ICTON 2010 Mo.A.1
978-1-4244-7797-5/10/$26.00 ©2010 IEEE 1
European and American Research Toward Next-Generation
Optical Access Networks
Leonid G. Kazovsky
1
, Claus Popp Larsen
2
, Dirk Breuer
3
,Anders Gavler
2
, Mikhail Popov
2
, Kun Wang
2
,
Gunnar Jacobsen
2
, Erik Weis
3
, Christoph Lange
3
, ShingWa Wong
1
, She-Hwa Yen
1
, Vinesh Gudla
1
,
Pegah Afshar
1
1
Stanford University, USA;
2
Acreo AB, Sweden;
3
Deutsche Telekom
ABSTRACT
Next-generation optical access networks will deliver substantial benefits to consumers including a dedicated
high-QoS access to bit rates of hundreds of Megabits per second. They must also deliver significant benefits to
network owners/operators to justify the needed infrastructure investment expected to reach billions of Euros.
Benefits to network owners/operators are expected to include reduced total cost of ownership, due to higher
reliability, lower energy consumption, better flexibility and efficiency, and a smaller number of sites needed to
support network operations. This paper will describe recent progress toward that goal including R&D efforts in
Europe under the FP7 projects ALPHA and OASE and in the US at Stanford University under the SUCCESS
and DAN projects.
1. INTRODUCTION
With ever-increasing demand for bandwidth, optical access networks such as Passive Optical Networks (PONs)
and active optical networks (AON) are being deployed by many service providers worldwide. Current PONs can
support rates of 2.5 Gbps and a reach of 20 km and AONs can support up to 1 Gbit/s per subscriber symmetrical
connections. While this is sufficient to meet the near-term user-bandwidth requirements, higher rates and newer
technologies such as WDM are required for next generation PONs to support future demands. The ideal scenario
would be to have the increased bandwidth at similar or lower cost compared to present-day networks. One
technology that could provide a solution keeping in mind the demands on future networks is the Active Optical
Network (AON)-based architecture. With the in-built redundancy of Ethernet, AONs have in particular the
advantage of keeping networks localized and can reduce the load on backbone servers. Another important factor
in migration to newer architectures is to bring down the total operating costs (OpEx) of the network and the
OASE (Optical Access Seamless Evolution) project is a step in this direction which envisages Central Office
(CO) consolidation that could bring down the Total Cost of Ownership (TCO) of the network. Reducing the
OpEx would also allow the operator greater freedom to provide more services to the customer. The challenge is
to integrate these technologies in already-deployed infrastructures and to fulfil the upgrades in an incremental
and non-interruptive manner, i.e. graceful evolution over the next several years. Increasing integration of
wireless technologies and the use of novel components are important to continue offering flexible and cost
effective broadband access. One of the key PON challenges is to design and introduce low-cost reconfigurable
devices that also preserve the passive nature of PON. With the rapid development within optical access
networking it is impossible to foresee exactly what “next generation” will bring us, which is exemplified in this
paper. The business model can dictate the preferred technology, and today the extremes are the incumbent
vertical integration versus the municipal open access. It is probably safe to assume that both models will coexist
and be further developed within a foreseeable future, and they may even cross-fertilize each other. The paper is
organized as follows. Section 2 presents Acreo’s work in the ALPHA (Architectures for fLexible Photonic
Home and Access networks) project on dynamic access/aggregation networks for AON architectures. Section 3
discusses the Deutsche Telekom’s (DT) activities within the OASE project and the proposed work for a
consolidated network model at a lower TCO. Section 4 describes optical access research at Stanford and presents
the S tanford U niversity Access (SUCCESS) architectures, which enable graceful evolution from current to next-
generation optical access networks
2. ACTIVE OPTICAL NETWORKS-BASED APPROACH (ACREO)
Active optical networks (AON, a.k.a. active or point-to-point Ethernet) are predominantly based on Ethernet that
is thoroughly standardised within the IEEE concerning physical interfaces as well as control and management
planes. Whereas end user transmission speeds are not an issue in AON (as opposed to PON) the next generation
activities within the ALPHA project are focused on dynamic control and management in the access and
distribution (a.k.a. aggregation or metro) network as well as integration with network control in the core
network. Current access and aggregation networks mainly focus on having an aggregative function where traffic
mainly flows between end users and the core parts of the network, see Fig. 1a). Also seen in this picture is the
fact that a CO node is in principle the same as an aggregation switch and this leads to added flexibility in the
AON Ethernet designs and in the dynamics of adding new customers. From an Ethernet point of view the
networks are tree-shaped, but from an optical point of view rings are often deployed for redundancy. Meshed