By Jagat Ram, Head – Data Center Operation and Delivery, L&T
Automation is no longer a futuristic concept—it’s a necessity for modern data centers. As we manage increasingly complex infrastructures across multiple locations, power densities, and technologies like liquid cooling, the need for intelligent automation has never been greater.
At L&T, and across the industry, we’re witnessing a shift from manual oversight to autonomous systems capable of self-optimization and self-healing. This transformation, however, must be approached in a structured and strategic manner.
The Three Layers of Automation
When we look at automation opportunities in data centers, I see three core layers where technology can make a transformative difference:
Operational Layer: This is where we manage day-to-day operations—monitoring systems, analyzing performance, and ensuring uptime. Automation here can drive consistency, efficiency, and faster response to anomalies.
Infrastructure Layer: Covering servers, storage, power, and cooling systems, this layer is critical for ensuring stability and performance. Automation can reduce human intervention, eliminate errors, and improve resilience.
Business Process Layer: At the top lies the interface with customers—CRM systems, chatbots, and digital workflows that enhance service delivery and responsiveness.
It’s important to clarify: automation isn’t about replacing people. It’s about enhancing efficiency, resilience, and scalability within the ecosystem.
The Evolution Toward Autonomy
True autonomy doesn’t happen overnight. It’s a three-phase evolution:
Phase 1: Reactive to Assisted Operations – Systems operate under manual and rule-based automation.
Phase 2: Semi-Autonomous Operations – Predictive AI and self-healing capabilities start emerging, helping us anticipate failures and optimize proactively.
Phase 3: Fully Autonomous Operations – An AI-driven ecosystem where systems make independent decisions to self-optimize and self-heal, ensuring seamless continuity.
The Roadmap to an Autonomous Data Center
In my view, building an autonomous data center involves a disciplined, step-by-step approach:
Assess the Current State: Understand where you stand in terms of uptime, PUE/WUE, energy optimization, workforce efficiency, and regulatory or ESG compliance.
Build a Unified Data Foundation: Bring together fragmented data across operations into a single source of truth.
Deploy IoT Sensors: Capture key parameters such as temperature, humidity, vibration, and pressure to enable real-time visibility and predictive maintenance.
Integrate IT and OT Systems: Connect the IT layer (compute, storage, and network) with OT systems like BMS, SCADA, and EPMS. Use open protocols such as SNMP, Modbus, BACnet, and MQTT to standardize data collection.
Implement Centralized Monitoring and Digital Twins: Digital twins create a virtual replica of the data center, allowing us to simulate scenarios, predict failures, and take preventive actions.
Automate Key Operations: Start small—with low-risk, high-impact areas such as AI-based dynamic cooling, UPS/genset automation, and routine patching—and expand progressively.
IT-OT Convergence with Cyber Isolation
Integrating IT and OT is crucial to achieving predictive optimization. However, it must be done securely. At L&T, we emphasize firewall-based isolation to ensure that while systems share intelligence, critical components remain protected.
This integration is essential, especially as data center customers demand deeper visibility and real-time alerts—capabilities that require seamless interaction between IT and OT environments.
Cloud or On-Prem? Balancing Efficiency and Security
A common question I get is whether AI/ML predictive systems should be cloud-based or hosted on-premises. My view:
Cloud-based platforms are cost-effective and ideal for pilots or early-stage implementations. They allow you to start small, validate use cases, and scale as needed.
On-premises platforms offer tighter control over data, ensuring better security and compliance—particularly for sensitive or mission-critical workloads.
Many organizations adopt a hybrid approach—starting with cloud-based proof-of-concepts and gradually moving mature models in-house.
Looking Ahead: From Automation to Intelligence
Technologies like digital twins, AI/ML engines, PLCs, SCADA systems, and virtual assistants are shaping the next frontier of data center management. The vision is clear—to evolve from automated operations to intelligent, self-governing ecosystems.
Automation today is not just about speed; it’s about intelligence. A truly autonomous data center will not only detect and resolve issues but continuously learn and optimize—delivering reliability, sustainability, and efficiency at an unprecedented scale.
As I often say, “Automation is not about replacing human capability—it’s about amplifying it. The future belongs to data centers that can think, adapt, and evolve.”