Browsing by Author "Umesh Bhatt"

Now showing 1 - 3 of 3

On Designing a Compaction Strategy for Enhanced LevelDB Read Performance
(Association for Computing Machinery, Inc, 2025) Umesh Bhatt; Sarvesh Pandey
Key-value stores based on a Log-Structured Merge-Tree, especially LevelDB, are extensively used for managing unstructured data with no predefined schema. LevelDB’s popularity stems from an optimized write performance, achieved through in-memory buffers, sequential disk writes, and background compaction. Although the write performance is satisfactory, the read performance is, at best, marginally acceptable, particularly for random (non-contiguous) reads. Read performance can be enhanced by optimizing the compaction strategy (i.e., the trade-off between the number of levels and the number of tables per level). LevelDB utilizes geometric progression to determine compaction table size, which often leads to increased read latency due to the inefficient distribution of compaction tables across levels. We propose utilizing power law distribution to determine compaction table size for each level, which in turn improves read efficiency. Our empirical results show a reduction in average query runtime (up to 2%), an increase in read throughput (up to 2.5%), and a decrease in average latency (up to 2%) on YCSB’s read-only workload. © 2024 Copyright held by the owner/author(s).
On Designing an Intelligent Shipping Algorithm for Decentralized E-Commerce Systems
(Springer Science and Business Media Deutschland GmbH, 2024) Suneel Kumar; Sarvesh Pandey; Umesh Bhatt
The foundation of online shopping lies in the efficient and timely delivery of products from sellers to customers directly through a streamlined (and user-friendly) digital platform. However, transitioning to the online shopping platform leads to the following two problems: higher shipping charges and product handling fees levied by these platforms. The shipping charge is typically exempted if the order value is equal to or higher than the predetermined threshold value. The existing shipping charge exemption rule does not favor customers with low and mid-range budgets who often place orders valued lower than the threshold. To address the inherent biasness, we propose the History Informed Shipping (HIShip) method, establishing a fair business environment for all parties involved – the online shopping platform provider, sellers, and customers. HIShip intelligently utilizes the order history data, i.e., the cumulative sum of orders’ value placed in the recent past, to make the shipping charge exemption rule friendly to low and mid-budget customers. Furthermore, it reduces biasness against vendors selling products that cost less than the threshold amount. Such a win-win scenario for the seller and customer eventually generates more revenue for the online shopping platform. We simulate the blockchain environment and use the TPC-H dataset to assess the performance. Our algorithm outperforms the threshold-based traditional approach. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
On Enhancing E-Commerce Shipping Policies with Blockchain and Recommender Systems
(Springer, 2025) S. Dasaratha Kumar; Sarvesh Pandey; Umesh Bhatt
E-commerce systems aim to deliver products on time and at a competitive price to customers through the Internet. Though transformational, adapting to an internet-based online system led to higher shipping charges (being borne by customers) and overwhelming options (requiring endless time to make purchasing decisions). Most existing shipping policies exempt the shipping fee only when the customer’s order value exceeds a pre-set threshold; they do not consider the frequency of orders made by a customer when deciding on a shipping fee exemption. First, to address the shipping charge problem, we propose a History Informed Shipping (HIShip) method, which utilizes the customer’s transaction history in making the shipping charge exemption decisions. HIShip mainly benefits low (and mid) order-value customers who frequently order and sellers with a product cost lower than the pre-defined shipping exemption threshold amount. The greater customer and seller participation eventually contribute to higher revenue from the e-commerce platform. Furthermore, we store order history in the blockchain to ensure decentralization and immutability in a trustless environment. HIShip’s shipping policy is evaluated against a naive threshold-based shipping policy on the TPC-H dataset, and results confirm that 21.5% and 21.06% increase in the percentage of orders (placed by low and mid ’order-value’ customers, respectively) qualify for the shipping fee exemption. Second, we integrated an ML-based recommendation mechanism to suggest appropriate product(s) further in case the actual order does not qualify for shipping fee exemption. Evaluation results show that SVD is the best model with a minimum RMSE of 0.765364 and MAE of 0.508519 on the ‘All Beauty’ dataset. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.