Cisco FCoE Switches Help Hospital Save Power Page 2


Want the latest storage insights?

Download the authoritative guide: Enterprise Data Storage 2018: Optimizing Your Storage Infrastructure

Continued from Page 1

Big Power Savings

While FCoE is seen as the future, the organization hasn't eliminated its traditional Fibre Channel fabric. Older servers continue to run FC and are being switched out gradually. Him estimates that it might be three years before they are entirely changed out. But the results of the switch to FCoE have been immediately apparent in terms of greater reliability and lower energy consumption. As a result, the organization no longer has a power crunch.

"Since we put Nexus in place six months ago, we've had no network downtime," said Him. "The Nexus 2000 consumes 89 watts per box with data running across it. We were very short on power before, but now we have more margin."

The Nexus and MDS both use the NX-OS, which helps make management and training easier. Salem Hospital has two engineers trained on this operating system. It provides them with visibility into the network and storage, so a networking technician can troubleshoot without having to call upon his storage equivalent, and vice versa. Previously, far more interaction was needed between the networking and storage disciplines to resolve problems.

Another element of the Cisco plan for global data center domination is the Cisco Unified Computing System (UCS). This compact unit combines storage, server and networking gear in one form factor and does away with a lot of cabling. The hospital currently has Cisco UCS under testing.

"We are excited about the initial numbers from Cisco UCS," said Him. "We are looking at the possibility of using it for our entire data center."

FCoE Learning Curve

While Him is excited about the potential of FCoE, he admitted that getting it up and running on the first couple of servers was far from plain sailing. He ran into issues with drivers and difficulties in configuration.

"As the Windows drivers were not working properly, we weren't provisioning correctly," he said. "It was mainly configuration glitches and a general lack of understanding."

With the experience gained from the first couple of machines, he said setting up servers and getting VMs running over FCoE now saves significantly on overall deployment time. Him gives the example of an imaging system that ran on an older server. The hospital moved it onto a VM running over FCoE in a few hours.

His advice to others on FCoE implementation is to extensively test provisioning of LUNs to ensure the various channels are working properly.

"Once you get comfortable with the technology, it becomes so much easier," said Him. "It's all about testing."

Follow Enterprise Storage Forum on Twitter

Submit a Comment


People are discussing this article with 0 comment(s)