Breaking Them Up and Putting Them Back Together


  One of the most common problems faced by Information Technology (IT) departments today is the persistent growth of applications and data. With high growth, employee turnover, and multiple acquisitions, administrators have difficulty managing heterogeneous applications, servers and storage. While tens of servers and applications with a few terabytes of storage could be handled by a single administrator, hundreds of servers and applications with petabytes of storage takes a team effort for secure, uninterrupted, application availability. The simple SAN has often grown into multiple complex SANs spread across multiple sites.

Instead of building one large Fabric or multiple disparate Fabrics, Fibre Channel has standardized techniques to break up large Fabrics and selectively put them back together. A large Fabric can be broken up with the concept of Virtual Fabrics that lets multiple Fabrics reside on a physical piece of hardware. Instead of having an 800-port Fabric on a few Directors and Switches as seen in Figure 1-1, Virtual Fabrics lets the administrator carve up the SAN into four 200-port Fabrics on the same infrastructure as shown in Figure 1-2. As with any type of organization, an intelligent division of the resources is required at some point.

Fig. 1-1 Large Unruly Fabric
Figure 1-1: When Fabrics span multiple sites and have hundreds or thousands of switch ports, the Fabric becomes difficult to manage and Fabric Services become cumbersome.

Virtualized Fabrics
Figure 1-2: Virtual Fabrics divides the management of the SAN into multiple Fabrics that are more manageable than one large Fabric.

Having multiple small Fabrics on the same physical hardware enables the administrator to break the SAN management into comprehensible pieces that can be managed more effectively. For example, instead of having one large Zone Set that many administrators alter, each administrator can be responsible for a limited set of devices and applications in a smaller Zone Set. This divide and conquer mentality of Virtual Fabrics will be explained in Chapter 3.

The story of Virtual Fabrics doesn’t end there. After the Fabric is carved up into manageable pieces, selective devices can be configured to communicate between the separate Fabrics via Inter-Fabric Routing. Inter-Fabric Routing allows communication between devices from different Virtual Fabrics back without merging the Fabrics. As shown in Figure 1-3, an Inter-Fabric Router (IFR) sits above Virtual Fabrics and proxies devices in and out of Fabrics so that the devices can communicate while the Fabrics remain separate. Inter-Fabric Routing is discussed in Chapter 5.

Inter-Fabric Routing
Figure 1-3: The goal of Inter-Fabric Routing is to allow selective communication between a subset of devices in different Fabrics. In this example, Device 2’ is proxied into Fabric 1 and Device 1’ is proxied into Fabric 2. The Inter-Fabric Router on the left makes the devices look like they are directly attached to the Fabric as shown on the right.

Source:Extract from Chapter 1 of "Fibre Channel Advances" by Scott Kipp

Back to Article List