Learning-based Two-tiered Online Optimization of Region-wide Datacenter Resource Allocation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916442892402688 |
|---|---|
| author | Chen, Chang-Lin Zhou, Hanhan Chen, Jiayu Pedramfar, Mohammad Lan, Tian Zhu, Zheqing Zhou, Chi Ruiz, Pol Mauri Kumar, Neeraj Dong, Hongbo Aggarwal, Vaneet |
| author_facet | Chen, Chang-Lin Zhou, Hanhan Chen, Jiayu Pedramfar, Mohammad Lan, Tian Zhu, Zheqing Zhou, Chi Ruiz, Pol Mauri Kumar, Neeraj Dong, Hongbo Aggarwal, Vaneet |
| contents | Online optimization of resource management for large-scale data centers and infrastructures to meet dynamic capacity reservation demands and various practical constraints (e.g., feasibility and robustness) is a very challenging problem. Mixed Integer Programming (MIP) approaches suffer from recognized limitations in such a dynamic environment, while learning-based approaches may face with prohibitively large state/action spaces. To this end, this paper presents a novel two-tiered online optimization to enable a learning-based Resource Allowance System (RAS). To solve optimal server-to-reservation assignment in RAS in an online fashion, the proposed solution leverages a reinforcement learning (RL) agent to make high-level decisions, e.g., how much resource to select from the Main Switch Boards (MSBs), and then a low-level Mixed Integer Linear Programming (MILP) solver to generate the local server-to-reservation mapping, conditioned on the RL decisions. We take into account fault tolerance, server movement minimization, and network affinity requirements and apply the proposed solution to large-scale RAS problems. To provide interpretability, we further train a decision tree model to explain the learned policies and to prune unreasonable corner cases at the low-level MILP solver, resulting in further performance improvement. Extensive evaluations show that our two-tiered solution outperforms baselines such as pure MIP solver by over $15\%$ while delivering $100\times$ speedup in computation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2306_17054 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Learning-based Two-tiered Online Optimization of Region-wide Datacenter Resource Allocation Chen, Chang-Lin Zhou, Hanhan Chen, Jiayu Pedramfar, Mohammad Lan, Tian Zhu, Zheqing Zhou, Chi Ruiz, Pol Mauri Kumar, Neeraj Dong, Hongbo Aggarwal, Vaneet Networking and Internet Architecture Online optimization of resource management for large-scale data centers and infrastructures to meet dynamic capacity reservation demands and various practical constraints (e.g., feasibility and robustness) is a very challenging problem. Mixed Integer Programming (MIP) approaches suffer from recognized limitations in such a dynamic environment, while learning-based approaches may face with prohibitively large state/action spaces. To this end, this paper presents a novel two-tiered online optimization to enable a learning-based Resource Allowance System (RAS). To solve optimal server-to-reservation assignment in RAS in an online fashion, the proposed solution leverages a reinforcement learning (RL) agent to make high-level decisions, e.g., how much resource to select from the Main Switch Boards (MSBs), and then a low-level Mixed Integer Linear Programming (MILP) solver to generate the local server-to-reservation mapping, conditioned on the RL decisions. We take into account fault tolerance, server movement minimization, and network affinity requirements and apply the proposed solution to large-scale RAS problems. To provide interpretability, we further train a decision tree model to explain the learned policies and to prune unreasonable corner cases at the low-level MILP solver, resulting in further performance improvement. Extensive evaluations show that our two-tiered solution outperforms baselines such as pure MIP solver by over $15\%$ while delivering $100\times$ speedup in computation. |
| title | Learning-based Two-tiered Online Optimization of Region-wide Datacenter Resource Allocation |
| topic | Networking and Internet Architecture |
| url | https://arxiv.org/abs/2306.17054 |