IILM 2026

Every cluster in production today exists because someone decided to build it, how many nodes, how much network fabric, how much storage, and when. That planning decision is as technically demanding as the systems research that follows it, yet receives a fraction of the scholarly attention. IILM 2026 fills that gap, focusing on the intelligence layer that drives infrastructure provisioning: how organizations predict demand, model constraints, automate workflows, and make high-stakes capacity decisions at hyperscale.

At scale, capacity planning means reconciling demand signals from telemetry, sales pipelines, and supply chains against hard constraints on power, cooling, and procurement — all under uncertainty. The cost of miscalibration is severe: over-provisioning strands capital; under-provisioning degrades throughput for months. Most organizations still rely on fragmented spreadsheets and siloed legacy systems. IILM 2026 brings together the researchers and practitioners building what comes next.

The workshop draws on original research and production deployment experience at hyperscale, and the design and deployment of an enterprise-scale AI-augmented planning platform managing capacity decisions across a global network build program. It is structured to transfer that operational knowledge into the research community while simultaneously surfacing open problems that practitioners cannot yet solve. The result is a workshop where academic rigor and production reality are in active dialogue throughout.

Target Audience:

• Infrastructure architects and capacity planning leads at hyperscale cloud providers, national labs, and research computing centers

• Technical program managers responsible for large-scale cluster provisioning and network deployment lifecycles

• Applied ML researchers developing forecasting, optimization, and anomaly detection systems for infrastructure operations

• Platform engineers building unified tooling to replace fragmented legacy planning systems at organizational scale

• Systems researchers at the intersection of cluster management, operations research, and applied AI

• IEEE members exploring production-grade AI deployment in operational infrastructure contexts

Topics of Interest:

The workshop solicits original research contributions and practitioner experience papers on the following topics:

AI/ML for Capacity Demand Forecasting

• Time-series models for cluster resource demand prediction across 3–18 month horizons (LSTM, Transformers, ensemble methods)

• Multi-source demand signal fusion: integrating telemetry, workload fingerprints, sales pipelines, and product roadmaps

• Uncertainty quantification and confidence-bounded forecasting for high-stakes provisioning decisions

• Evaluation frameworks and ground truth methodology for infrastructure demand forecasting

Constraint-Aware Provisioning and Optimization

• Reinforcement learning for infrastructure provisioning under power, cooling, procurement, and topology constraints

• Multi-objective optimization: balancing capital cost, redundancy, latency, and build lead times at hyperscale

• Scenario simulation and what-if analysis frameworks for capacity investment decisions

Workflow Automation and Platform Engineering

• Unified planning platform architecture: consolidating fragmented legacy tools into AI-augmented enterprise systems

• End-to-end workflow automation from capacity signal to build order to deployment tracking

• Legacy modernization patterns: retiring spreadsheet-era planning tools at organizational scale

• Multi-tenant platform design for cross-functional planning across network, compute, and supply chain teams

Anomaly Detection and Operational Risk

• Early warning systems for capacity saturation, demand spikes, and supply chain disruption signals

• Root cause analysis using ML on infrastructure telemetry and operational event logs

• Adaptive capacity response: automated and semi-automated remediation workflows

Human-in-the-Loop Systems and Governance

• Explainable AI for capacity planning recommendations in operator-facing decision support tools

• Override mechanisms, audit trails, and accountability frameworks for AI-driven provisioning decisions

• Organizational change management: adoption patterns for AI-assisted planning in large infrastructure teams

Keywords: Infrastructure lifecycle management  •  Capacity planning  •  AI/ML for infrastructure  •  Hyperscale systems  •  Demand forecasting  •  Constraint-aware optimization  •  Workflow automation  •  Platform engineering  •  Reinforcement learning  •  Explainable AI  •  Cluster management  •  Anomaly detection

Deadlines :

Paper Submission Deadline

July 17, 2026

Notification of Acceptance

July 24, 2026

Camera-Ready Deadline 

August 3, 2026

Workshop Date

September 22, 2026

Paper Submission:

Click here to submit papers on Easy Chair

5-7 Pages, IEEE double-column format

Click here for IEEE Author Guidelines & Policies for Paper Submission

Call for Reviewers:

Please fill the form to volunteer as a reviewer to help us ensure the quality of the papers presented at this special session.