This paper is available on arxiv under CC 4.0 license.
Authors:
(1) Hatem A. Alharbi, CSchool of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom;
(2) Taisir E.H. Elgorashi, School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom;
(3) Jaafar M.H. Elmirghani, School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom. Table of Links Abstract & Introduction Related Works Repealing Net Neutrality Profit-Driven Model Results Conclusions & References Biographies Abstract Network neutrality (net neutrality) is the principle of treating equally all Internet traffic regardless of its source, destination, content, application or other related distinguishing metrics. Under net neutrality, ISPs are compelled to charge all content providers (CPs) the same per Gbps rate despite the growing profit achieved by CPs. In this paper, we study the impact of the repeal of net neutrality on communication networks by developing a techno-economic Mixed Integer Linear Programming (MILP) model to maximize the potential profit ISPs can achieve by offering their services to CPs. We focus on video delivery as video traffic accounts for 78% of the cloud traffic. We consider an ISP that offers CPs different classes of service representing typical video content qualities including standard definition (SD), high definition (HD) and ultra-high definition (UHD) video. The MILP model maximizes the ISP profit by optimizing the prices of the different classes according to the users demand sensitivity to the change in price, referred to as Price Elasticity of Demand (PED). We analyze how PED impacts the profit in different CP delivery scenarios in cloud-fog architectures. The results show that the repeal of net neutrality can potentially increase ISPs profit by a factor of 8 with a pricing scheme that discriminates against data intensive content. Also, the repeal of net neutrality positively impacts the network energy efficiency by reducing the core network power consumption by 55% as a result of suppressing data intensive content compared to the net neutrality scenario. Index Terms— Net neutrality, AT&T, IP over WDM networks, profit, power consumption. I. INTRODUCTION Network (net) neutrality regulations prohibit ISPs from applying different treatment to IP packets based on their content e.g. prioritizing, blocking or throttling certain Internet content or allowing quality differentiation. Net neutrality, which was scrapped by the US Federal Communications Commission (FCC) in December 2017, has been the subject of remarkable debate in recent years between ISPs and CPs with each side trying to exploit their assets and expand their profit and influence. The debate is fueled by the rapidly escalating demand for CPs services as a result of the interconnection between Internet and broadcasting markets. Cisco forecasts [1] that by 2021, annual global Internet traffic will hit 2.2 Zettabytes per month and CPs datacenters will be the source of 71% of this traffic. Online video services are the primary cause of this accelerated growth in Internet traffic. Video streaming is poised to consume 78% of the total CPs bandwidth with 75% of Internet video traffic originating from higher video services quality (HD and UHD). Proponents of preferential treatment of Internet traffic complain that the increasing demand for data-intensive content creates a significant burden on the communication network. They argue that removing net neutrality will give ISPs further control of their infrastructure, which is crucial in order to improve QoS and reduce security threats. Another argument is that a significant fraction of the profit of this tremendously growing market is seized by CPs whereas ISPs act as a transit or transport medium into CPs customers. In the US, the quarterly profit margin of AT&T (an ISP) has been almost stable over the last six years whereas Netflix (a CP) profit margin has risen up in rapid pace from 0.7% to 9.8% within the same period [2], [3]. In contrast, advocates warn that removing net neutrality will slow down the innovation in the Internet and its content and will limit the content competition by disadvantaging small businesses, and subsequently, diminish online services. Deploying traffic discrimination in video delivery services has many challenges, e.g. detecting video packets and enforcing a policy on a certain video quality. Traffic discrimination in IP communication networks has been surveyed intensively in the literature. Several traffic management practices have been surveyed in [4]. The authors highlighted that traffic discrimination taxonomy has four features: (i) characteristics or condition of the traffic (e.g. based on content, protocol or source/destination); (ii) traffic classification (e.g. based on flow rate, header information or routing); (iii) mechanism of discrimination (e.g. modify, delay, drop or block); and (iv) perceived discrimination by end-users. Video traffic can be analyzed using two mechanisms; deep packet inspection (DPI) [5], [6] or traffic profiling [7], [8]. DPI examines the data packets that are sent over the network and traffic profiling detects abnormal network traffic