Tools
Tools
Somavaram, Vijayanand (2023) An Investigation Into Performance Efficiencies In AWS DynamoDB Configurations For Various Serverless Application Workloads. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (1MB) | Preview |
Preview |
PDF (Configuration Manual)
Download (2MB) | Preview |
Abstract
NoSQL database systems are being widely used and undoubtedly gained a lot of popularity in the last decade or so. Many factors affect the performance of the NoSQL database system. Among these factors, the database system’s throughput settings, including the read and write capacity units, can affect its performance. In many NoSQL databases, especially in cloud-based systems like Amazon DynamoDB, Azure Cosmos DB, or Google Cloud Big-table, users must provision the read and write capacity units. These systems allow users to manually provision the capacity units and provide services to auto-manage the throughput settings. This research studies how these capacity units affect the performance of database systems and focuses on identifying the optimal provisioned throughput settings considering the incoming traffic requests with predictable payloads, to eventually optimize the performance of serverless environments that use these databases. If these values are under-provisioned it can lead to increased latency, throttling, and poor performance during high-traffic periods. If over-provisioned, it can lead to unnecessary costs. As part of this research, AWS DynamoDB which is one of the prominent NoSQL databases was considered. This study is about evaluating the performance of DynamoDB under the default settings provided by AWS and investigating whether these settings could be improved further to assess if latencies can be reduced even further. The experimental results indicate that read-intensive workloads showed improvement in latencies through additional capacity provisioning. However, write-intensive workloads did not exhibit improvement with additional capacity provisioning. This research utilized data collected from tests conducted and developed a machine-learning model that predicts the optimal read capacity units for requests under predictable workloads.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Kazmi, Aqeel UNSPECIFIED |
| Subjects: | Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Cloud computing Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning |
| Divisions: | School of Computing > Master of Science in Cloud Computing |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 11 Apr 2025 09:46 |
| Last Modified: | 11 Apr 2025 09:46 |
| URI: | https://norma.ncirl.ie/id/eprint/7417 |
Actions (login required)
 |
View Item |