Fabric Switches
Fabric Switch
Performance
In similar fashion to directors, fabric switches also provide high-
performance, dynamic connections between end devices in a Fibre
Channel switched network. Fabric switches also support mainframe
and OSI computing environments.
Through non-blocking architecture and limited FRU redundancy,
fabric switches also offer high availability and high-performance
bandwidth. Although switches do not offer the redundancy,
availability, or port count of an enterprise-class director, they offer a
much lower-cost connectivity option. Fabric switches should be
installed for:
•
Implementation as the principal building block of a small-scale
SAN or as a consolidation point for enterprise-class SANs.
•
Departmental and workgroup connectivity.
•
Applications where distributed storage predominates.
Fabric switches also provide connectivity between servers and
devices manufactured by multiple OEMs. To determine if an OEM
product can communicate through switch connections or if
communication restrictions apply, refer to the product publications or
contact McDATA.
Fabric switches provide an availability of 99.9% through a redundant
configuration of power supplies and cooling fans. When an active
FRU (power supply or fan) fails, the backup takes over operation
automatically to maintain switch and Fibre Channel link operation.
Availability is also provided through concurrent firmware upgrades
and spare or unused Fibre Channel ports.
Excluding an availability of 99.999%, fabric switches offer the same
general performance features as directors, including high bandwidth,
low latency, local control, low communication overhead, multiple
topology support, and multiple service class support.
Introduction to McDATA Multi-Protocol Products
Introduction to McDATA Multi-Protocol Products
1
1-15