A Research Framework for the Clean-Slate Design of Next-Generation Optical Access

Abstract

Abstract-A comprehensive research framework for a comparative analysis of candidate network architectures and protocols in the clean-slate design of next-generation optical access is proposed. The proposed research framework consists of a comparative analysis framework based on multivariate noninferiority testing and a notion of equivalent circuit rate taking into account user-perceived performances, and a virtual test bed providing a complete experimental platform for the comparative analysis. The capability of the research framework is demonstrated through numerical results from the study of the elasticity of hybrid TDM/WDM-PON based on tunable transceivers.Index Terms-Next-generation optical access, comparative analysis, equivalent circuit rate (ECR), statistical hypothesis testing, non-inferiority testing, quality of experience (QoE).

The effect of ISP traffic shaping on user-perceived performance in broadband shared access networks

“…The comparative analysis framework proposed in [5] addresses these drawbacks of the original ECR framework by extending it to multiple performance metrics and taking into account statistical variability of measured data in comparison using multivariate non-inferiority testing. The non-inferiority testing is a one-sided variant of the equivalence testing used in Medicine and Biology for the establishment of the equivalence between two different clinical trials or drugs [34].…”

“…To take into account the statistical variability in measured data and process multiple performance metrics in an integrated way during the comparison, we adopt the comparative analysis framework that we proposed for the clean-slate design of next-generation optical access in [5]. In this framework we use the user-perceived performance of a single subscriber under no traffic shaping as a reference case against which we compare the user-perceived performance of all other shaped configurations either with a single subscriber or with multiple subscribers.…”

“…Unlike metro and backbone networks, however, access networks directly interface with end-users, so it is important to base the study and design of access networks on the behaviors of and the actual performance perceived by end-users [5]. The major goal of our study in this paper, therefore, is to investigate the effect of ISP traffic shaping on user-perceived performance, i.e., the quality of experience (QoE), and thereby to provide ISPs further insights into the design, deployment, and operation of the next-generation access networks from end-users' perspective.…”

“…1 The current study was specifically triggered by the results from recent investigations of "PowerBoost" in [1] and [2], the feature present in some cable broadband networks that enables sharing of unused capacity by giving customers extra bursts of speed whose duration is controlled by the token bucket size. In this paper we extend those investigations of the effect of the token bucket size on packet-level performance at network/transport layers to those on user-perceived performance at application/session layers based on the research framework which we proposed for the clean-slate design of next-generation access networks [7,5].…”

“…An overview of multivariate non-inferiority testing procedure[5]. User-perceived performance metrics with 95 percent confidence intervals for a single subscriber with access line rate of 100 Mbit/s: (a) Average session delay of HTTP traffic, (b) average session throughput of FTP traffic, and (c) decodable frame rate (Q) of UDP streaming video.…”