Loading...
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 57149 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Habibi, Jafar
- Abstract:
- Content delivery over the Internet is limited by the distance between the website or application server and the end user, so the further the server is from the end user, the longer it takes for the website content to load. Content delivery servers play an important role in improving the performance of Internet services by placing the content needed by users in cached servers close to users and reducing the distance that data must travel. In a content delivery network, a large number of geographically distributed edge servers are deployed, and whenever a user requests a content, the desired content is downloaded from the nearest edge server. Currently, more than half of the internet traffic reaches the users through these networks, which shows the importance and necessity of these networks in the ecosystem of providing broadband services, and cost reduction along with providing service quality and optimizing the response time to users’ requests are important. Most of the researches related to the optimization of content delivery networks do not use the big data available in the network edge servers. The reason for this is the large volume of data and their meaning at the moment. Due to the high volume as well as changes in the input load of requests in a short period of time, flow analysis methods should be used to create general knowledge about the network. In this research, the use of flow analysis to optimize content delivery networks is investigated. Flow analytics provides a real-time, data-driven approach to optimizing content delivery networks by analyzing network traffic and user behavior patterns. In this research, the use of deep learning algorithms and data processing techniques to analyze the data flow on different network nodes and find relevant parameters in creating general knowledge of the system, aggregating these parameters in real time to create general knowledge about the content delivery network and The use of this cumulative knowledge created in improving the performance of the content delivery network in delivering content to end users is investigated
- Keywords:
- Content Delivery Network (CDN) ; Optimization ; Streaming Algorithm ; Deep Learning ; Optimum Allocation
-
محتواي کتاب
- view
- مقدمه
- مقدمه
- تعریف مسئله
- اهمیت موضوع
- ادبیات موضوع
- اهداف پژوهش
- ساختار پایاننامه
- جمعبندی
- مفاهیم اولیه
- مقدمه
- شبکههای تحویل محتوا
- شبکههای تحویل محتوای مبتنی بر ابر
- جمعبندی
- کارهای پیشین
- مقدمه
- رویکرد تخصیص ایستا
- رویکرد تخصیص پویا
- رویکرد تخصیص پویا واکنشگر
- رویکرد تخصیص پویا پیشکنشگر
- رویکرد تخصیص پویا ترکیبی
- جمعبندی
- سازوکار پیشنهادی
- مقدمه
- پیشبینی بار شبکه
- مجموعهی داده
- معیارهای ارزیابی
- مدلهای پیشبینی
- مدل آریما
- مدل حافظه طولانی کوتاه مدت
- مدل شبکه عصبی چرخشی
- مدل واحد بازگشتی دروازهای
- مدل تبدیلکننده
- مقایسه پیشبینیهای انجام شده
- بهینهسازی مدلهای پیشبینی
- مهندسی ویژگیها
- غنی سازی مجموعه داده
- انجام پیشبینی
- قسمت واکنشگر
- جمعبندی
- نتیجهگیری
- مراجع
