The major objectives of DISCUS are to:
- Produce an end-to-end network architecture (including access, metro and core networks) based on LR-PONs directly connected to a flat optical core that economically scales when sustained customer bandwidths grow by three orders of magnitude or more compared to today’s DSL sustained rates. The LR-PON will have a target physical layer split of 512 with a logical split of at least 1024 to take advantage of future technology improvements as they become available, for example coherent detection in the access network. The reach target will be 125 km with a minimum target of 100km. however designs that trade split and range will also be investigated. The solution will be tested via cash flow and performance comparisons against conventional FTTX access architectures (e.g. GPON, FTTCab, point to point), Ethernet metro access (backhaul) network and
IP-MPLS core networks. Target improvements over conventional architectures are: reduction in buildings ~98%, reduction in customer network ports >99%, reduction in network port cards ~70%.
- Produce a modelling framework that enables the architecture to be tested within a wide range of different geographies/countries, including those with minimal data sets available. These will be cash flow models and traffic and service scenario models. Much of the model structure will be adopted from earlier work and adapted to the DISCUS architecture and measurement and comparison requirements.
- Produce a solution that can gracefully grow sustained bandwidth over the installed network infrastructure without service disruption or major infrastructure modification/upgrades. The initial single wavelength solution will operate at 10Gb/s symmetrically. For a full 512 way physical layer split this offers a sustained bandwidth of 20Mb/s per customer. With typical statistical multiplexing gains customers would experience ~200Mb/s with peak rates capability of 10Gb/s depending on ONT customer interfaces. Additional wavelengths will add this capacity again for each wavelength added. If only the C-band is exploited (initial target) then up to ~40 additional channels could be added. This means that the same network architecture could easily upgraded over time to potentially offer a massive 32 Gb/s per user (considering statistical multiplexing gains) with 40Gb/s peak rate if 40 wavelengths are added into the system (using 40 Gb/s transceivers). However some of these wavelengths would also be used for core wavelengths over the access network, target for these wavelengths will be 100Gb/s (see below).
- Produce a network architecture that has lowest power consumption with minimal and reducing power per unit bandwidth as network bandwidth grows. Target will be at least one order of magnitude reduction in power consumption compared to incremental growth of today’s architectures.
- Produce a flexible control plane that allows multi-layer (fibre, wavelength, layer 2 and layer 3) switching of traffic flows at the core nodes.
- Demonstrate that the architecture can successfully evolve from today’s architectures.
- Ensure the solution is able to adopt new technologies as they arise while co-existing with previous generation technologies – graceful evolution. Once the fibre infrastructure is in place the target for the architecture will be to grow bandwidth, add services, new technologies etc. with service disturbance to customers no greater than normal protection/restoration recovery events. For example one target will be to enable physical layer access via reused fibres to the LR-PON amplifier nodes for future technologies operating at different wavelength or even wave bands (such as the L-band).
- Ensure the solution provides service resiliency mechanisms in case of faults in the physical infrastructure including total metro-node failure (i.e. disaster recovery will be considered as an inherent part of the design).
- Demonstrate the ‘Principle of Equivalence’ for connections, showing that all network access points have equal potential bandwidth and service capability via a demonstration of 100Gb/s over the Access path to the Metro node and into the core network.
- Produce a network architecture with a combined wireless edge with ubiquitous optical access and demonstrate the benefits to customers that the combined architecture produces.
- Produce white papers to help guide regulatory policy that enables a truly competitive environment that is controlled by the customers and users rather than network operators and service providers.
DISCUS takes a clean slate perspective for the future broadband network architecture. It will:
- Ensure reuse of the physical layer infrastructure wherever possible in order to avoid the very high trenching and cabling costs that would occur in a new build scenario.
- It will determine a target architecture that remains economically viable as demand and bandwidth scales to three or more orders of magnitude compared to today’s copper and electronic based networks.
- It will determine evolution paths from today to the future architecture. It will provide hooks at minimal upfront cost for future evolution as new technologies emerge.
- It will provide an architecture that enables a fully competitive service provision environment
without the wasted resources or contractual lock-in caused by simple physical layer only
- It is anticipated that the new architecture will pose challenges for the regulatory environment
and therefore, DISCUS will be active in stimulating debate and formulating regulatory policy and strategy to enable the benefits the architecture can provide.
It will be important to ensure that the architecture can be economically deployed in as wide a range of geographies/territories and geo-types as possible and used standardised approaches and technology where necessary. However the architecture will avoid unnecessary standardisation where possible and enable competing service providers to differentiate themselves as much as possible. DISCUS will therefore be active in opinion forming activities and relevant standardisation activities.