Saturday, April 16, 2011

Comparing DAS, NAS, iSCSI, SAN


Purpose:

The purpose of this document is to briefly explain the different types of storage options available and there advantages and disadvantages.

Storage Use Considerations Factors:

  • Available budget
  • Data security requirements
  • Network infrastructure
  • Data availability requirements, etc.

Storage Types:

  • Direct Attached Storage (DAS),
  • Network Attached Storage (NAS),
  • Storage Area Networks (SAN).

Direct Attached Storage:

Direct Attached Storage is a system of hard drives addressed directly via system buses within the computer (IDE, SCSI); the network interface is managed by the operating system. As these buses can only bridge short distances within the decimeter range, DAS solutions are limited to the respective computer casing. Depending on the bus type, DAS systems are also restricted to a relatively small number of drives - Wide-SCSI achieves the maximum of 16 directly addressable drives. Due to these limitations and the need for more flexible storage, the importance of DAS is declining. Although DAS in terms of terabyte is still growing by 28% annually, the need for storage is increasingly being covered by networked storage like NAS and iSCSI systems.

Network Attached Storage:

NAS systems are generally computing-storage devices that can be accessed over a computer network (usually TCP/IP), rather than directly being connected to the computer (via a computer bus such as SCSI). This enables multiple computers to share the same storage space at once, which minimizes overhead by centrally managing hard disks. NAS devices become logical file system storage for a local area network. NAS was developed to address problems with direct attached storage, which included the effort required to administer and maintain “server farms”, and the lack of scalability, reliability, availability, and performance. They can deliver significant ease of use, provide heterogeneous data sharing and enable organizations to automate and simplify their data management.



NAS Application Uses (Low Performance):

  • File/Print server
  • Application specific server
  • Video Imaging
  • Graphical image store
  • Centralized heterogeneous file sharing
  • File system mirroring
  • Snap shot critical data
  • Replacement of traditional backup methods
  • Medical imaging
  • CAD/CAM
  • Portable centralized storage for offsite projects
  • Onsite repository for backup data

Advantages of NAS:

  • NAS systems offer a number of advantages:
  • Heterogeneous OS support. Users running different types of machines (PC, Apple iMac, etc.) and running different types of operating systems (Windows, Unix, Mac OS, etc.) can share files.
  • Easy to install and manage. NAS appliances are “plug-and-play” meaning that very little installation and configuration is required beyond connecting them to the LAN.
  • NAS appliances can be administrated remotely, i.e. from other locations.
  • Less administration overhead than that required for a Unix or Windows file server.
  • Leverages existing network architecture since NAS are on LANs.
  • NAS server OSs are smaller, faster, and optimized for the specialized task of file serving and are therefore undemanding in terms of processing power.
  • A NAS appliance is a standalone file server and can free up other servers to run applications. Compared to iSCSI an additional host server is not necessary.
  • Compared to iSCSI, NAS appliances already include integrated mechanisms for backup, data synchronization and data replication.

Disadvantages of NAS:

  • Heavy use of NAS will clog up the shared LAN negatively affecting the users on the LAN. Therefore NAS is not suitable for data transfer intensive applications.
  • Somewhat inefficient since data transfer rides on top of standard TCP/IP protocol.
  • Cannot offer any storage service guarantees for mission critical operations since NAS operates in a shared environment.
  • NAS is shared storage. As with other shared storage, system administrators must enforce quotas without which a few users may hog all the storage at the expense of other users.
  • NAS is less flexible than a traditional server.
  • Most database systems such as Oracle or Microsoft Exchange are block-based and are therefore incompatible with file-based NAS servers (except for SQL).

Storage Area Network:

Storage Area Networks (SAN), which also include iSCSI, are distinguished from other forms of network storage by using a block based protocol and generally run over an independent, specialized storage network. Data traffic on these networks is very similar to those used for internal disk drives, like ATA and SCSI. With the exception of SAN file systems and clustered computing, SAN storage is still a one-to-one relationship. That is, each device (or Logical Unit Number (LUN)) on the SAN is “owned” by a single computer (or initiator). SANs tend to increase storage capacity utilization, since multiple servers can share the same growth reserve. Other benefits include the ability to allow servers to boot from the SAN itself. This allows for a quick and easy replacement of faulty servers since the SAN can be reconfigured so that a replacement server can use the LUN of the faulty server.


iSCSI/SAN Application Uses(High Performance):

  • Offer power users disk space on demand
  • Databases (Oracle, MS-SQL, MySQL)
  • Video Imaging
  • Graphical image store
  • File system mirroring
  • Snap shot critical data
  • Replacement of traditional backup methods
  • Medical imaging
  • CAD/CAM
  • Onsite repository for backup data
 
Advantages of iSCSI:

