Performance evaluation of video server replication in metro/access networks

Abstract

Internet traffic is increasingly becoming a media-streaming traffic. Especially, Video-on-Demand (VoD) services are pushing the demand for broadband connectivity to the Internet, and optical fiber technology is being deployed in the access network to keep up with such increasing demand. To provide a more scalable network architecture for video delivery, network operators are currently considering novel metro/access network architectures which can accommodate replicated video servers directly in their infrastructure. When servers for VoD delivery are placed nearer to the end users, part of the traffic can be offloaded from the core segment of the network, and the end users can experience better Quality of Service (QoS). While the deployment of caching systems for traffic offloading has been studied in the core network, no work has already investigated the potential performance gains by replicating the content in the metro/access segment of the network, even closer to the users. In our work, we will compare the performance of video server replication in different metro/access network architectures, i.e. a metro ring architecture and a tree-based architecture, by considering both active and passive technologies. We will evaluate using both simulative and analytical methodologies how content providers could benefit from the deployment of replicas of video servers in terms of blocking probability of the VoD requests.

Latency- and capacity-aware placement of chained Virtual Network Functions in FMC metro networks

“…Passive Optical Networks (PONs) or Active Optical Networks (AONs) [5] are more and more adopted for the aggregation of fixed or mobile users' traffic in the metro/access segments, but they still offer bandwidth capacity that is typically much scarcer than the available bandwidth in the core segment. This means that the metro/access networks can become a bottleneck if too much traffic is kept within the metro/access domain (i.e., by hosting too many VNFs), potentially causing disruptive effects [6]. This is true also for computational resources, which are in general much more constrained at the edge of the network than in remote DCs [7].…”

A QoS-enhanced data replication service in virtualised cloud environments

“…As cloud users are paying more and more attentions on quality-of-service (QoS), an effective data replication service should reduce access latency and bandwidth consumption, also be capable of achieving load balancing and improves reliability by creating multiple data copies (Victer Paul et al, 2012;Inacio and Dantas, 2014;Mansouri, 2016). Unfortunately, due to the dramatic increasing size of datasets, many traditional data replication services become more and more inefficient, which is especially true when the Big Data age is coming (Balasangameshwara and Raju, 2013;Mokadem and Hameurlain, 2015;Savi et al, 2015).…”