Bill Burnham, CTO for US Public Sector at Hewlett Packard Enterprise, speaks with Katya Mijatovic, Principal Data Scientist at Data Society and CDO Magazine Editorial Board Member, about technologies that made the biggest impact, implementing new technology and approach to choosing locations for setting up data centers.

Mijatovic interviewed Burnham on the sidelines of the recently conducted AFCEA CeVA Data Centric Summit.

Highlighting a takeaway from his presentation at the summit, Burnham mentions that the purpose of the future enterprise is to run artificial intelligence and data science. The traditional enterprise is ill-equipped to run artificial intelligence and provide access to data where it resides. His presentation covered the evolution from a traditional three-tiered environment into an AI-capable enterprise.

Speaking of technologies that made the biggest impact in the past decade, Burnham mentions the miniaturization of computing which has changed enterprise’s ability to bring computing to the edge where most of the data is collected. It has been enabled by graphical processing units (GPU) that parallelize data processing.

Burnham further mentions that post-2016 once the compute got small enough, it enabled enterprises to build models that support AI and made it possible to fine-tune existing models rapidly.

He goes on to discuss new technologies he is looking forward to. Burnham mentions private 5G boosting the ability to connect edge devices at longer distances in a private network, connecting to data collectors like IOT, and bringing smaller and more capable compute to the edge.

As Burnham emphasizes, the volume of data generated at the edge is expanding exponentially, encompassing everything from 4k video streams to sensor data. This surge isn't confined to defense; it extends to retail, where vast quantities of data are amassed and analyzed. Consequently, processing must occur at the edge.

He adds that investments in edge computing will enable operations beyond traditional data centers. Leveraging GPUs, processing tasks can be executed where the data originates. This shift marks a departure from the conventional practice of centralizing data in data centers.

The emerging paradigm according to Burnham involves decentralized data, with high-performance computing disseminating workloads to the data's location. By harnessing the connectivity of 5G and bolstering edge computing capabilities, insights can be gleaned more swiftly.

Responding to Mijatovic’s question about what goes into a decision to build data centers, Burnham points out that the size of a data center hinges entirely on its power consumption. A data center's scale is measured in megawatts and not in terms of racks or servers, as these dimensions are inconsequential compared to power usage.

Further Burnham explains that Virginia is an ideal destination for data centers because of affordable electricity rates, a crucial factor in the evolving landscape of sustainable data centers. He adds that the availability of affordable power and efficient cooling systems will dictate the industry's direction. Consequently, there's a growing likelihood of data centers gravitating towards colder climates, where cooler natural air can assist in cooling infrastructure at lower costs.

CDO Magazine appreciates Bill Burnham for sharing his insights with our global community.

Public Interest Technology

Related Stories