMC, IBM, DELL etc, and all of them have their own optimized OS and file system to support different clients.

Storage Area Networks (SAN) : Storage Area Networks (SAN) are, as the name implies, back-end storage networks that connect multiple hosts through a switched fabric such as Fibre Channel or iSCSI. A SAN provides flexibility for "carving" out storage for multiple servers where the servers can be spread out across a data center. A typical SAN (As shown in Figure 3 below) is configured with multiple switches and multiple server Host Bus Adapters (HBA) to create a high-availability storage configuration and various RAID solutions are used to protect data at the disk level. Figure 4: A typical SAN

Building a SAN is more expensive than DAS or NAS and requires expertise with specific hardware and software used to configure the SAN. A SAN should be considered when supporting many servers as part of an overall data center storage design concept. The real strength of a SAN is that storage can be assigned and later reassigned as needed to support the changing needs of specific servers. This results in the efficient use of storage and minimizes unused storage capacity for a given server.

Advantage: Like NAS systems, SAN systems allow for better resource capacity utilization by sharing server growth space. SANs also allow servers to boot directly from the SAN. This allows new servers to access LUNs on faulty servers, thereby restoring system functionality without losing data.

Disadvantage: SANs can, however, be more expensive because they require the creation of a fibre channels to transfer data between end users and the SAN.

Basic differences in these three (DAS, NAS & SAN) technologies : Basically, a SAN does block I/O just like having a disk directly attached to a server (DAS) . A NAS is really remote file system I/O where the file request is redirected over a network to a device (really a processing entity with its own file system) where the file I/O is actually performed. Following pictures 5 & 6 will give you a general idea of how the I/O is done between block (SAN) and remote file I/O (NAS): Figure 5: NAS (File I/O based)

Figure 6: SAN (Block I/O based)

Conclusion: Deciding whether to use a SAN vs. NAS has many factors to consider so there's not a specific answer. In general if your application requires block I/O or there is a significant performance requirement, use a SAN. If it's file based I/O for the application or you need to share files and you want simple administration, use NAS.

Part 2:

Databases on NAS vs. SAN

Lets take deep dive and see how NAS and SAN works architecture wise, NAS technology has a place in the IT organization for general file sharing applications, using industry standard protocols such as NFS over traditional networks. Lets us now examine how it works as a database access model in Figure 7 below: Understanding the architecture of Network Attached Storage (NAS) Database access model :

Figure 7:

Anatomy of a NAS Database request: 1 Client makes data request over network to database server. 2 Database server performs I/O to Network Attached Server (NAS) over network. 3 NAS performs I/O to disks over I/O bus (SCSI, IDE, SSA, Fibre Channel). 4 NAS returns I/O to Database server over regular LAN or Network. 5 Database server sends results back to client over network.

Advantages of this access method: 1 Storage sharing among two or more host systems is possible. 2 Storage can be physically located further away from host systems. 3 Storage can be brought online and reconfigured dynamically. 4 Backups can be simplified. Disadvantages of this access method: 1 Extra workload is placed on LAN network. 2 Networks may need upgrading to support extra I/O activity. 3 Extra steps are involved in performing I/O operations. 4 There is a lack of mirrored cache for end-end data protection. 5 There is no direct or server attach storage capability exists. 6 There is no support for RAW volume partitions for fast I/O. 7 There is no support for parallel or clustered databases like Oracle Parallel Server. 8 I/O sizes and RAID configurations are limited or fixed. 9 Extra burden is placed on NAS server to perform I/O, RAID and disk rebuilds.

Understanding the architecture of Fibre Channel SAN, Server Attached Storage, Database Block Access model: In this model, you have the low latency and high performance characteristics of the traditional proven model bcos of the dedicated I/O network . However, you also have the flexibility of a networking solution with distance, sharing and multiple-host connectivity. Let us examine the client server database access request procedure in Figure 8: Figure 8:

Anatomy of a Server Attached Storage (SAS) Block Database request: 1 Client makes data request over network to database server. 2 Database server performs I/O over Fibre Channel I/O bus to hardware RAID. 3 Database server sends results back to client over network.

Advantages of this access method: 1 Storage sharing among two or more host systems is possible. 2 Storage can be physically located further away from host systems. 3 Storage can be brought on-line and reconfigured dynamically. 4 Backups can be simplified. 5 It is possible to expand and add more positives. 6 Workload is off-loaded from LAN networks. 7 You can postpone or delay updates and improve performance of LANs. 8 There are no extra steps involved in performing I/O to reduce latency. 9 There is fully redundant hardware RAID with mirrored cache. 10 Integration with volume managers and other storage utilities is possible. 11 There is support for RAW volumes for Oracle and other applications. 12 Dynamic storage allocation and expansion can be done. 13 Variable I/O size, RAID levels and other performance enhancement tools are avail able. 14 Hardware RAID offloads server from RAID operations and disk drive rebuilds. Disadvantage of this access method: File or data sharing requires global file system/NAS like NFS or CIFs.

-Prepared By, Ashwin Pawar [email protected]