Next Generation Optical Network Architecture Featuring Distributed Aggregation, Network Processing and Information Routing Theofanis G. Orphanoudakis, Chris Matrakidis, Alexandros Stavdas Department of Informatics and Telecommunicatios University of Peloponnese Tripolis, Greece {fanis, cmatraki, astavdas}@uop.gr Abstract— The ontology of communications is rapidly changing, shifting interest to machine-to-machine (M2M) interactions and the internet of Things (IoT). These are becoming vital for sustainability of social life and the revitalization of the economy providing the infrastructure to new production forms like distributed manufacturing, cloud robotics while becoming important to grid-based energy systems. Adding to them the voracious needs for data of the traditional broadband users, residential or business, together with the back/front hauling requirements of mobile operators, one is expecting a significant strain in the access. A multitude of heterogeneous access networks are emerging and the integration of them in a single platform ensuring seamless data-exchange with Data-Centres is of major importance. In this paper we describe HYDRA (HYbriD long-Reach fiber Access network), a novel network architecture that overcomes the limitations of both long-reach PONs as well as mobile backhauling schemes, leading to significantly improved cost and power consumption figures. The key concept is the introduction of an Active Remote Node (ARN) that interfaces to end-users by means of the lowest cost/power consumption technology (short-range xPON, wireless, etc.) whilst on the core network side it employs adaptive ultra-long reach links to bypass the Metropolitan Area Network. The scheme leads to a higher degree of node consolidation, network convergence and Access-Core integration. The proposed architecture can enhance performance while supporting network virtualization and efficient resource orchestration based on Software Defined Networking (SDN) principles and open access networking models. Keywords—Ubiquitous network infrastructure, software defined networking, network convergence, optical networks I. INTRODUCTION Interconnecting a vast number of sensors, actuators and various artifacts is becoming an indispensable tool to Cloud- based economic activity. New forms of interaction and communication like machine-to-machine (M2M) interaction and the Internet of Things (IoT) promote new production forms and methods like distributed manufacturing and cloud robotics for highly automated factories while providing the backbone to grid-based energy systems. Moreover, the data-rate needs of residential and business users keep increasing at an astonishing rate. Finally, the proliferation of pico/femto-cell wireless networks also puts important pressure to mobile operators. Apparently, a large number of heterogeneous networks are emerging and building an independent, communication infrastructure for each one of them is not an option. Moreover, the essence of Cloud-supported production and social welfare requires the constant data exchange of these networks to centralized, distributed or federated data centres. Moreover, this poses stringent performance requirements in terms of QoS performance (mainly latency and packet loss) to the underlying infrastructure [1] while it requires global optimization of IT and Telecom resources followed by resource slicing policies for “commoditization”. Currently, access networks rely upon centralized packet multiplexers to perform network and protocol processing functions, manage subscriber state, and aggregate network traffic. However, this monolithic model can lead to capacity and space bottlenecks, poor reconfiguration capabilities and limited interfaces for users to control and configure the network infrastructure and limit the revenue that data services can produce for operators. Today, the last drop to end-user is provided by means of either fixed-line access technologies, such as cable modems, xDSL as well as wireless technologies such as mobile 3G/4G/LTE-A networks (HSDPA/HSPA+), WiFi and WiMax. Wireless networks feature high flexibility in terms of broad area coverage and end-user network accessibility but face the limitations of relatively low bandwidth (10’s to a few 100’s of Mb/s shared between all the users in a cell) and limited bandwidth connections to backhaul networks. Hence, it seems inevitable that the edge of the telecom network will be based on a combination of a wireless part for ubiquitous service provision to mobile users, but also fiber for adequate capacity which is not possible with the wireless infrastructure alone. In the fixed-line front, Passive Optical Networks (PONs) are widely recognized as the dominant broadband scheme [2]- [4]. This technology has also the potential to allow network convergence, e.g. when exploited for mobile network front/backhauling (MFH/MBH). It is worth noting that in order to achieve the required data rate over wireless/mobile networks, cell distances have to be small, which results in a large number of cells in urban areas. This makes a PON backhaul architecture very attractive, since serving a large number of cells with point to point links is prohibitively costly. However, the cost of deploying and operating such high capacity next generation access networks is significant. Thus,