A Switched Fabric SAN topology is the most widely used and scalable architecture in Fibre Channel (FC) Storage Area Networks. In this design, servers and storage devices are connected through Fibre Channel switches, forming a fabric that allows any device to communicate with any other device efficiently, reliably, and at high speed.
Unlike Direct-Attach or Arbitrated Loop topologies, FC-SW provides high performance, redundancy, scalability, and intelligent traffic management, making it the preferred design for enterprise data centers.
A fabric is a collection of one or more interconnected FC switches that:
-
Provide a path between servers (initiators) and storage (targets)
-
Manage data traffic using intelligent switching
-
Maintain device information through Name Server
-
Ensure security using zoning
-
Offer multiple paths for redundancy
| Component | Description |
|---|---|
| HBA (Host Bus Adapter) | Connects server to FC network |
| FC Switch | Core device that builds the fabric |
| Storage Array Ports | Target ports for data access |
| Fibre Channel Cables (SFP + Fiber) | Physical connectivity |
| Fabric Services | Name Server, Zoning, RSCN, Routing |
-
Server HBA logs into the fabric (FLOGI)
-
Device registers with the Name Server
-
Zoning rules are applied
-
Server discovers storage ports
-
Data transfer begins through the shortest available path
-
Fabric manages traffic, congestion, and failover automatically
| Port Type | Purpose |
|---|---|
| F_Port | Connects end device (HBA/Storage) |
| E_Port | Connects switch to switch (ISL) |
| G_Port | Generic, auto becomes E or F |
| U_Port | Uninitialized port |
| TE_Port | Trunking E-Port for ISL trunking |
-
Scalability
-
Supports thousands of devices
-
Easy to add new switches, servers, or storage without disruption
2. High Performance
-
Dedicated bandwidth per port (16G/32G/64G/128G)
-
No shared loop like FC-AL
-
Low latency switching
3. Redundancy & High Availability
-
Multiple paths between devices
-
Fabric re-routes traffic automatically if a link fails
4. Intelligent Switching
-
Frame-based switching
-
Shortest path selection
-
Congestion control
5. Security with Zoning
-
Controls which server talks to which storage
-
Prevents unauthorized access
-
When multiple switches are connected, they form ISLs.
Benefits:
-
Fabric expansion
-
Load balancing
-
Redundancy
-
ISL Trunking increases bandwidth
| Service | Function |
|---|---|
| Name Server | Maintains WWPN database of devices |
| RSCN | Notifies devices about fabric changes |
| Zoning | Access control |
| FSPF | Routing protocol for path selection |
| Feature | FC-AL | Direct Attach | FC-SW |
|---|---|---|---|
| Scalability | Low | Very Low | Very High |
| Performance | Shared | Limited | Dedicated |
| Redundancy | No | No | Yes |
| Management | Hard | Simple | Intelligent |
| Enterprise Use | No | No | Yes |
Enterprise SAN always uses two fabrics:
-
Fabric A
-
Fabric B
Each server and storage connects to both fabrics.
Why?
If Fabric A fails, traffic continues via Fabric B without downtime.
1. FSPF (Fabric Shortest Path First)
-
Routing protocol used inside FC fabric
-
Chooses best path dynamically
2. ISL Trunking
-
Combines multiple ISLs into one logical high-bandwidth link
3. NPIV (N_Port ID Virtualization)
-
Multiple virtual WWPNs on a single physical HBA port
-
Used in virtualization (VMware)
4. NPV (N_Port Virtualizer)
-
Access switch mode without full fabric services
-
Reduces domain IDs
5. QoS and Traffic Isolation
-
Prioritizes critical storage traffic
6. Fabric Binding & Security Policies
-
Prevent unauthorized switches from joining fabric
-
Zero single point of failure
-
Predictable performance
-
Massive scalability
-
Secure communication
-
Easy to manage large environments
-
Ideal for virtualization, databases, and critical workloads