Achieving 100G Transmission Rates in Packet Transport Networks

Abstract

A new method for scaling packet ring network transmission capacity from 10Gbits/s to 100Gbits/s is presented in this article. The concept of high capacity packet rings is introduced as a way of enabling service providers to fully utilize the installed network capacity for the scaling and distribution of high definition IP video services. It describes a hashing mechanism for load balancing and enhanced resiliency over the high capacity links. IntroductionDuring recent years, content distribution networks have become a prominent business opportunity for service providers. The distribution of on-demand high-definition (HD) content such as HD Video on Demand (VOD) and IP Television (IPTV) imposes new scalability requirements on Multi-Service Operators (MSO) and Telcos to scale their transport infrastructure beyond 10Gbits/s. In many cases, packet transport networks have been built to enable the distribution of broadcast standard definition (SD) VOD and IPTV services [1]. This design, however, is currently being challenged by a shift towards HD ondemand services with concurrency rates varying in some cases from 10% to 20%, increasing scalability requirements beyond 10Gbits/s and in some cases up to 100 Gbits/s.In this paper we present a definitive way of guaranteeing in-service scalability of packet transport ring architectures beyond 10 Gbits/s, without compromising neither the performance nor survivability of the transport network. Scalable Packet Transport ArchitecturesAs mentioned above, one of the main challenges service providers face when trying to increase network capacity to satisfy the demand of HD on-demand services is to provide non-disruptive and gradual scaling of ring topologies.Currently there are two possible solutions to this problem: 1) to partition the ring into smaller rings capable of accommodating the bandwidth requirements of a reduced number of subscribers, or 2) to increase the basic transmission rate of the ring infrastructure beyond initial capacity. The first option increases the long term OPEX, as it requires the operator to manage an increasingly larger number of ring networks serving smaller regions. On the CAPEX side, it increases costs as it requires more network elements to distribute the content to a larger number of rings. Over time, this solution is unsupportable as it would require the operator to continuously divide its rings as the demand for bandwidth and concurrency rates increase. The second alternative is highly disruptive and involves a vast investment in new equipment to support the higher transmission rate.A third alternative is the introduction of high capacity (HC) packet transport rings. This alternative represents an evolutionary approach which allows in-service scaling of the transmission rate in ring architectures beyond 10Gbits/s and upto 100Gbits/s. The infrastructure may be upgraded as the demand for bandwidth capacity grows without disrupting existing services on the ring. A segment of a typical triple-play network relying on high capacity packet transport ring...