ImpalaE: Towards an optimal policy for efficient resource management at the edge
Edge computing is an extension of cloud computing where physical servers are deployed closer to the users in order to reduce latency. Edge data centers face the challenge of serving a continuously increasing number of applications with a reduced capacity compared to traditional data center. This pa...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Academy of Cognitive and Natural Sciences
2022-11-01
|
| Series: | Journal of Edge Computing |
| Subjects: | |
| Online Access: | https://acnsci.org/journal/index.php/jec/article/view/572 |
| _version_ | 1797361806266597376 |
|---|---|
| author | Tania Lorido-Botran Muhammad Khurram Bhatti |
| author_facet | Tania Lorido-Botran Muhammad Khurram Bhatti |
| author_sort | Tania Lorido-Botran |
| collection | DOAJ |
| description |
Edge computing is an extension of cloud computing where physical servers are deployed closer to the users in order to reduce latency. Edge data centers face the challenge of serving a continuously increasing number of applications with a reduced capacity compared to traditional data center. This paper introduces ImpalaE, an agent based on Deep Reinforcement Learning that aims at optimizing the resource usage in edge data centers. First, it proposes modeling the problem as a Markov Decision Process, with two optimization objectives: reducing the number of physical servers used and maximize number of applications placed in the data center. Second, it introduces an agent based on Proximal Policy Optimization, for finding the optimal consolidation policy, and an asynchronous architecture with multiple workers-shared learner that enables for faster convergence, even with reduced amount of data. We show the potential in a simulated edge data center scenario with different VM sizes based on Microsoft Azure real traces, considering CPU, memory, disk and network requirements. Experiments show that ImpalaE effectively increases the number of VMs that can be placed per episode and that it quickly converges to an optimal policy.
|
| first_indexed | 2024-03-08T15:59:40Z |
| format | Article |
| id | doaj.art-8a1ea259ba654873aff218d81b22ed46 |
| institution | Directory Open Access Journal |
| issn | 2837-181X |
| language | English |
| last_indexed | 2024-03-08T15:59:40Z |
| publishDate | 2022-11-01 |
| publisher | Academy of Cognitive and Natural Sciences |
| record_format | Article |
| series | Journal of Edge Computing |
| spelling | doaj.art-8a1ea259ba654873aff218d81b22ed462024-01-08T12:04:11ZengAcademy of Cognitive and Natural SciencesJournal of Edge Computing2837-181X2022-11-011110.55056/jec.572ImpalaE: Towards an optimal policy for efficient resource management at the edgeTania Lorido-Botran0Muhammad Khurram Bhatti1RobloxInformation Technology University Edge computing is an extension of cloud computing where physical servers are deployed closer to the users in order to reduce latency. Edge data centers face the challenge of serving a continuously increasing number of applications with a reduced capacity compared to traditional data center. This paper introduces ImpalaE, an agent based on Deep Reinforcement Learning that aims at optimizing the resource usage in edge data centers. First, it proposes modeling the problem as a Markov Decision Process, with two optimization objectives: reducing the number of physical servers used and maximize number of applications placed in the data center. Second, it introduces an agent based on Proximal Policy Optimization, for finding the optimal consolidation policy, and an asynchronous architecture with multiple workers-shared learner that enables for faster convergence, even with reduced amount of data. We show the potential in a simulated edge data center scenario with different VM sizes based on Microsoft Azure real traces, considering CPU, memory, disk and network requirements. Experiments show that ImpalaE effectively increases the number of VMs that can be placed per episode and that it quickly converges to an optimal policy. https://acnsci.org/journal/index.php/jec/article/view/572edge computingpolicy gradientreinforcement learningefficient resource management |
| spellingShingle | Tania Lorido-Botran Muhammad Khurram Bhatti ImpalaE: Towards an optimal policy for efficient resource management at the edge Journal of Edge Computing edge computing policy gradient reinforcement learning efficient resource management |
| title | ImpalaE: Towards an optimal policy for efficient resource management at the edge |
| title_full | ImpalaE: Towards an optimal policy for efficient resource management at the edge |
| title_fullStr | ImpalaE: Towards an optimal policy for efficient resource management at the edge |
| title_full_unstemmed | ImpalaE: Towards an optimal policy for efficient resource management at the edge |
| title_short | ImpalaE: Towards an optimal policy for efficient resource management at the edge |
| title_sort | impalae towards an optimal policy for efficient resource management at the edge |
| topic | edge computing policy gradient reinforcement learning efficient resource management |
| url | https://acnsci.org/journal/index.php/jec/article/view/572 |
| work_keys_str_mv | AT tanialoridobotran impalaetowardsanoptimalpolicyforefficientresourcemanagementattheedge AT muhammadkhurrambhatti impalaetowardsanoptimalpolicyforefficientresourcemanagementattheedge |