Optimized planning for managed wireless network management

Raghunandan, Arpitha

Permalink

https://hdl.handle.net/2142/120372 Copy

Description

Title

Optimized planning for managed wireless network management

Author(s)

Raghunandan, Arpitha

Issue Date

2023-04-17

Director of Research (if dissertation) or Advisor (if thesis)

Caesar, Matthew

Department of Study

Computer Science

Discipline

Computer Science

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• Managed Wireless Networks
• RAN
• AI Planning
• Constraint Programming
• Management
• Synthesis
• Sequence of Actions
• Optimization
• Modeling

Abstract

A significant majority of modern wireless networks are centralized, composed of a set of base stations operated by a single administrative entity with a wired backhaul, referred to as managed wireless networks. Their underlying infrastructure is growing ever more complex, which in turn makes management tasks difficult. Constituent protocols and systems must work in concert to perform actions, yet may suffer from complex and conflicting interdependencies between management goals and operational steps. Today, operators deal with these challenges through manually maintained playbooks, which contain instructions to be followed when a change to the network is required or observed. However, the instructions in these playbooks, whether automated or human-managed, typically consider management goals individually rather than jointly for optimizing network-wide operations and thus can lead to degraded performance, increased operational expense, and proneness to failure. In this thesis, we explore the possibility of generating an automated playbook to synthesize an optimal sequence of operational steps for efficient management of managed wireless networks while ensuring scalability. We leverage AI-based planning techniques to synthesize efficient system-wide management operations for large-scale managed wireless networks. To improve scalability, (a) we adopt a greedy constraint optimization approach to planning to guide the search process (b) develop compact formal representations of diverse network components and operational actions. We provide concrete use cases of our approach through simulation-based evaluation with Radio Access Networks (RAN), considering some representative management tasks. We find that our approach can compute plans for networks of a few hundred base stations within a few minutes.

Graduation Semester

2023-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/120372

Copyright and License Information

Copyright 2023 Arpitha Raghunandan

Owning Collections