by comparing new traffic against previous traffic profile. For example, an alarm can be triggered if the data rate transmitted over the network (measured in bps) spikes above the desired data rate, which could indicate an increase in data rate. QoS for video services delivery can be applied either by reserving network bandwidth for video packets (e.g. using IntServ) or labelling video content as high priority e.g. by applying Differentiated Services (e.g. using DiffServ) [9]. The Internet ecosystem is complex with many stakeholders. As illustrated in Fig. 1, the main stakeholders in the Internet ecosystem are; ISPs, CPs, content delivery networks (CDNs) and end-users. Users pay ISPs a subscription fee to get Internet access and subscribe to CPs (if required) to access their content. CPs subscribe to a CDN to access storage and processing capacity and to deliver their content to customers. CDNs are responsible for sending CPs content at large scale over ISPs network infrastructure, e.g. the CP Netflix collaborates with the CDN Amazon Web Services (AWS) to reach their customers [10]. ISPs play as the key intermediary in the delivery process as they provide the required connectivity between users and content. Most ISPs such as AT&T [11] and Comcast [12] are now providing CDN services in additional to networking services. To simplify our analysis, we consider a direct relationship between ISP and CPs. Due to net neutrality regulations, current pricing policy of ISP networking services applies a fixed charge which is not linked with bitrate usage. For example, in the US, AT&T uses a fixed pricing model by charging CPs $3,282 per 10 Gbps per month [13] regardless of the content type transferred to users (either UHD video content or a simple text message). The rest of this paper is organized as follows: Section II briefly summarizes related work. We describe the pricing scheme we used in this paper and the profit-driven model we adopted in Section III. Our results are presented in Section IV. In Section V, we provide concluding remarks. This paper is available on arxiv under CC 4.0 license. Authors: (1) Hatem A. Alharbi, CSchool of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom; (2) Taisir E.H. Elgorashi, School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom; (3) Jaafar M.H. Elmirghani, School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom. This paper is available on arxiv under CC 4.0 license. available on arxiv Authors: Authors: (1) Hatem A. Alharbi, CSchool of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom; (2) Taisir E.H. Elgorashi, School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom; (3) Jaafar M.H. Elmirghani, School of Electronic and Electrical Engineering, University of Leeds, LS2 9JT, United Kingdom. Table of Links Abstract & Introduction Abstract & Introduction Related Works Related Works Repealing Net Neutrality Repealing Net Neutrality Profit-Driven Model Results Profit-Driven Model Results Conclusions & References Conclusions & References Biographies Biographies Abstract Network neutrality (net neutrality) is the principle of treating equally all Internet traffic regardless of its source, destination, content, application or other related distinguishing metrics. Under net neutrality, ISPs are compelled to charge all content providers (CPs) the same per Gbps rate despite the growing profit achieved by CPs. In this paper, we study the impact of the repeal of net neutrality on communication networks by developing a techno-economic Mixed Integer Linear Programming (MILP) model to maximize the potential profit ISPs can achieve by offering their services to CPs. We focus on video delivery as video traffic accounts for 78% of the cloud traffic. We consider an ISP that offers CPs different classes of service representing typical video content qualities including standard definition (SD), high definition (HD) and ultra-high definition (UHD) video. The MILP model maximizes the ISP profit by optimizing the prices of the different classes according to the users demand sensitivity to the change in price, referred to as Price Elasticity of Demand (PED). We analyze how PED impacts the profit in different CP delivery scenarios in cloud-fog architectures. The results show that the repeal of net neutrality can potentially increase ISPs profit by a factor of 8 with a pricing scheme that discriminates against data intensive content. Also, the repeal of net neutrality positively impacts the network energy efficiency by reducing the core network power consumption by 55% as a result of suppressing data intensive content compared to the net neutrality scenario. Index Terms — Net neutrality, AT&T, IP over WDM networks, profit, power consumption. Index Terms I. INTRODUCTION Network (net) neutrality regulations prohibit ISPs from applying different treatment to IP packets based on their content e.g. prioritizing, blocking or throttling certain Internet content or allowing quality differentiation. Net neutrality, which was scrapped by the US Federal Communications Commission (FCC) in December 2017, has been the subject of remarkable debate in recent years between ISPs