Machine Learning in Computer Networking
📌 AI and Machine Learning in Software Defined Networking
Global Telecommunication Giant deploys Machine Learning-Enhanced Software Defined Network
Some news agencies say: Vodafone successfully completed an advanced software upgrade across its entire global transport network, which serves hundreds of millions of users and third-party internet and content providers in 28 countries across four continents.
The software enhancement is the foundation on which Vodafone and its wholesale and enterprise customers can progressively add capacity faster and more cost-effectively to meet future demand, as well as introduce new services and boost security at scale.
As part of its digital evolution, Vodafone directly applied Software Defined Networking (SDN) to the multi-vendor parts of the global network that orchestrates all mobile and fixed data and voice traffic. Specifically, Vodafone engineers applied the software to both the Optical Network, which converts data into light for transmission at very high speeds, and the Internet Protocol (IP) Network that securely allows computers and devices to connect to the Internet.
The new technology will lead to a network that is fully automated and programmable and behaves more like a super-computer. Changes to more than 620 multi-vendor network platforms can be achieved using software-driven commands, virtually stored in Vodafone’s secure private cloud.
Johan Wibergh, Chief Technology Officer of Vodafone, said: “This software upgrade gives us a single view on the section of the transport network connecting people and machines globally. It will allow us to provide even faster and more secure connectivity across Europe and to other regions. We can continually and automatically adapt to dynamic peaks in traffic worldwide, whether they are due to people returning to the office or live streaming major sports events.”
With data traffic over Vodafone’s fixed and mobile networks continuing to grow by more than 15% per year, Vodafone can now more easily meet demand. And through the introduction of advanced machine learning and artificial intelligence, the company will in the future be able to anticipate precisely where it is needed.
Read more on - ComputerWeekly.com
Software Defined Networking
Software-Defined Networking (SDN) is a method for networking that uses software-based controllers or application programming interfaces (APIs) to communicate with underlying hardware infrastructure and direct traffic on a network.
SDN differs from traditional networks, which use dedicated hardware devices (i.e., routers and switches) to control network traffic. SDN can create a virtual network or control hardware via software.
The most significant feature of SDN is its ability to allow flexible control of networks, which is essential for the operation of large-scale computer networks such as data centers and supercomputers.
Supercomputers mostly use MPI (Message Passing Interface), which offers a suite of APIs for inter-process communication among parallel processes. In the past, some of us had tried to speed up the MPI collective communications as MPI_Bcast and MPI_Allreduce, taking advantage of network programmability brought by Software Defined Networking (SDN).