Throughput-Optimal Cross-Layer era
Leandros Tassiulas and Anastasios Ephremides laid the foundation for throughput-optimal cross-layer control in multi-hop networks by developing the backpressure (max-weight) scheduling framework, proving stability under general arrival processes and heterogeneous traffic. This era also saw the receiver-initiated soft-state reservation framework exemplified by RSVP, with Braden, Zhang, and Berson among the primary designers, enabling scalable QoS provisioning through soft state across hops. The Generalized Processor Sharing GPS model for fair service-rate sharing and QoS constraints was advanced by S. Parekh and R. Gallager, providing a rigorous queuing-theory basis for proportional resource allocation in shared links. Together, these strands—backpressure stability theory, RSVP-based soft-state provisioning, and GPS-based fair queuing—translated theoretical throughput guarantees into practical provisioning metrics for heterogeneous, multi-class networks through 2008.
Energy-Efficient Orchestration era
Rajkumar Buyya [1] is a leading figure active across Carnegie Mellon University [3] and National University of Singapore [4] during this era. Buyya [1] contributed optimal online deterministic algorithms and adaptive heuristics for energy and performance efficient dynamic consolidation of virtual machines in Cloud data centers [6], a cornerstone for energy-aware orchestration in the Energy-Efficient Orchestration era. Rui Zhang [2] is a prominent researcher affiliated with Duke University [5] and National University of Singapore [4] during this era. Zhang [2] advanced Throughput Maximization for UAV-Enabled Mobile Relaying Systems [7], providing design guidance for high-throughput and energy-efficient wireless backhaul in heterogeneous networks. Learning-Driven Edge Offloading era
Representative authors in this era include Y. Mao, C. You, and K. Zhang, whose early mobile edge computing work shaped resource optimization through joint offloading, edge-cloud orchestration, and cross-layer design. Mao and coauthors grounded the field with models comparing latency, energy, and transmission constraints, articulating multi-objective optimization frameworks that would be later addressed by learning-based methods. C. You contributed scalable, cross-layer optimization approaches and architectures that accommodate heterogeneous devices and stochastic workloads, enabling adaptive offloading policies. K. Zhang contributed theoretical and algorithmic insights into scheduling and resource allocation at the network edge, laying groundwork that inspired reinforcement-learning driven offloading and energy-latency trade-off solutions in the 2018–2024 era.