Skip to main content

Future Optical Networks for Innovation, Research and Experimentation

Periodic Reporting for period 1 - ONFIRE (Future Optical Networks for Innovation, Research and Experimentation)

Reporting period: 2017-10-01 to 2019-06-30

Efficiency and cost-effectiveness of the future transport networks will play a key role in enabling the digital equality. Prices for ICT equipment and connection time are required to decrease. Otherwise, cost will remain a significant barrier to attain the aforementioned equality, since digital exclusion affects some of the most vulnerable and disadvantaged groups in society. Artificial intelligence (AI)-enabled automation of optical networks with white box switches is arguably the most significant challenge that needs to be addressed in order to further increase efficiency and cost-effectiveness. The ONFIRE project addresses these challenges via two research themes to take advantage of flexibility and modularity in the hardware and software domains, i) Disaggregation in optical networks and ii) Cognition in optical networks. This combination aims at bringing a reduction in capital expense (CAPEX) through the use of cost-effective optical systems for optical disaggregated networks with white boxes, and well as in operational expenses (OPEX) through the network optimization brought by artificial intelligence. The specific technological-oriented objectives of the ONFIRE project are:
• leverage on the flexibility and modularity that have been fostered by means of software techniques, exploiting software-defined networking (SDN) concepts, aiming at deploying an AI-enabled control scheme across heterogeneous optical transport networks.
• leverage on the flexibility from elastic interfaces and programmable white boxes, transforming the operation of today’s networks infrastructure and reducing over-provisioning and margins, in order to increase overall network equipment utilization.
• develop cost-effective subsystems for non-intrusive in-band monitoring of advanced modulation formats to guarantee both quality of transmission (QoT) and quality of service, QoS (on a per flow basis) for end-to-end services.
• use data mining techniques to leverage massive monitoring data from the physical layer, provided by advanced mechanisms in coherent interfaces, white boxes and monitoring subsystems, to continuously adjust and re-optimize network settings supported by SDN control allowing full network programmability with accurate performance prediction.

To achieve these technical objectives, ONFIRE deploys a three-year research programme centred on two European industrial PhDs.
Individual Research Project on data plane technologies for disaggregated optical networks
Fabiano Locatelli (ESR1) joined on May 2018. He has been involved in WP1 activities (disaggregated optical networks), including the state-of-the-art (SoA) analysis in order to identify open problems and methods leading to the ONFIRE goals. In particular, a deep study of three main subjects has been conducted: i) low-margin optical networks, ii) OSNR monitoring techniques and iii) disaggregated optical networks. Regarding the first research activity, during the first year of the project, the ESR has been involved in the design of a possible low-cost scheme for high-resolution signal spectrum analysis agnostic of the modulation format. On the other hand, with regard to the second identified topic, ESR has developed an algorithm able to understand, under certain circumstances, several spectral features that define the optical spectrum, such as OSNR, central frequency and bandwidth. This method performs properly, however returned deteriorated results when considering the filtering effects. Thus, in addition to this “classic approach” a second solution based on machine learning-based method for OSNR estimation has been developed relying on the support vector machine (SVM) algorithm. During the last months, the ESR has also been involved in the preparation of his thesis proposal titled “Data Plane Architectures and Monitoring Techniques for System Margins Reduction in Disaggregated Optical Networks”.

Individual Research Project on cognitive optical networks
Ankush Mahajan, joined ONFIRE project on September 2018 as ESR2. Since then, he has been directly involved on conducting WP2 activities such as elaborating the SoA analysis to allow identifying open problems and methods leading to the ONFIRE goals. Macroscopically the research topics studied by ESR2 focus on: adopting and exploiting Machine Learning (ML) techniques, enhancing QoT estimations, providing prompt failure detection & localization. These three topics are to be investigated through within both traditional and disaggregated optical networks. As a follow-up to D2.1 ESR2 in D2.2 narrows the scope to the identified research challenges and objectives. In this regard, D2.2 first outlines the limitations on applying Machine Learning (ML) to assist traditional QoT tools. In addition to the above summarized research topics, ESR2 has been also involved in the preparation of his PhD thesis proposal. In January 2019, he presented his doctorate thesis proposal entitled, “Machine Learning-Assisted Management of SDN-Enabled Disaggregated Packet & Optical Network” in the framework of the PhD program of the Signal Theory and Communications (TSC) department of the Universitat Politècnica de Catalunya-Barcelona Tech (UPC).
Recently, the approach of disaggregation of a chassis-based design into commodity (off-the-shelf) components has been gaining popularity, since it allows telecom operators and service providers to appropriately size their infrastructure and grow as needed. For the sake of an optimal network configuration and management, this paradigm implies a functional disaggregation, rather than block-by-block disaggregation. Thus, different data plane aspects have to be redesigned in order to favour an appropriate integration that provides the scalability pursued by the service providers. Additionally, interoperability aspects should be taken into account, since a smooth migration can be expected, moving from chassis-based (proprietary) network elements to network white boxes.

On the other hand, the adoption of cognitive and machine learning strategies appears as an appealing strategy aiming at providing the means to autonomously learn from the applied decisions and automatically plan and act (according to the current status) to (re-)configure the system with the objective to deal with the demands and requirements of the conveyed services. The adoption of cognition solution in networking is not novel approach but important research and innovation are still deserved to be produced to make cognition as a feasible asset to be effectively leveraged and rolled out within upcoming optical networks.

ONFIRE will significantly advance the state of the art in disaggregated optical networks by:
- developing novel modular programmable transceiver architectures and interfaces;
- introducing novel optical performance monitoring techniques agnostic to the optical signal waveforms
- defining the architecture to enable the deployment of cognitive solutions communicating with advanced monitoring systems and the SDN control / orchestration entities;
- devising novel cognitive algorithms adopting decisions and actions to enhance/ensure specific targets demanded by each service being accommodated.
The achievement of these goals will significantly increase the data plane flexibility and programmability, coping with increased transmission capacity and reach, while reducing CAPEX and energy consumption.
ML-based penalty estimator architecture (i.e., train/test/estimate)
Project Logo
Schematic diagram of the experimental setup for the high-resolution optical spectra acquisition