More information
Description
Data on network utilization and traffic is valuable to optimize network performance. However, collecting direct network measurements or other related raw data from SDN controllers is currently too expensive to be feasible, especially in the case of large-scale high-speed networks (such as 5G). Moreover, the available disk-based or in-memory cache solutions found in SDN controllers are not optimized for the particular data encountered in this context.
The R&D project 5G-data by AppArt, aims to address this issue by investigating the properties of network traffic data available in Software Defined Networks and evaluate/combine various streaming compression techniques to reduce their storage requirements. The key objective is to develop a novel SDN controller that utilizes the most appropriate compression techniques to minimize the storage and bandwidth requirements needed to handle network traffic data, which can then be further analyzed to gain insights on how to optimize the overall network utilization and performance.
5 new jobs will be created by carrying out the project. Additionally, it will enhance the competitiveness of the company in the beneficiary country and increase the revenues.
Summary of project results
5G networks have largely replaced the previous generation 4G networks in Europe. The new networks have the capacity for higher download speeds, with a peak speed of 10 gibabits per second (Gbit/s). 5G also has higher bandwidth to deliver faster speeds than 4G and can connect to more devices, thus improving the quality of Internet services in crowded areas. 5G, as well as high-speed fiber optic cable internet, has also contributed to the advancement of applications related to the Internet-of-Things (IoT) as well as Machine-Learning (ML) and Artificial Intelligence (AI).
An SDN (Software Defined Networks) controller is the application that acts as a strategic control point in a Software-Defined Network. Essentially, it is the “brain” of the network. SDN controllers enable cloud-like computing within a network. This allows network engineers and administrators to respond quickly to changes in business requirements through a centralized control console that is not physically linked to the hardware of the network. In other words, SDN acts a centralized "brain" for the network that can communicate and command the rest of the network.
Data related to the a network''s utilization and traffic is valuable to optimize network performance. However, collecting direct network measurements or other related raw data from SDN controllers is currently too expensive to be feasible, especially in the case of large-scale high-speed networks (such as 5G). Moreover, the available disk-based or in-memory cache solutions found in SDN controllers are not optimized for the particular data encountered in this context.
The R&D project 5G-data by AppArt, addressed this issue by investigating the properties of network traffic data available in Software Defined Networks and evaluated/combined various streaming compression techniques to reduce their storage requirements so they are more easily stored and processed. The key objective was to develop a novel SDN controller that utilizes the most appropriate compression techniques to minimize the storage and bandwidth requirements needed to handle network traffic data, which can then be further analyzed to gain insights on how to optimize the overall network utilization and performance.
The project comprised of the following activities:
- Study of the communication protocols between the control layer and the data and application layers (through an SDN controller) and generation of large volume synthetic data to be used during development and evaluation of the compression techniques.
- Study and analyse different compression techniques to identify common properties of network information (identify patterns in the information in the control plane of SDNs), select low-cost compression schemes in terms of requirements, and determine the most appropriate schemes for each type of network information.
- Implement appropriate compression techniques in low-level libraries and ultimately in SDNs
- Evaluation of the improvements provided by the proposed solution in terms of the achieved compression and the reduced bandwidth required.
- Develop the interfaces and finalise the SDN controller with integrated compression capabilities
- Implementation of a "demo" version of the platform (SDN controller), suitable for demonstrations in conferences or exhibitions. Developed a lightweight, portable environment to host and execute the offering in such a setting.
The novel SDN controller with compression capabilities will allow the analysis of the operation of software-defined networks such as 5G, and will offer valuable insights to network operators for its use and optimisation, therefore offering a better service to its users as well.
The project also led to the creation of at least 3 new permanent full-time positions for AppArt, and it will contribute to increased turnover and profitability for the company.