Browsing by Author "Sarker, Debashish"

Now showing 1 - 3 of 3

ENDOSCOPIC DIAGNOSIS AND THERAPEUTIC APPROACHES OF THE UPPER DIGESTIVE TRACT IN ANIMALS
(Chittagong Veterinary and Animal Sciences University Chittagong – 4225, Bangladesh, 2022-06) Sarker, Debashish
Flexible endoscopy is a minimally invasive technique of the visualization, investigation and biopsy of gastrointestinal (GI) tract. Practicing gastroscopy in small animals is still apparently new in Bangladesh. Diagnostic implications include the evaluation of structural abnormalities, inflammatory conditions, intraluminal masses, injuries, and foreign bodies (FB). Due to difficulties of visual examination of the upper GI tract, various diseases of the GI system remain undiagnosed. The aim of this study was to diagnosis and management of complications in upper GI tract by minimally invasive method in appropriate anesthesia. In addition, to determine the digestive health of an animal compared with its history and physical examination. The present study was conducted on 30 animals (10 goats, 10 dogs and 10 cats) during July 2021 to March 2022. Depending on the conditions of the animal, all goats (n=10) were sedated with diazepam while, most of the dogs (n=8) and few number of cats (n=4) were gone for general anesthesia with xylazine premedication and few dogs (n=2) and most cats (n=6) were without premedication. Ketamine was administered in a dog and four cats, a combination of ketamine with diazepam was used in six dogs and a cat and propofol was administered in three dogs and five cats. The fasting duration in goats, the time of hospitalization and procedural time in dogs and cats were statistically significant (P≤0.05) in subgroups. Gastroscopy broadly 80% of goats (n=8), 40% of dogs (n=4) and 80% of cats (n=8) had normal ruminal or gastric mucosal appearance; 20% of goats had abnormal ruminal nature, mild to moderate gastritis obtained on 50% of dogs and 10% of cats, and severe gastritis documented on 10% of dogs and cats. Therapeutically, 80% of the FB in cats (n=4/5) were successfully retrieved by endoscope. On endoscopic examination, 30% of dogs and 50% of cats were diagnosed as healthy with history of anorexia or FB obstruction while, 20% of dogs had gastritis without clinical illness. All animals returned to its normal behavior with minimum difficulties.
Scenario of Waste Management of Poultry in Bagmara Upazila, Rajshahi
(Chattogram Veterinary and Animal Science University, 2019-06) Sarker, Debashish
The study attempts to determine the scenario of waste and biosecurity management of poultry farms located in Bagmara upazila, Rajshahi, Bangladesh. The study was conducted in 8th to 28th December 2018 at two village in Bagmara upazila. A total number of 31 commercial poultry farms was considered for the study. Each farm was visited individually, and a structural questionnaire was used to collect data related to poultry waste management system and biosecurity management in order to determine the public health concern surrounding farms. In this study, the percentage of broiler, layer and turkey was 9.68, 87.10, 3.23, respectively. In those farms 93.55% farm owner used poultry litter as fish feed whereas, 3.23 and 3.23% farm owner used litter as fertilizer and biogas. Majority of the farmer (87.10%) used cage rearing which is followed by shelf rearing (9.68 %) and floor rearing (3.23 %). About 97.77% farm had no foot bath. Most of the farms (77.24%) caused the water pollution while 54.84% farms produced noise and smell. About 83.7% farms used chemical sanitizer for disinfection purpose. Only 35.5% farms properly managed their dead birds and eggs. Most of the farms were surrounded by long tree and bush except 19.35% of farms. In conclusion, the litter and waste management system and biosecurity procedures in the study farms were not adequate. Poultry farm owners should undertake proper waste management systems to reduce the environment and public health risk from poultry waste.
Using machine learning to predict optimal erasure coding policies for object storage system in OpenStack Swift
(BRAC University, 2026-01) Ankon, Amio Malakar; Chaki, Boloy; Sayan, Mashrur Shakhawat; Sarker, Debashish; Mukta, Jannatun Noor
Erasure coding helps to reduce storage overhead and improve fault tolerance. But the procedure to select an appropriate erasure coding policy is complex, which often involves tradeoffs among various metrics such as access latency, recovery behavior, storage efficiency, etc. Generally, these are handled using static or heuristic-based configurations, which can not account for variations in workload. In the industry, service providers like Ceph do benchmarking based on throughput/latency without considering workload diversity. OpenStack Swift, one of the most widely used open-source object storage systems, supports erasure coding, but it allocates policy selection in a manual, static way - without it being workload-aware. To address this problem, this thesis showcases a data-driven performance modeling framework for erasure coding in object storage systems using machine learning. A structured dataset- ABDS-30k was constructed in a controlled execution of varying workload conditions. It was created on a Swift All-In-One (SAIO) testbed under different erasure coding policies with the addition of failure injections. The dataset collects several empirically observed performance metrics such as read and write latency, tail latency, success rate, and reconstruction time in case of disk failures. For the job of selecting the optimal erasure coding policy, we adopt two distinct Machine Learning paradigms - a regression-based approach of performance modeling and a classification-based approach for direct data-driven policy recommendation. For regression, CatBoost, XGBoost, and Random Forest Regression are used to predict target metrics in a given workload context and failure scenario, and the optimal policy is chosen based on a weighted score. In the classification-based approach, CatBoostClassifier, XGBoostClassifier, and Logistic Regression are used to label workload–policy pairs using an oracle cost function, and the most optimal erasure coding policy is directly recommended. Experimental results show that the regression models have moderate prediction errors for latency and recovery metrics because of the inherently noisy and heavy-tailed nature of the system, but they remain effective in optimal policy recommendation by exhibiting top-1 accuracy with 48.16% in XGBoost Regression and a top-3 accuracy 100% in Random Forest Regression model. Also, the regret mean (0.010 - 1.81) and regret median (0.00216 - 0.089) values showed a very low margin of error. Classification-based approach shows comparatively weaker metrics, indicating that it is not optimal for data-driven policy recommendation. Overall, this shows the feasibility of using ML-based performance modeling as opposed to static and heuristic-based policy selection, which can be later used as a foundation for future control-plane automations.