What is ONE
ONE is a Collaborative Project STREP from the competition FP7-ICT-2017-5, which started on Sep 1, 2017 with the duration of three years.
In a nutshell, ONE addresses three important trends in networking: programmability, semantic adaptation and orchestration. ONE’s main goal is to contribute to these trends, especially in the context of integration of high-speed optical transmission and switching with the future Internet. ONE is an architecture, system and tool at the same time. ONE unique focus on network management systems (NMS).
• Coordination-without-integration of NMSs for a multi-vendor, multi-layer network management system coordination, including the emerging third party systems, such as Path Computation Element (PCE);
• Semantic adaptation between the NMSs over standardized interfaces (Web Services/MTOSI, etc) and as such as vendor-independent;
• Combination of manual and automated network management processes, as well as smart analytics to create, manage and redefine network management procedures custom tailored to each telco’s business process;
• Programmability of legacy management workflows and systems involving multiple NMSs, adapting them semantically, and orchestrating services and business processes;
• Technology and vendor independence with automated and programmable business processes.
In comparison to the above features, however, ONE has not been designed to be…
• a new NMS; it is designed to be a light weight system independent of any NMS;
• a control plane client; but is control-plane friendly and can talk to any control plane implementation;
• yet-another virtualization or router programmability tool, nor it changes anything on legacy systems; it only adapts the semantics of whatever vendor is deployed in a carrier network;
Key Innovation Example: Orchestration & Programmability through coordinated (and automated) management operations
ONE envisions network management of the future to be a combination of Manual Programmability, Smart Analytics and Autonomic Network Management. Due to the existence of human-in-the-control-loop, this combination will likely complement efforts for full automation of network management, just like an autopilot compliments (and does not replace) a live presence of a pilot. To enhance dynamic provisioning, configuration, fault detection and response in complex multi-service networks, a combination of manual workflow programmability (as defined by humans) and automation of many common functions is critical. Both manual programmability and automation can be combined with sophisticated analytics, so that many elements in the network and human organizational procedures, together with resources and traffic behaviors, can be continually monitored, analyzed and effectively executed. Any types of special conditions can be detected and responses can be provided depending on requirements, and (human) business processes.
Take an example of Internet and transport network management. The automated interaction between the Internet network infrastructure and carrier-grade transport systems has been at the heart of carriers’ efforts to lower capital and operational expenses and improve overall network performance. However, in practice even the simplest processes remain fragmented and lack orchestration. For instance, the transport network NMS is usually designed to deliver a small number of services with fairly static demands on network operation, with simple management procedures. On the other hand, the IP network was expected to support a large number of services and quickly adopt new upcoming services to reduce time to market.
As a result, telecom carriers have been forced to find a reasonable balance between the complexity and associated cost of the operations required at the IP layer, and the simplicity and cost savings of operating and configuring the equipment at the transport layer. On top of this, both IP NMS and transport NMS need to interoperate with third party systems, such as PCE. Such a complex ecosystem requires smart orchestration of business procedures, decision making process based on analytics, and openness to third party systems with semantics adaptation. This is a great challenge that ONE is addressing.