Storage Area Networking

Todd L. Montgomery, Stephen W. Olson · 2018

This chapter focuses on storage networks, their unique requirements, and many of the architectures and protocols that are used on a storage area network (SAN). Storage systems are intolerant to loss and require a lossless underlying network. The conflict with this is that Ethernet was never intended or designed to be lossless and has always relied on upper-layer protocols for recovery and retransmission of any data lost across the network. The SAN is defined as a high-performance network dedicated to storage traffic that is independent of all other traffic in the cloud. The SAN allows for shared storage resources to be remotely accessed by servers in the cloud data center. The SAN uses a specifically designed storage protocol called Fiber Channel and extends the SCSI command set across the network to remote storage arrays. Fiber Channel uses the concept of credits where the initiator communicates to the storage target when there are buffers available to send additional storage traffic.

Read the paper · More papers on PaperTik