How will the management of data centers today compare to the management of data centers tomorrow?

Download the presentation

The methods in managing the data center must adapt, as more and more economies grow around the utilization of data created through the internet of things. The more data acquired will require greater processing power which in turn puts an increased strain on the supporting infrastructure. Balancing efficient data center management with increased level of processing can all too easily become an impossible situation if the current methods remain the same.

The traditional approach to data center management has been fairly static over the last number of years, consisting of IT and facilities teams, each responsible for the upkeep of their respective systems. In most cases both teams operate in separate silos with very little overlap, with a more reactive approach as opposed to a proactive one. While this approach has worked well enough over the years, the requirement for truly efficient data centers requires a greater and far more cooperative approach.

With the advent of the IoT age, we can see a substantial increase in the sheer volume of data being created. Not only must we increase data management capabilities in order to utilize the data created, we must also adapt the overall data center management method. Focusing purely on how to best utilize the rivers of data as a result of IoT would be shortsighted. There is a definite need to change the overall management of the data center so that the storage and processing of this data can be efficiently managed. In its true form, the basic principles around the structure of the data center will not change. The goal of supporting the data applications through redundant powered, environmentally controlled environments, using redundant network paths will remain the same. We can easily compare data centers to the factory, assembling components into a finalized product. Data centers are bringing together different systems to support the application. But just like factories, over time the focus on efficiency has revolutionized the management of the overall process.

The greatest shift in data center management is the introduction of machine learning to the supporting infrastructure. Major players such as Google have already taken this leap, pointing their Deep Mind AI machine learning at their data center infrastructure. This shift in the management and control in the data center, moving from human managed and controlled environments to complex computing machine learning environments, results in a greater deal of accuracy and intricate control in the management of data center infrastructure in balance to the application layer. The ability to achieve balance in fine-tuning the power and environmental systems is a major leap in the drive to overall efficiency. This treatment of the data center as a single system allows for overall efficiency instead of efficiency based on individual systems.

The impact of machine learning in the management of the data center will result in the change in the role of the data center support staff. In order to cope with the management of vast amounts of data produced by the growing number of connected devices, there is a definite requirement for the moving of trust in human hands-on approach handling the operation of the separate systems in the data center to full artificial intelligence/machine learning controlling a single system. But exactly where people and the required skillsets will fit into this remains to be seen. Most definitely we will see a reduction in headcount required to manage the facility; so most likely we will have smaller cross functional teams with a broader range of skills from electrical engineering through to data scientists managing the day-to-day operations of the data center.

To conclude, the data center must be managed as a single system and machine learning will play a major factor in this. This reliance on machine learning will greatly impact the level of efficiency the data center is required to achieve, however, the efficiency focus within the data center will range from the application layer down through to the supporting infrastructure and not on individual systems in the chain. The impact of this change will mean far less human interaction with the day to day operations of the data center resulting in a larger technician to host ratio, but also an eradication of operating silos.
 

About the panel speaker:

Dennis O'Sullivan has over 15 years of experience in the data center industry. He joined Eaton’s data center automation development team in 2014 and, as Data Center Solutions Operations Specialist, is responsible for providing insight into the day-to-day challenges of data center operators and managers to aid in the development of new products.

Before joining Eaton, he worked for Yahoo! for over nine years as part of its EMEA data center operations team, overseeing the build-out and operation of several data centers and numerous network point-of-presence sites. His other previous positions include working for BT Ireland as part of its data center network operations team.