Top Six Recommended Fibre Channel Physical Layer Best Practices

Most of the Global 2000 deploy and depend on fibre channel SANs to run their most mission-critical applications, and this is done for good reason.  First of all, fibre channel was built as a channel architecture, quite like data center mainframes. A channel generally has a high bandwidth and low protocol overhead.  Fibre channels were designed with internal mechanisms, like buffer-to-buffer credits, to reduce the effects of congestion.  Fibre channel also has services for scaling and management, including fabric routing protocols, policy-based routing, hardware trunking, virtual fabrics, and fault isolation.

Even with all this, I’m sure you’ve experienced some exasperating problems.  When applications start to act “funny,” we often get called in to help find problems.  It’s surprising and a little tragic how often the problems can be traced to simple physical layer issues.  To provide insight on how to avoid many of these physical layer issues, we got together with the Taneja Group to share some real-life stories for their new white paper on best practices to avoid physical layer issues.

Taneja covers these six best practices:

Design and Implementation

1. Manage Links to Protocols
2. Don’t Mix and Match Cables
3. Cheap Glass isn’t Maintenance

Operations and Maintenance

1. Test Links (or Fail)
2. Clean Cables (or Fail)
3. Bend Glass Only Intentionally

Highly recommend reading!