Some professions enjoy a regular 9-to-5. schedule. Data center management is not one of them. A restful evening may come to a sudden end due to a text or a frantic phone call. So, what are the factors that can keep a data center manager up at night?
William Hunter, member of the data center industry professionals association AFCOM, is a data center management expert who has worked for the likes of Amazon Web Services, Disney, and Cingular/AT&T Wireless. He named power as one of his biggest concerns.
To avoid late-night headaches, he worked hard to keep backup power systems functioning. This included ensuring UPS batteries could carry the load during an outage, and that generators not only had enough fuel but were properly maintained.
“It’s vital to do regular testing of UPS systems and stay up-to-date with manufacturer-recommended preventative maintenance,” said Hunter.
The level of change that permeates the industry keeps Donna Jacobs awake at night. This AFCOM member is the Director of Infrastructure Operations at University of Pennsylvania. She oversees two data centers and is in the process of consolidating them into one smaller-footprint facility. In addition to housing university administrative systems, these data centers serve as an internal colocation facility for spaces requiring conditioned space.
“The constant modification to on-prem and virtual configurations makes oversight and planning of capacity a moving target,” said Jacobs.
Her way to resolve this is full engagement with change management. This acts as a forum to expand the knowledge throughout the organization of the downstream effects of any technology changes. After all, capacity is not always available. Higher-density equipment randomly placed throughout the room, for example, can alter the disposition of cooling, power, and other key control variables.
Hardware failure is the big challenge that faces Jaynene Hapanowicz. As senior VP for IT infrastructure at Dell, she is responsible for 12 data centers globally, running a mix of Dell EMC server, storage, and networking gear. The data centers support more than 65,000 server instances, 3,000 applications, 19,000 database instances, and 145,000 users. They are about 80 percent virtualized.
“Hardware failures can really impact business services,” she said. “As my team provides infrastructure for multiple applications, we monitor our farm or cluster density because a hardware failure can magnify the impact to all those applications.”
For more factors that keep data center managers up at night, read this article in full on the AFCOM website.
The article is free to AFCOM members. To learn more about AFCOM or to join the association, visit here.
AFCOM is a sister organization to Data Center Knowledge.