إطار عمل ذكي في الزمن الحقيقي لإدارة حركة مرور البيانات الضخمة استناداً إلى الحوسبة الضبابية

الباحث

هشام هاشم محمد

المشرف

ضحى بشير عبدالله البزاز

الجامعة

جامعة الموصل

الكلية / القسم

كلية علوم الحاسوب والرياضيات / علوم الحاسوب

الشهادة

دكتوراه

السنة

2022

لغة الرسالة

انكليزي

الكلمات المفتاحية

Fog Computing ; Big Data ; Traffic Management ; Real-Time Framework

Cloud computing regarded as the most powerful computing architecture in the distributed computing field. However, it is not suitable for some type of applications such as delay-sensitive and computation intensive applications. When user device encounter computation and/storage intensive application it may offload it, either partially or entirely, to a third resourceful entity. Offloading tasks from the users' devices to the cloud will take a relatively long time and consumes network bandwidth. This motivates the appearance of fog computing. Fog computing consists of three-tier architecture these tiers, from bottom to top, are: user devices, fog layer, and cloud layer. As a result of the fog layer's closeness to the user device, offloading tasks to it will be faster than offloading them to a remote cloud, and this will also save the network bandwidth. Fog computing has spread widely in last decade, however it is challenging to build and test fog computing-based systems in the real world. This prompted the developers to simulate and test their systems using fog simulation frameworks. In this thesis, a four-tiered fog computing framework with the name DHFogSim has been suggested. DHFogSim is a smart real time framework built for managing traffic of any fog computing system. Fog computing system, especially the real time one; produces a huge number of traffic during its life cycle. Such traffic needs to be handled carefully in order to achieve two related goals. The first one is to reduce the delay for user to increase system Quality of Service QoS, and the second is to distribute the workload over working nodes to utilize the resources as much as possible. DHFogSim adds a smart agent layer between the users' devices layer and the fog computing layer. At the Ethernet layer, DHFogSim proposes and implements the use of multi-level queues rather than a single queue; these queues are managed using a weighted round robin mechanism. Tasks are dispatched to various queues based on the value of the Type of Service (ToS) bits, which are located at the second byte of the IP header. The value of ToS bits is determined by the smart agent layer based on the task's time constraints. To evaluate the proposed DHFogSim framework, a traffic management system was simulated and evaluated on both the mFogSim framework and the DHFogSim framework. According to the results, the suggested framework reduces task delay at fog brokers from (0.5450 ms) to (0.3675 ms). The task delay at fog nodes has been reduced from (0.5482) to (0.3626). Task latency also decreased from (0.3515 ms) to (0.3226 ms). Furthermore, mean packet drop count is slightly improved form (0.1957) to (0.1772). DHFogSim is open source and available at the following site: https://github.com/HeahamShakarchi/DHFogSim

طلب نسخة من الملف

الملفات غير منشورة على الموقع. اترك بياناتك وسيتواصل معك موظفو قسم الشؤون الثقافية والفنون لإرسال النسخة مجاناً.

الاسم

رقم الهاتف

البريد الإلكتروني

الغرض من الطلب (اختياري)

إرسال الطلب