and CPs with each side trying to exploit their assets and expand their profit and influence. The debate is fueled by the rapidly escalating demand for CPs services as a result of the interconnection between Internet and broadcasting markets. Cisco forecasts [1] that by 2021, annual global Internet traffic will hit 2.2 Zettabytes per month and CPs datacenters will be the source of 71% of this traffic. Online video services are the primary cause of this accelerated growth in Internet traffic. Video streaming is poised to consume 78% of the total CPs bandwidth with 75% of Internet video traffic originating from higher video services quality (HD and UHD). Proponents of preferential treatment of Internet traffic complain that the increasing demand for data-intensive content creates a significant burden on the communication network. They argue that removing net neutrality will give ISPs further control of their infrastructure, which is crucial in order to improve QoS and reduce security threats. Another argument is that a significant fraction of the profit of this tremendously growing market is seized by CPs whereas ISPs act as a transit or transport medium into CPs customers. In the US, the quarterly profit margin of AT&T (an ISP) has been almost stable over the last six years whereas Netflix (a CP) profit margin has risen up in rapid pace from 0.7% to 9.8% within the same period [2], [3]. In contrast, advocates warn that removing net neutrality will slow down the innovation in the Internet and its content and will limit the content competition by disadvantaging small businesses, and subsequently, diminish online services. Deploying traffic discrimination in video delivery services has many challenges, e.g. detecting video packets and enforcing a policy on a certain video quality. Traffic discrimination in IP communication networks has been surveyed intensively in the literature. Several traffic management practices have been surveyed in [4]. The authors highlighted that traffic discrimination taxonomy has four features: (i) characteristics or condition of the traffic (e.g. based on content, protocol or source/destination); (ii) traffic classification (e.g. based on flow rate, header information or routing); (iii) mechanism of discrimination (e.g. modify, delay, drop or block); and (iv) perceived discrimination by end-users. Video traffic can be analyzed using two mechanisms; deep packet inspection (DPI) [5], [6] or traffic profiling [7], [8]. DPI examines the data packets that are sent over the network and traffic profiling detects abnormal network traffic by comparing new traffic against previous traffic profile. For example, an alarm can be triggered if the data rate transmitted over the network (measured in bps) spikes above the desired data rate, which could indicate an increase in data rate. QoS for video services delivery can be applied either by reserving network bandwidth for video packets (e.g. using IntServ) or labelling video content as high priority e.g. by applying Differentiated Services (e.g. using DiffServ) [9]. The Internet ecosystem is complex with many stakeholders. As illustrated in Fig. 1, the main stakeholders in the Internet ecosystem are; ISPs, CPs, content delivery networks (CDNs) and end-users. Users pay ISPs a subscription fee to get Internet access and subscribe to CPs (if required) to access their content. CPs subscribe to a CDN to access storage and processing capacity and to deliver their content to customers. CDNs are responsible for sending CPs content at large scale over ISPs network infrastructure, e.g. the CP Netflix collaborates with the CDN Amazon Web Services (AWS) to reach their customers [10]. ISPs play as the key intermediary in the delivery process as they provide the required connectivity between users and content. Most ISPs such as AT&T [11] and Comcast [12] are now providing CDN services in additional to networking services. To simplify our analysis, we consider a direct relationship between ISP and CPs. Due to net neutrality regulations, current pricing policy of ISP networking services applies a fixed charge which is not linked with bitrate usage. For example, in the US, AT&T uses a fixed pricing model by charging CPs $3,282 per 10 Gbps per month [13] regardless of the content type transferred to users (either UHD video content or a simple text message). The rest of this paper is organized as follows: Section II briefly summarizes related work. We describe the pricing scheme we used in this paper and the profit-driven model we adopted in Section III. Our results are presented in Section IV. In Section V, we provide concluding remarks.

Too Long; Didn't Read

Making security fun one episode at a time

Impact of the Net Neutrality Repeal on Communication Networks: Abstract & Introduction

Too Long; Didn't Read

Net Neutrality: Unbiased Internet Access for All!

About Author

TOPICS

THIS ARTICLE WAS FEATURED IN...

Impact of the Net Neutrality Repeal on Communication Networks: Abstract & Introduction

Too Long; Didn't Read

Net Neutrality: Unbiased Internet Access for All!

About Author

TOPICS

THIS ARTICLE WAS FEATURED IN...

RELATED STORIES