National and Kapodistrian University of Athens
Economics and resource management in content delivery networks
Abstract
dc:descriptionWe focus on the management and pricing of resources in Content Delivery Networks (CDNs), which are extensively adopted nowadays as a very efficient mechanism for Internet information provision. We establish a multi-stage framework that can be used by Origin Servers (OSs), that is, content-generating entities, for claiming and reserving CDN resources in a fine-grained way. The basis of our work concerns the trading and pricing of resources and lies in the use of stocks, a Secondary Market (SM), and Stock Options (SOs); concepts, tools, and products commonly used in modern capital markets.We treat resources as stocks. Using a Predictive Reservation Scheme (PRS), network resources, i.e., traffic volume (TrV), offered by a CDN are being monitored through well-established techniques, i.e., Kernel Regression Estimators (KREs), in a given time frame (typically in the order of an hour). This information is dynamically updated to reflect transient behavior and longer-term trends. The model we establish relies initially on the resource use forecasting that the kernels can provide but also on other complementary techniques. To cope with inaccuracies in the Load Prediction Mechanism (LPM) predictive capability that would imply inefficient use of the claimed resources, we establish a SM for the stock trading. This scheme allows the fast offloading of unused reserved resources between OSs. This exchange occurs either by implementing socially optimal practices or by allowing automatic electronic double auctions at the end of the day (EoD) or at shorter intervals. Then, SOs are used to address the potential scarcity of resources, not adequately addressed by the stock trading on the SM. SOs are purchased and exercised depending on the lack of resources at the EoD. SOs can also be traded in the SM. As a result, OSs may acquire resources (if required) at a standard price, avoiding penalizing tariffs for last-minute requests.The proposed framework is examined through extensive trace-driven simulations, with the mechanisms that comprise it being gradually activated. Corresponding trace-driven simulations are also performed for the dominant market pricing plans, i.e., the Pre-costed (PrC) and Pay As You Go (PAYG) plans. The optimal limits of PrC and PAYG plans, as well as the conditions under which the efficiency of these plans can be improved, are also explored. Among the metrics considered during the simulations are: (a) the cost for the OS; (b) the volume of resources committed and the percentage of them that remain unused; (c) the cost for the CDN; (d) the profitability of the CDN; and (e) the cost-benefit ratio for the CDN.Defining the efficiency of a mechanism under consideration as the distance of the above metrics of the mechanism from the corresponding metrics of the ideal plan (IdP), we conclude that the proposed PRS (with all sub-mechanisms activated) has the best efficiency among all the considered plans. At the same time, PRS is a Pareto - Optimal solution.
Degree
thesis:*- Grantor dc:publisher
- National and Kapodistrian University of Athens
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Vathias, Elias
- Βαθειάς, Ηλίας
Subjects
dc:subject × 42- Δίκτυο Διανομής Περιεχομένου
- Διακομιστής Προέλευσης
- Μετα-Δίκτυο Διανομής Περιεχομένου
- Κεφαλαιαγορά
- Μετοχή
- Δευτερογενής Αγορά
- Δικαίωμα Προαίρεσης
- Σχέδιο Τιμολόγησης
- Αυτόματες Ηλεκτρονικές Δημοπρασίες
- Παρακολούθηση
- Μοντελοποίηση
- Πρόβλεψη
- Σχήμα Πρόβλεψης Κρατήσεων
- Πόροι
- Κατανομή πόρων
- Διαχείριση πόρων
- Τιμολόγηση Πόρων
- Κοινωνικά Βέλτιστες Πρακτικές
- Content Delivery Network
- Origin Server
- Meta-CDN
- Capital market
- stock
- Secondary Market
- Stock Option
- Pricing Plan
- Automatic Electronic Auctions
- Monitoring
- Modeling
- Forecasting
- Predictive Reservation Scheme
- Resources
- Resource allocation
- Resource management
- Resource Pricing
- Socially Optimal Practices
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Δίκτυα ηλεκτρονικών υπολογιστών και Επικοινωνίες
- Computer and Information Sciences
- Natural Sciences
- Computer Networks and Communications
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/52333