DICE EAST 2023: Experts Discuss Operational Intelligence and the Path to Progress for Data Centers
Held inside the #1 data center market in the world, DICE EAST was a prime location for a discussion on the future of data centers and a shift to operational intelligence.
DICE EAST is a national two-day Data Center Event bringing together industry experts, professionals, and stakeholders to discuss and address key topics and challenges in the data center industry. Held in Tysons Corner, VA this past May 24-25, the event promotes knowledge sharing, networking opportunities, and collaboration among participants. It provides a platform for discussing emerging trends, innovations, and best practices in data center operations, sustainability, power solutions, workforce development, and public perception.
In an insightful discussion, industry experts gathered to shed light on the future of the data center, including energy efficiency, operational intelligence, and the systems to manage and automate them.
Members of the panel included:
- Yuval Bachar, the CEO of EdgeCloudLink (ECL), previously the Chief Architect of LinkedIn data centers at Microsoft Azure.
- Peter Gross, an electrical engineer with extensive tenure in the industry, beginning as a power systems designer and evolving into consulting work related to data centers.
- Ben Stewart, retired Coast Guard member, serving as part of the NTT Data Centers team, overseeing operations with his extensive experience in five colocation operations.
- Michael Dongieux, the CEO and Chief Architect at MCIM, a software company specializing in operations and maintenance software for data centers. During his tenure at Wells Fargo as a data analyst, Michael played a pivotal role in developing their global critical facility information management system.
The Evolution of Data Centers: From Energy Efficiency to Operational Intelligence
First, the panel recognized the significant role of energy efficiency in shaping the data center industry. It emerged as a response to the growing concerns about the environmental impact of data centers.
Initiatives such as a congressional inquiry and environmental organizations like Greenpeace pressured the industry to prioritize energy efficiency. The introduction of the PUE (Power Usage Effectiveness) metric by the Green Grid provided a standardized measure for evaluating energy efficiency and encouraged organizations to strive for better performance. Energy efficiency became the initial focus in the industry’s efforts to reduce carbon emissions brought on by climate change and remains a crucial aspect today.
The Shift towards Operational Intelligence
While energy efficiency continues to be important, the panel next discussed a shift towards a new driving force: operational intelligence. The exponential growth of data centers, especially within the colocation industry, has presented staffing challenges. Additionally, the impact of the COVID-19 pandemic has further complicated workforce dynamics, with many employees preferring remote work arrangements. This has created a pressing need for innovative solutions.
Operational intelligence, a subset of AI, offers a promising solution. By integrating business intelligence into data center management tools, organizations optimize resource allocation and potentially reduce the dependency on a large workforce. This transformative approach allows for better matching available resources to the anticipated workload. AI-driven technologies will enable data centers to adapt to changing workforce preferences and efficiently address staffing challenges.
The Rise of Data Center Operating Systems
The next topic discussed was how the data center market needs help creating consistent definitions and fragmented data sources. By establishing centralized management of manufacturers, models, and OEM data sheets, data centers streamline processes and improve data integrity. This standardization allows for better comparability and sets the foundation for data-driven decision-making.
“If you look back across the years, we’ve all had disparate, siloed CMMS systems. You’ve got DCIM systems, EPMS systems, and ITSM systems. They go on and on and on. There’s always some SharePoint and some Excel mixed in there for fun. But what I’ve noticed and heard repeatedly with colos is that their data’s stranded. It’s not very actionable. You want mission critical systems that are connected; the data needs to be cleaned, the data needs to be interrelated. And once you’ve got those two ingredients, then you can start to take automated actions based on that data instead of having to be reactive and a lot of manual discretionary effort, which just slows everything down.”— Michael Dongieux
Clean data benchmarking enables data center operators to gauge the performance of their equipment against industry peers. Metrics such as mean time to failure and maintenance and repair costs can be compared, providing valuable insights for performance optimization. Real-time access to benchmarking data empowers operators to make informed decisions promptly, replacing the need for lengthy reporting cycles.
Automation and Intelligence: Next-Gen Operating Systems
The power of an operating system lies in its ability to provide valuable insights and predictions. Rather than relying solely on traditional preventative maintenance schedules, which are time-based, a next-gen operating system enables a shift towards predictive maintenance. It would identify the precise components that require attention or replacement based on their actual condition, optimizing the utilization of resources and minimizing downtime.
“It’s going to look at all my data, all my alarms. It’s going to learn my data center, and it’s going to learn the environment it’s operating in, my client load, load profile, what it’s doing, and it’s going to learn about it, and it’s going to tell me things. It’s going to enable me to move from preventative maintenance, which is time-based to predictive maintenance.”— Ben Stewart
This operating system’s level of automation and intelligence is truly remarkable. For instance, when a component needs to be replaced, it would generate a detailed maintenance operation procedure (MOP) with accurate breaker labeling and step-by-step instructions specific to the data center. This eliminates the need for manual setup and ensures seamless maintenance procedures.
While this technology presents immense benefits, its implementation in the colocation industry poses additional challenges. Convincing clients of digital infrastructure advantages and the potential reduction in manpower is a crucial step. Improving efficiency, reducing downtime, and enhancing reliability will be vital in gaining client trust and support.
The Potential of Knowledge Models: Data Center Efficiency and Transformation
Many have embraced using chatbots powered by next generation models like GPT. These models revolutionize the way we interact and seek information. By scouring vast amounts of unstructured data, they provide remarkably informed responses, surpassing our traditional expectations for automation. With advancements like GPT-4, the capabilities of these models have taken a giant leap forward.
“If you take that concept and you apply it to what I would call a large knowledge model (instead of a large language model), based on all your data about your data center portfolio, think of the kind of insights and intelligence you’ll be able to get out of this.”— Michael Dongieux
This paradigm shift towards knowledge models marks a new era in data center management. By harnessing the power of data and advanced models, you can drive efficiency, make informed decisions, and propel your organization toward a future where data centers are optimized, intelligent, and transformative.
The Significance of Incremental Improvements and Risk Avoidance
The data center industry has traditionally focused on incremental improvements and risk avoidance to ensure reliability. However, over the past decade, the necessity for faster innovation has been driven by various factors.
In conclusion, there is a shift towards operational intelligence driven by the exponential growth of data centers and staffing challenges. AI and automation are promising solutions to optimize resource allocation and adapt to changing workforce dynamics. The rise of data center operating systems and clean data benchmarking enable operators to make data-driven decisions and improve performance. Additionally, implementing large knowledge models powered by advanced models like GPT, can unlock valuable insights and intelligence for enhanced operational efficiency. By embracing these advancements and transformative approaches, the data center industry can continue to thrive and meet the evolving demands of the digital era.
Listen to the full panel discussion here: