Storage Provisioning and Networking - Deploying and Managing a Cloud Infrastructure - page 251

number (LUN) of the faulty server. Another advantage is the effective disaster recovery
process through storage replication, where SAN can span a distant location containing a
secondary storage array.
A SAN utilizes a high-speed network, such as Fibre Channel (FC) through host bus
adapter (HBA), to connect computer data storage. FC technology commonly supports oper-
ating rates of 2, 4, 8, or 16 gigabits per second (Gbps). It uses Fibre Channel Protocol (FCP)
as the transport protocol that transports Small Computer System Interface (SCSI) com-
mands over FC networks.
A SAN makes available several devices and ports, but restricts access for each connected
system to a subset of devices/ports. Smaller subsets restrict interference, add security, and
simplify management. This is called FC zoning, and it can be done only in switched fab-
ric networks. However, FC can also be used over Ethernet (called FCoE); FC frames are
encapsulated over Ethernet networks, such as 10 Gigabit Ethernet (GigE), while preserving
the FCP protocol. Zoning differs from zoning in a virtual SAN (VSAN) in the sense that a
device or port can be part of multiple zones but only one VSAN. A VSAN is in fact a sepa-
rate network (think of VLAN).
There are a couple of types of zoning: soft and hard. Soft zoning restricts a machine to
see only the devices that it is allowed to see. However, any machine can contact any other
device using the device address. This helps in establishing security through obscurity. On
the other hand, hard zoning restricts communication physically across the network fabric,
which ensures security.
In an FC setup, devices are identified using a unique identifier called the World Wide
Number (WWN) or World Wide ID (WWID). Access to a device can also be restricted
using its WWN in a SAN environment. This is called name zoning and is more secure than
using port zoning. The port to which the host is connected can be moved while preserving
the access, whereas in port zoning, the access can be gained inadvertently to any resources
the previous host had access to.
Cloud Storage
As most other things in the cloud, storage is based on a highly virtualized infrastructure
and inherits the same characteristics as cloud computing in terms of agility, scalability, elas-
ticity, and multitenancy. It can be made available in both onsite and offsite configurations.
The biggest advantage of cloud storage is that it acts as a single entity despite the fact
that the infrastructure would most likely be geographically scattered. This makes many
distributed storage services behave and act as one, which is obviously advantageous in
terms of visibility (of data), transfer of data sources, and troubleshooting. Cloud storage is
agnostic when it comes to where the data is actually located. The administrator does not
have to worry about setting up a block or file storage configuration to determine where the
data actually resides (as in the case of a SAN).
Cloud storage provides high fault tolerance through redundancy and distribution of
data. Although SAN also provides redundancy, it is actually set up using techniques such as
Redundant Array of Independent Disks (RAID). Even though RAID has been used widely,