We are living an era of cyber-physical systems, where the physical and digital worlds are closely connected. While COVID-19 has affected most if not all of us, data centers have ameliorated its impact by keeping people connected and the economy afloat. The growing reliance on digital infrastructure has created an urgent need to intelligently monitor and manage data centers, improving their operational continuity, effectiveness, productivity, and sustainability while reducing their risk of failure.
Data centers are complex and dynamic environments. The advances in intelligent monitoring, robotics, and AI allow pattern recognition and behavior predictions almost immediately. It’s these three trends in data center management automation that reduce human interaction.
1) Intelligent Monitoring and Analytics
Intelligent monitoring tools employing data science and AI for analytics are rapidly gaining traction in the data center management arena. Data science, more than a technology, is a discipline that aggregates, normalizes, and interprets large amounts of data using tools in math, statistics, advanced computing, data engineering, and visualization, among others. It is being adopted in multiple processes across the data center industry to facilitate data interpretation and to perform actions based on previous knowledge and real-time analytics.
Figure 1 shows the cycle intelligent data center monitoring and management systems should undergo. The first step is data collection and aggregation, enabled through the use of IoT, sensing devices, and the equipment itself. Data collected must be reliable and aligned with the business objectives to ensure it’s relevant and can be understood. Real-time data collection can be a challenge, especially in legacy facilities that lack adequate instrumentation. The next step is data processing, leveraged by data science and/or AI-enabled analytics. These tools help to estimate parameters and comprehensive key indicators to better understand data center behavior, anticipate potential problems, and detect anomalies or vulnerabilities. And, the final step is using analytics for decision-making as well as creating action plans that address prioritized needs and challenges.
This is an iterative process, as the outcome of actions undertaken may modify data center behavior and predictions, opening up a window of opportunities for continuous improvement. Effective optimization strategies must include all assets to support end-to-end resource management, establishing an integrated approach toward performance improvement and risk mitigation.
2) Simulation Tools
Data center professionals are interested in quantitative predictions on never-before-seen data based on what happened in the past to evaluate strategies for achieving desired goals. That’s where the second trend comes in — the use of simulation tools incorporating AI and digital twinning. Simulation models can be based on theoretical models, which are often inaccurate or obsolete, or they can use AI to learn information directly from collected data without relying on predetermined equations. Simulations and predictions are used to better understand numerous scenarios.
Data collected by sensing devices and machine learning has been used to improve efficiency, reducing cooling consumption and overall emissions in data centers. There are multiple practical applications of AI in data centers, including behavior prediction; anomalies detection; intrusion detection; like-things identification; resource optimization; metrics and benchmarking; image, speech, and text recognition and classification; interaction personalization; fraud detection; risk assessment; proactive maintenance; video recognition; robotics; and navigation, among others.
Simulations are enhanced using digital twins, which are simulated 3D replicas of real data centers that analyze the current context. Digital twin technology is enabled by IoT and sensing devices gathering data in real time to monitor the facility and simulate the desired scenarios. For more rational decision-making, digital twins enhanced with AI not only offer real-time knowledge but can also make predictions based on assumptions about how different resources change or behave. It relies on a bidirectional, real-time connection between physical resources and digital twins to offer physical realism without compromise.
While fixed sensing devices coupled with IoT have proven to be effective for monitoring tools, data centers are demanding more mobile and interactive technologies to adequately automate monitoring and control the facility. So, it makes sense that the third emerging trend is the use of robotics. A robot is in charge of automating a specific process or repetitive task. Robotics perform tasks that workers would have to do otherwise, increasing process automation and productivity while reducing human error. We can also refer to software robots, digital robots, or nonphysical robots. Robots such as “chatbots” are able to notify and/or communicate with users on their computers or smartphones.
Robots have been used in data centers to perform tasks, such as storage retrieval, plugging and unplugging network connections, and destroying hardware, among others. Nonphysical robots are capable of extracting business insight out of large data sets for smart troubleshooting by creating better user interactions and experiences. Multiple experiments have been conducted with robots to navigate a facility and gather relevant data. AI-enabled intelligent robotics offer improved decision-making for tasks, such as collision avoidance, optimal resource deployment, and routes selection.
The global data center landscape is changing, and it requires more automation. To stay competitive, legacy data centers need to be upgraded with new technologies. This article explored three innovative trends that will ultimately contribute to more effective and efficient operations of unmanned data centers — a goal visualized long ago that has become much more pressing to achieve after the global pandemic.
New technologies have made the adoption of intelligent monitoring and analytics, simulation tools, and robotics a reality, and their implementation in mission critical facilities requires a reliable, well-orchestrated strategy to create business value. AI, considered immature by many, plays a key role in analyzing data and providing insights that humans are not capable to decipher. The adoption of these trends can be a game changer for the data center industry. They will transform the way facilities are managed by assessing designs or strategies prior to their implementation, enabling faster and more effective decision-making.