One of the main purposes of Mohammadhassan Safavi’s Ph D thesis is to leverage machine learning algorithms to predict where users activities should be done to incur minimum cost. Cost could be energy consumption, service quality degradation, service delay and so on. The thesis also covers many other aspects from reducing energy usage in data centers to how machine learning can be used in 5G mobile operator networks. On Friday 16 October 2020 Hassan defends his thesis ”Content and Resource Management in Edge Networks”
A big portion of Mohammadhassan Safavi’s thesis is focusing on user generated content such as texts, photos or videos that everyone takes and shares with friends and family via Internet. And for this digital content he has developed algorithms to find out where in the network the best host server should be
“Consider yourself traveling as a tourist to a beautiful European city”, he explains. “You take a picture of a historical building and share it on Instagram. Your friends start seeing it once you upload it; What many people may not know is the complicated algorithms that run behind the curtains from the time you share your photo until your friends view it.”
“Imagine yourself answering this question in which city the server should be?” he continues. “In the city where the photo is taken; in the city of your residence; Or in the city where you have grown up? These are just simplified versions of real technical questions that we as researchers deal with, because it was not asked in what aspects a place is good or bad or what the consequences of a wrong decision are.“
Both enterprises (e.g. media companies, NEWS agencies) and ordinary people can produce web contents. Content delivery networks (CDN) are used to distribute them to the population. Telco-CDNs, thanks to their local presence near target populations, can cache contents to enhance quality of the experience.
What made you want to pursue a PhD?
“I have always enjoyed being emerged in an academic environment. Even before starting my PhD journey, I was a university lecturer for five years in my home country. PhD research is very exciting as you get to work on a project of your own and experience the feeling of delivering something at a world-class level. For example, when I get promising results in my research, I envision how my findings can be published beside the findings of best authors and researchers from around the world”, Hassan says.
What is the most fascinating or interesting with your thesis subject?
“The large part of my research focuses on handling user generated data and IoT data. I believe one day these data will be kind of asset being the main part of every individual, enterprise or even society capital. Therefore, extracting information and managing these data and making critical decisions from them are the most fascinating tasks in my opinion.”
How will your research com to use?
“I believe my results or similar findings from other researchers will be used in practice. These days all enterprises are constantly looking for ways to minimise their cost and shift towards making their solutions more optimum. In one part of my work, I concluded a strategic case that can be useful in decision making for video-based product lines. In another part of my work, I proposed machine learning-based analytic systems that motivate enterprises not to use dedicated personnel for analytics and parameter tuning. Moreover, in my content delivery research, I proposed how Internet operators can have shares in revenue chain of user-generated video delivery systems. In my opinion, PhD students should think of ways to transform commercializing their ideas and I believe I am potentially among those.“
What are your plans?
“I am currently looking for opportunities both in academia and industry. My dream job will be having an academic position as well as running my own company to patent my ideas. As of now, I am offered a post-doctoral position and I am considering accepting, as it is a way for me to further contribute to the scientific society and promote my research, Hassan Safavi concludes.“