IBM DS3500, Введение и руководство по внедрению

Страница: 26 / 790

7914DiskAttach0908.fm
Draft Document for Review March 28, 2011 12:24 pm
2
IBM System Storage DS3500: Introduction and Implementation Guide
1.1 Fibre Channel disk attachment
Fibre Channel (FC) is a high-speed disk attachment technology primarily used for storage
networking. It is designed to connect a large number of storage devices to a number of host
servers across a Storage Area Network (SAN). Fibre Channel is a transport Protocol (FCP)
which transfers SCSI commands and data over Fibre Channel networks.
FC supports a much higher number of devices and much longer cable lengths than SCSI. It
has become the preferred disk attachment technology in midrange and large scale datacenter
solutions.
At the time of writing, DS3500 Storage maximum FC throughput is 8 Gbps. In fact, 10 Gbps
links can be used today, but only for SAN switch interconnection.
Host servers contain one or more FC Host Bus Adapters (HBA). The HBAs provide
connectivity to the storage devices using FC cabling and SAN Switch.
For more information about Fibre Channel and SANs, see Introduction to Storage Area
Networks, SG24-5470.
FC topologies
There are three major Fibre Channel topologies, describing how a number of ports are
connected together. A port in Fibre Channel terminology is any entity that actively
communicates over the network, not necessarily a hardware port. This port is usually
implemented in a device such as disk storage, an HBA on a server or a Fibre Channel switch.
򐂰 Point-to-point
Two devices are connected directly to each other. This is the simplest topology and
provides a direct link between an FC HBA inside a host server and a storage device, with
limited connectivity.
򐂰 Arbitrated loop
This topology can be used to interconnect several FC devices. A typical example would be
to attach a certain number of host servers to an FC storage subsystem. A loop can consist
of up to 127 devices.
A minimal loop containing only two ports, while appearing to be similar to FC-P2P, differs
considerably in terms of the protocol. Only one pair of ports can communicate
concurrently on a loop. This means the devices share bandwidth, so the arbitrated loop
topology is not suitable for high performance requirements.
Arbitrated loops were commonly implemented with the use of an FC hub. Even though this
is physically a star topology, logically it will be a loop. Alternatively, devices can be
connected in a daisy chain manner.
Arbitrated loops are rarely seen these days, as switched fabrics have become the norm.
򐂰
Switched fabric
The most commonly used topology in a typical SAN today is switched fabric. SAN
switches are used to provide FC connectivity between the host servers and storage
devices. Switched fabrics can become very complex in large scenarios, connecting
hundreds of host servers to a very large number of storage subsystems.
SAN switches provide optimized traffic flow and increased performance by allowing
concurrent data transfers between many connected hosts and storage devices. Switched
fabrics can provide dedicated bandwidth, as opposed to arbitrated loop technology, where
the bandwidth is shared among all the devices in the loop.