  • Ease of scaling disk storage. With iSCSI the disks are remote from the server, therefore adding a new disk just requires the use ofdisk manager or if replacing the whole server re-mapping the data to the server using iSCSI. With iSCSI you can easily create huge storage pools with volumes in the range of several tera- or petabytes.
  • In comparison to NAS, which provides a file-level interface, iSCSI provides a block level interface and are therefore compatible with database applications such as Oracle or Microsoft Exchange, that also use a block level interface. Leverages existing network architecture since iSCSI are on LANs.
  • iSCSI storage appliances can be seamlessly integrated into existing SAN environments, since it also runs on block level storage.
  • iSCSI can provide significant benefits for providing failover for high availability configurations. iSCSI allows IP based replication, mirroring and clustering of data and offers integrated MPIO (Multi-Path I/O).
  • iSCSI can also be configured as a particularly flexible DAS system – the local SCSI bus is so to speak extended by the network.
  • As iSCSI is an underlying technology to the OS and uses the native file system of the applications, it is fully compatible with all file systems.



Disadvantages of iSCSI:

  • The demands of accommodating SCSI commands and SCSI data packets in TCP/IP packets require extensive hardware resources: CPU performance should be at least that of a 3 GHz Pentium processor, Gigabit Ethernet (GbE) should accordingly be used as a network interface, and the RAM requirement is also significant.
  • For sharing iSCSI targets with multiple initiators, additional server or specific (client) software for shared data access is required. Known providers of SAN data sharing software are Adic, Sanbolic, IBM, Polyserve, Dataplow and SGI.
  • As iSCSI is an underlying technology to the OS and application, anything that an organization currently has can be used. On the other hand this means that extra licenses for OS and software applications might be needed.
  • Comparing to NAS, an iSCSI Target is not a standalone device and an additional host server is necessary.
  • For sharing centralized storage pool disk among heterogeneous OS requires additional sharing software.
  • In iSCSI appliances mechanisms for backup, data synchronization and data replication are not integrated and must be configured. Comparing to NAS, iSCSI behaves like a local hard drive in the network.
 
Advantages of SAN:

  • A SAN will have all the advantages of iSCSI.
  • A SAN has higher throughput since it runs over dedicated fibre channel topology and not the LAN.
  • A SAN will reduce the overhead that iSCSI places on system resources.

Disadvantages of SAN:

  • Implementing a SAN infrastructure will cost more then NAS or iSCSI due to the additional equipment needed to build out the fibre channel topology.

Summary:

  • Direct Attached Storage is probably not going to feasibly allow us to grow our disk capacity in the future.
  • Network Attached Storage (NAS) is the obvious choice for a storage solution wherever the main focus is on storing and archiving files and shared access to these over a network – even from different client operating systems. Small and medium-sized businesses, typing pools, legal or agency offices, and even end users with large amounts of multimedia files will find an affordable storage solution for their needs in NAS.
  • For storing database systems - other than SQL-based database systems - on a network, NAS is however not a feasible solution. For requirements of this type the industry has developed the Storage Area Network (SAN) technology, which can often be implemented using iSCSI components. Advantages of iSCSI: An IP-based SAN allows administrators to use their familiar management tools and security mechanisms and rely on their existing know-how. However, iSCSI only makes sense in connection with a fast LAN infrastructure: At a throughput of approximately 120 Mbyte/s, the performance of a 1 Gbit Ethernet will be sufficient for database applications for approximately 100 users (data volume: approx. 15 MByte/s). Only high-end storage systems will require a 10 GbE infrastructure. Somewhat inefficient since data transfer rides on top of standard TCP/IP protocol. In contrast, SAN uses protocols designed especially for data transfer (though the advantage disappears if a server on the LAN is used to provide a file interface to a SAN).

Mapping Global Zone Name Inside Solaris 10 Zone

Log into global zone where local zone(s) reside.

Using zonecfg, configure the following for each zone:
# zonecfg -z (zone)
add fs
set type=lofs
set options=ro
set special=/etc/nodename
set dir=/etc/globalname
end
verify
commit
exit

Create mount point within local zone directory structure:
# touch /zones/(zone)/root/etc/globalname

Mount lofs file system manually:
# mount -F lofs -o ro /etc/nodename /zones/(zone)/root/etc/globalname (or path where the root of the zone resides)

Confirm local zone can access file:
# zlogin (zone) cat /etc/globalname

Change /tmp size on Solaris 10 Zone

Log into global zone where local zone(s) reside.

Using zlogin, login to the local zone:
# zlogin –z (zone)

As root edit the vfstab:
# vi /etc/vfstab

Find the line in vfstab for the /tmp filesystem:
swap - /tmp tmpfs - yes size=512mb (this could be any value, 512 is example)

Change the value of size=512mb to the requested value (in MB):
swap - /tmp tmpfs - yes size=2048mb

Save the vfstab and exit back to the global zone.

To make the changes take effect the zone must be stopped and booted:
# zoneadm –z (zone) halt
# zoneadm –z (zone) boot

  Log back into local zone and confirm changes by reviewing df –h output:
# zlogin –z
# df –h|grep /tmp
swap                   2.0G     0K   2.0G     0%    /tmp