ATH has developed a mobile broadband connectivity solution for enterprises that employs Commercial Off-The-Shelf (COTS) servers and 4G/5G radio equipment called Bubblecloud. Awarded with 4 GSMA’s GLOMO awards, BubbleCloud is a click-to-deploy solution that leverages the power of AWS public cloud to deliver mobile broadband connectivity to enterprises with the same installation and maintenance simplicity as a regular corporate intranet based on traditional wireless technology such as Wi-Fi. The key aspect of the solution is exploiting cloud computing to run all the control plane procedures as well as OAM part through the so-called “Control Center”, whereas the user plane is handled on premise by the “Edge Node”, able to route traffic to local application as well as to the global Internet. The following figure depicts such architecture. The Control Center is a SW platform running in AWS that customers can purchase through the corresponding subscription system. On the contrary, the edge node is a small form-factor server to be installed on premise and connected on the one hand to the radio access network (e.g., small cells network) and to the other hand to a broadband data network, e.g., cable, DSL or fiber, in order to establish the connections to the Control Center and to provide access to the Internet. BubbleCloud can be installed on a private cloud, which is employed by 5Genesis, 5G-CONNI and successfully tested at the 4th ETSI NFV Plugtests. The 4G version of the solution has been selected by the French Ministry of Interior for the PC STORM project, in the context of the French Roadmap – PPDR Broadband LTE.
CMC 5GC is based on micro-services and includes 4G and 5G NFs as shown in the following figure. The solution has been deployed as part of 5G Test Network Finland providing network infrastructure to other Finnish national projects such as 5G-VIIMA focused in smart industry and harbours. 5G NSA and SA has been developed and deployed for trials and pilots in other EU projects such as PriMO-5G and 5G-SMART. Its 5GC supports the 3GPP SBA that allows to utilize new NFs and distribute them across the network based on the service needs. The 5GC features a distributed UPF with SBA support, terminating the IP/GTP tunnels and introduce the traffic classes into dedicated L1/L2 links to be propagated through the switched Ethernet network. On top of the existing 5GC, CMC provides a mobile backhaul orchestrator (MBO) designed using SDN to deploy and manage network resources and deploy slices. The MBO design includes the modules: a) Network monitoring to check available resources and b) Machine learning engine that will apply different traffic engineering to calculate optimal rules to be set in the network switches though SDN controller.
Open5GCore toolkit is a worldwide first practical implementation of the 3GPP 5G core network. It mirrors in a prototype form the 3GPP Rel-15 and 16 as well as own Fraunhofer research for the core network functionality and its integration with 5G NR (Standalone and Non-Standalone). As shown below, Open5GCore implements the new 5G components as standalone, independent of the previous 4G EPC functionality. Open5GCore is currently used by ICT-17 5G-VINNI and 5Genesis as basis for the experimental facilities and for the 5G-PPP German node, as well as by 30 different R&D labs across the world for experimenting on new local 5G networks. Open5GCore Rel. 5 integrates with 5G NR SA, off-the-shelf LTE and NB-IoT LTE and non-3GPP access networks such as WiFi. It can be deployed with containers or virtual machines on top of a large number of virtualization environments. The required HW for a setup highly depends on the expected capacity.
One2many's Cell Broadcast Centre Function (CBCF), shown in below, is a function in the 5GC introduced in Rel-15. It supports GERAN, UTRAN, E-UTRAN and NR in parallel, and can be deployed in the cloud, e.g. an Amazon data centre (AWS). The CBCF sends its public warning messages to the selected RAN Nodes via Access and Mobility Function and uses services of the Network Repository Function (obtaining tokens for its secure communication with AMF and to be notified of status updates of AMF). The Public Warning Portal (PWP) is used to compose the warning messages, which are disseminated, amongst others, to CBCFs in mobile operator's networks.
UBITECH has developed a Vertical Applications Orchestrator (VAO) that is responsible for the deployment and runtime management of 5G-ready applications over application-aware network slices. It has been developed in MATILDA 5G-PPP project, and is going to be used in SPIDER. The VAO is characterized by a top-down approach where applications design and development leads to the instantiation of application aware-network slices, over which vertical industrial applications can be optimally served. The VAO on-boards the vertical application and evaluates a provided set of requirements (e.g. QoS/QoE constraints, computational requirements). Based on this set of requirements, the creation of a network slice can be requested to the eSBA platform (through the 3GPP Common API framework). Upon the creation of a new network slice instance, the deployment of the vertical application may take place. VAO supports a set of intelligent orchestration mechanisms, covering the real time management, analysis, and runtime policies enforcement of vertical application components (i.e. service functions). A slice intent metamodel has already been defined and described and can be found in “D1.4 Network-aware Application Graph Metamodel.,” February 2018.
InterDigital has provided input into the (enhanced) Service-Based Architecture (eSBA) work in 3GPP since its formation in Rel-15, through key issues and solutions provided to core aspects of the realization of control planes for 5G networks. Since 3GPP, “TS 23.501, System Architecture for the 5G System (5GS)”, its realization of eSBA is recognized as one of three deployment choices for eSBA. For this, its solution provides a flexible service routing solution that utilizes the name resolution of the Network Repository Function (NRF) of the 3GPP 5GS to initiate a suitable routing the CP service request from the initiator to the (virtualized) service instance, realized through the service communication proxy (SCP), as defined in 3GPP, “TS 23.501, System Architecture for the 5G System (5GS)”. The next figure depicts InterDigital eSBA platform.
This deployment choice for 5G control plane is being realized in a platform that combines the 3GPP-compliant SBA realization with a lifecycle management (LCM) solution that targets the provisioning of micro-services, e.g., realized through containers or other lightweight virtualization techniques, over the SCP component of 3GPP. The orchestration component utilizes templates compliant with TOSCA. The Service Function Endpoint Management & Control (SFEMC) components allows a flexible control capability during service lifetime, specified through initial and state change policies within the orchestration template. The 3GPP service framework labelled component implements the SBA capabilities of 3GPP, “TS 23.501, System Architecture for the 5G System (5GS)” with the NRF and SCP components realizing the service registration & discovery as well as message routing functionality. InterDigital’s platform is already being utilized for (micro-)services beyond control plane services only, as indicated in the figure above with showing control as well as user plane services. For this, the SBA interprets ANY services as a micro-service that is able to not only be managed and controlled through the platform but also to route service requests across the (operator) network.
HWDU 5G Resource Controller is capable of making request-scheduling decisions over the available 5G infrastructure. It is capable to properly address the critical aspects such as efficiency and effectiveness of the underlying SBA NFs components, especially in heterogeneous domains. HWDU 5G Resource Scheduling Platform manages the resource controller instances (and their deployment locations) to maximize the reliability and scheduling performance, ensuring the atomicity of 5G operations. Due to the scale of the network and the diverse infrastructure, a single scheduler instance is not efficient to run a large network. As pre-provisioning is intrinsically limited, all of them should operate on a common, or at least overlapping set of resources.
