building block solutions
An Introduction to iSCSI Storage with
Wasabi Storage Builder® for IP-SAN
It's no secret that the need for data storage is increasing day by day. It is no longer just large corporations that require a lot of storage; even Small and Medium Businesses (SMBs) are faced with the need to store more data.
There are several factors that are contributing to the need for more storage:
**Data is increasingly being stored digitally. Everything from audio to x-rays, including photos, movies, music, medical images, security video, and nearly all documents, are now stored in a digital format.
**Hard disks, due to their cost and performance benefits, are quickly replacing tape for backing up and archiving data
**Legislation such as the Sarbanes-Oxley requires that certain documents and e-mail be kept for a period of several years.
It's not simply the storage of data that needs to be taken into consideration, that data must also be secure, protected, easy to manage, and accessible over vast distances. Current storage methods have been too expensive, too complex, and fundamentally inadequate for most SMBs. Consequently, their storage choices have been limited.
Until recently, storage options for the SMB have been too expensive, too complex, or too limited. All too often storage products offered features that SMBs didn't need, and certainly didn't want to pay for.
To view in more detail the problems SMBs face with regards to storage, let's look at what their storage options are. Storage today can be broadly categorized as one of two types - Direct Attach Storage (DAS) and Storage Area Networks (SAN).
DAS
As the name implies, this is storage that is directly attached to a server. DAS can be internal - hard drives mounted internally inside the server, usually attached to a RAID controller; or it can be storage that is attached externally to the server, most commonly to a SCSI or Fibre Channel HBA.
Figure 1: Typical external DAS configuration
Internal DAS has two major drawbacks. The first drawback is lack of scalability. If the server's storage capacity becomes full, additional hard drives must be added to the server. This can present a problem in that there may not be any available space in the server chassis for more hard drives.
Even if there is additional space, this leads us to the second drawback of internal DAS storage. In order to add more hard drives to the server, the server needs to be taken offline and shutdown in order to open up the chassis to install the additional hard drives. If the server is expected to be running 24/7, such as a web server, then the server's data will not be accessible while the system is offline.
External DAS, while more flexible than internal DAS, still has some drawbacks. Cable length is limited so the storage must be located relatively near the server. In addition, the server requires extra hardware, such as a SCSI or Fibre Channel HBA, in order to be able to communicate with the storage, which greatly increases the cost of a server.
SAN
SAN storage allows multiple servers to have access to the same storage resources.
Figure 2: In a SAN, multiple servers can access a shared storage system.
Sharing storage offers several benefits including improved scalability, more efficient use of disk space, and easier management and maintenance due to consolidation of storage resources.
Traditionally, SAN storage has used a Fibre Channel (FC) interface. Fibre Channel is a data transmission protocol and includes components such as HBAs, switches, and cables. FC technology has generally been out of reach for SMBs; primarily due to the high cost of implementation, but also because many SMBs do not want to pay for the specialized knowledge required to configure and manage FC networks. Consequently, SMBs have not been able to take advantage of the benefits offered by SAN storage.
As the need increases for SMBs to face the same storage issues as large companies, but without the budgets or knowledge needed to implement FC, a lower-cost solution is required.
iSCSI: How it Works
iSCSI is a protocol for connecting external storage to a server. It takes SCSI data packets and encapsulates them into IP packets, which are then transmitted over Ethernet. iSCSI devices are classified as either an initiator or a target. The initiator is the computer system that will be using the iSCSI storage. For example, the computer system could be a file or e-mail server. The target is the iSCSI storage device. The following figure shows a basic iSCSI configuration, with the initiator being directly attached to the target.
Figure 3: The Ethernet cable is plugged into standard Gigabit Ethernet ports.
iSCSI can also be used as part of an IP SAN, with multiple initiators and multiple iSCSI targets all connected through an Ethernet switch, as shown in the following figure.
Figure 4: An IP-SAN consisting of multiple servers (initiators) and iSCSI targets.
In this configuration, all of the storage is available to any of the servers. This allows storage to be more efficiently allocated. It also provides better scalability. Should more storage capacity be required, simply attach an additional iSCSI storage target into the switch. There is no need to shutdown any of the servers, and the iSCSI targets will remain online and continue to be accessible.
All that is required for the server to communicate with the iSCSI target is for the server to have an iSCSI initiator installed. An iSCSI initiator is used to establish communication between the server (initiator) and the iSCSI target. While there are iSCSI initiator host bus adaptors available from several hardware vendors, there are software iSCSI initiators that are freely available and typically included with virtually every major operating system, including Windows and Linux. These software iSCSI initiators allow standard Ethernet ports to be used to communicate with an iSCSI target. So, unlike DAS or Fibre Channel storage, which require an HBA or a motherboard with embedded SCSI or FC support (which add to the cost of the server), using a server as an iSCSI initiator does not require the purchase of any specialized hardware.
Once an iSCSI initiator is installed, it is used to login to and establish a connection with the iSCSI target. Once the initiator has logged in, the iSCSI target is a virtual hard disk, and appears to the operating system exactly as a local hard drive. For example, in Windows, you use the standard Disk Administration tool to create partitions and format the iSCSI virtual disk, and it gets assigned a drive letter.
Figure 5: Once logged in, the iSCSI targets appear just as if they were local storage.