Riyanto, Nur Rahmat Dwi (2025) Optimasi Object Relational Mapping: Integrasi Profiling Kueri Berdasarkan Frekuensi dan Penerapan Timestamp dalam Algoritma LRU. Masters thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 6025222008-Master_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
6025222008-Master_Thesis.pdf Restricted to Repository staff only until 1 April 2027. Download (11MB) | Request a copy |
Abstract
Aplikasi Point of Sale dikembangkan dengan menggunakan pendekatan Object Relational Mapping untuk mengatasi ketidaksesuaian impedansi (impedance mismacth) antara paradigma pemrograman berbasis objek dan basis data relasional. Meskipun Object Relational Mapping (ORM) menawarkan keuntungan, performance overhead tetap menjadi perhatian karena penerjemahan operasi berorientasi objek ke dalam kueri SQL dan pengelolaan pemetaan objek relasional. Namun demikian belum banyak penelitian mempertimbangkan untuk mengatasi permasalahan overhead yang terjadi pada framework ORM. Beberapa penelitian telah melakukan optimasi pada framework ORM dengan mengusulkan pendekatan berdasarkan pembagian kerja untuk memisahkan thread, pengambilan keputusan yang tepat, partisi, pemetaan data dalam memori, pra-pemrosesan entitas, dan penghindaran produk Cartesian. Penelitian ini bertujuan untuk mengintegrasikan metode profiling kueri berdasarkan frekuensi serta strategi caching dengan penerapan timestamp dalam algoritma Least Recently Used (LRU) untuk meningkatkan performa dan memitigasi performance overhead pada kerangka kerja Object Relational Mapping. Metode yang diusulkan oleh penulis disebut sebagai OptiORM-LRU, yang diterapkan pada 5 dataset sistem point of sale yang diperoleh dari repositori GitHub. Hasil dari penerapan metode tersebut pada web api point of sale 5 dataset tersebut memiliki tingkat response time rata-rata lebih cepat 25.85% di bandingkan web api native dan 29.35% lebih cepat di bandingkan web api Object Relational Mapping. Sedangkan OptiORM-LRU memiliki rata-rata throughput lebih tinggi 0.02/s dibandingkan web api native dan 0.01/s lebih tinggi dibandingkan web api Object Relational Mapping.
=====================================================================================================================================
The Point of Sale application is developed using the Object-Relational Mapping (ORM) approach to address the impedance mismatch between object-oriented programming paradigms and relational databases. Although ORM offers significant advantages, performance overhead remains a concern due to the translation of object-oriented operations into SQL queries and the management of object-relational mappings. However, limited research has been conducted to specifically address the overhead issues associated with ORM frameworks. Some studies have proposed optimizations for ORM frameworks through approaches such as workload partitioning to separate threads, optimal decision-making, partitioning, in-memory data mapping, entity pre-processing, and Cartesian product avoidance. This study aims to integrate query profiling methods based on frequency and caching strategies by implementing timestamps within the Least Recently Used (LRU) algorithm to enhance performance and mitigate performance overhead in ORM frameworks. The proposed method, referred to as OptiORM-LRU, is applied to five POS system datasets obtained from GitHub repositories. The results of implementing this method on the web API of these five POS datasets demonstrate an average response time improvement of 25.85% compared to the native web API and 29.35% faster than the ORM-based web API. Furthermore, OptiORM-LRU achieves an average throughput that is 0.02 requests per second higher than the native web API and 0.01 requests per second higher than the ORM-based web API.
Actions (login required)
 |
View Item |
