What is the typical employment density of a data center?
Better data center management tools mean an exponential shift in the number of IT staffers that can adequately support data center operations.
The conventional data center staffing model -- before automation tools were developed -- dictated that an IT team hire one system administrator (sys admin) for every 20 servers and that no more than 12 people reported to any other single person in the IT department. So, if your IT shop ran 100 servers, you would need five sys admins and a manager. For 1,000 servers, 50 sys admins would report to five supervisors and a manager. In addition, you'd have an equivalent number of storage and network administrators. On the test and development side of IT, you'd have another set of employees.
Looking for a
new job in IT?
Meet the CDO -- chief digital officer
Work on your third platform IT skills
Learn what experts want in a new admin
The IT staff could soon amount to a massive cost center for the business. Well, so much for theory, as this rarely would be the case.
In today's data center, a lot of staffing decisions come down to how well-tooled the data center is. If it uses a full management system, comprising a data center infrastructure management (DCIM) system and systems management tool set, then the IT team can automate the tasks around running platforms. This is how many service providers run their businesses, with very few sys admins looking after the platform or the facility.
Test and development isn't immune to this change in data center staffing either. With DevOps, who is to say where a developer stops and a data center operator begins?
A modern data center should aim for a single sys admin supporting around 1,000 physical servers. In 2009, Microsoft Corp. managed its Chicago data center of 300,000 servers with just 30 IT staff members, including the maintenance and facilities group; expectations are that it will need fewer IT staff members per server in the future. However, this level of support is only possible for identically configured servers. If you had 2,000 servers, each with different configurations, one system administrator would burn out in around two days.
About the author:
Clive Longbottom is the co-founder and service director of IT research and analysis firm Quocirca, based in the U.K. Longbottom has more than 15 years of experience in the field. With a background in chemical engineering, he's worked on automation, control of hazardous substances, document management and knowledge management projects.
This was first published in November 2013