Видео

Thank you for your message. Conflict mitigation in Non-RT RIC is also currently being worked on in O-RAN WG2.

This is inhouse simulator with a simplified model for the RICs. It is developed fully in Python.

The dashboard and the incorporation of the map was our internal implementation.

Thanks for your question. Y1 interface for exposure of analytics information from Near-RT RIC to authorized consumers, like packet core functions, application servers or edge servers. Example use cases are (from ["The Evolution of Open RAN", A 5G Americas Whitepaper, Feb 2023]): • Predicted UE-level RAN throughput, exposed by the RIC, when consumed by an application server such as a video streaming server, shall enable the server towards proactively and intelligently optimizing the video bitrate resolution for the streaming video content to the UE that minimizes buffering, stalling, etc., thereby subsequently improving the average QoE of the video streaming UEs. • Predicted UE-level RAN latency, exposed by the RIC, when consumed by an application server such as a 360-degree virtual reality (VR) streaming server, shall enable the server towards proactively and intelligently optimizing the IP packet size for the VR content that minimizes screen freezing, screen blackouts etc., thereby subsequently improving the average QoE of the VR UEs.

This is a commercial demo, thus not available for public use. However, there is an open source Energy Saving with Traffic Steering demo available at SD-RAN platform: github.com/opennetworkinglab/smart5g-nonrtric-plt-ranpm/wiki

@@RimedoLabs Thanks! I tried this in an 18.04 VM but had some issues. Is this the right version of ubuntu to try?

This code is maintained by ONF under SMaRT-5G project (opennetworking.org/sustainable-5g/) please reach out to them.

Thank you for those questions. As the presented work is somehow in progress they will help us to develop our ideas. Now, touching on your points: 1. We have arbitrarily assumed that we deal with an IIoT network. While considering the public network it could be that there is a network slice dedicated to such a device that is apriori known to be stationary, i.e., this slice does not support mobility functions by default. Another case is a private network with a single BS to provide a 5G network for the factory. Under such conditions, IIoT devices are considered to be static sensors for some measurement purposes. But of course, it would be good to generalize the presented idea using some advanced classification of devices. 2. This is a challenging case if we assume that detection is done based on Random Access Response where only TA is present. However, if we consider a device connected to the network, more parameters can be observed from RRC measurement reports. On the other hand, long-term statistics of TA could show some cyclic properties, i.e., routes of industrial devices are similar each day, etc. From this perspective maybe it could be solved by improving the KPI profiles.

Thanks for your question and comment. In this use case the O-RU is switched off.

Thanks for reply @@RimedoLabs SO can we switch off the CU and DU in O-RAN or there is no need of this bcz CU and DU runs on COTS hardware and in traditional RAN we also switched off the some components of BBU in some scenarios

In O-RAN O-CU and O-DU are virtual entities to be run on O-Cloud platform resources. Thus, intelligent management and orchestration of O-CU/O-DU HW and O-Cloud SW platform resources, coordinated by SMO, are required for the sake of EE. For this purpose, in O-RAN NES Technical Report, there is also a feature called "O-Cloud Resource Energy Saving Mode". The aim of this use case is to enable energy savings in the O-Cloud by reducing the power consumption of various O-Cloud components without impairing the network performance. Given the network status, the O Cloud components' power consumption can be optimized through actions such as adaptive shutdown of hardware, scaling out Network Functions, and optimization of CPU/FPGA power, memory usage, CPU frequency, etc. This use case targets e.g., optimization of workload placement and automated scale-in / scale-out processes help to dynamically adapt the pool of active hardware resources to the actual workload needs and allow to free up HW resources which could be shut down in idle times. However, in our scenario, we are not considering this item.

Thanks for your question. Yes, the open source xApp is available under this link: github.com/onosproject/rimedo-ts

Hello, the xApp portal is our internal tool, thus cannot be downloaded by external parties. You can however download and run the open source code for this demo github.com/onosproject/rimedo-ts or a latest open source demo for energy saving use case: github.com/opennetworkinglab/smart5g-nonrtric-plt-ranpm/wiki

Thanks for the question. The point here is that the Near-RT RIC connection/functionality is broken. In addition, if the cloud infrastructure fails it only could fail where the RIC is deployed and that could be in a different place than the rest of the infrastructure.

Thanks for the comment.

Hello, thanks for the question. Yes, that is correct - the ES use case is not yet available in the public version of the Open RAN xApp Demonstrator. We're currently working on it.

​@@RimedoLabs Is there a prediction for it to be available?

Hello, now the ES-rApp is available in the demonstrator.

@@sergiomartins4652 Could you please clarify the question. What do you mean by "available"?

@@RimedoLabs available in the public version of the Open RAN . You say ES-rApp is now available in the demonstrator. Is it in public domain this demonstrator?

Thanks for the question. The problem with the threshold is how to obtain its level. What ML does can be understood as adjusting the threshold in an automatic manner. If we consider the real network user throughput depends on many functional blocks like scheduler, precoder, beam management etc. Moreover, the realistic massive MIMO radio channel has some correlations, and complicated reflections. Due to those facts it is hard to estimate the user throughput related to the given array configuration apriori, on the basis of standard optimization. It is necessary to capture the data (user troughputs, power consumption) related to the given antenna array, and train the ML model. Moreover, while considering massive MIMO not only the total number of users matters but their spatial distribution, especially in the angular domain. For example, the same number of users can require different array configuration when being cumulated at one specific direction, and when being spread over multiple directions. Our rApp takes multiple input features, and some of them are related to the per beam load distribution that reflects this spatial diversity of users in cell. We hope this clarifies the ML usage.

​@@RimedoLabs yes, thank you for the time :).

Yes, that is correct. There is a discussion to get interfaces between SMO or even Near-RT RIC to the NWDAF. I believe one of the important use case for this would be Network Slicing, where to get end-to-end SLA assurance this is beneficial.

@@RimedoLabs MaxLinear just has issued a new open ran 5g SoC chip called “Sierra”. A very powerful chip!!!!

Thanks, we're glad that you find the content useful. The slides are available under this link, after you subscribe to our newsletter: mailchi.mp/5cc584269396/w36l66fgmk

Thanks, we're glad you like it. The slides are available under this link, after you subscribe to our newsletter: mailchi.mp/5cc584269396/w36l66fgmk

Thanks, we're glad it's helpful.

We provide free access to the demonstrator for everyone who subscribes to our mailing list. To get access, use the following link: lnkd.in/d5mZPUMG

This is our inhouse simulator, we are using our internal processes for testing the apps.

You can find downloadable slide-decks with most of that content on our website dedicated to O-RAN: rimedolabs.com/o-ran/

Thank you!