09:09:24 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : if possible admin can mute all..and open during QnA 09:10:37 From Rogier Noldus (Ericsson & TUD) : I think that the pre-OPEN-RAN era did not pose a restriction to innovation. Now that we arrive at 5G RAN, the RAN has become more elaborate - the RAN is becoming a network on itself. This requires redistribution of the various functions (RRH, BBU-D, BBU-C). 09:14:08 From Sudhakar B, IISc : OpenFronthaul is only between DU-RU. Is there a reason to indicate the interface from SMO to RU?? 09:15:10 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Replying to "OpenFronthaul is onl..." RU Managament PLane is specific to each RU vendor as f now hence dedicated OFH-MPlane 09:16:19 From Sudhakar B, IISc : Replying to "OpenFronthaul is onl..." Thank you Bhavin 09:16:30 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Reacted to "Thank you Bhavin" with šŸ‘ 09:17:00 From Rogier Noldus (Ericsson & TUD) : The interface between BBU-D and RRH, namely eCPRI, is not standardised by O-RAN. Wouldn't it be better when eCPRI is part of O-RAN? 09:27:27 From Navid Nikaein : In the pre-Open-RAN era, developers has no role, and it was difficult to deploy a multi-vendor and change the network definition as the requirements change. It is true that RAN itself is becoming a network. Does it make sense for you Rogier? 09:29:48 From Navid Nikaein : The interface between SMO and RU is a new interface that O-RAN tries to specifies in an effort to converge O1. the benefit would be more for the operators to manage the lifecycle of the whole E2E network, let us know if you agree? 09:30:01 From Soliman, Mohammed | Mohamed | RMI : is the network element as cu, du, amf, etc are represented as pods? 09:31:11 From Navid Nikaein : Replying to "OpenFronthaul is onl..." the MPlane is the current status of how to manage an RU, which as I said O-RAN tries to converge it with O1. 09:31:33 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Reacted to "the MPlane is the cu..." with šŸ‘ 09:31:43 From Sudhakar B, IISc : Reacted to "the MPlane is the cu..." with šŸ‘ 09:32:50 From Navid Nikaein : eCPRI does not allow multiplexing gain to happen at DU, which is why it is not considered by O-RAN 09:34:21 From Rogier Noldus (Ericsson & TUD) : Replying to "In the pre-Open-RA..." It is true that pre-OpenRAN, the coponents in the RAN had to be from one vendor, but that did nt really pose a problem. I see the Open-RAN architecture as a natural evolution in the 5G RAN. With 5G, Open-RAN starts to make sense. 09:35:16 From Sajjad : Navid you are mute 09:37:01 From Alireza MOHAMMADI (EURECOM) : Yes. They are all deployed in Pods on top a Kubernetes cluster, including the simulated UEs. 09:39:05 From Navid Nikaein : Each network elements is deployed as containers in a pod, and placed in different compute nodes in the cloud infrastructure. Note that the infrastructure is distributed and the placement is automatic. 09:39:34 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : what is the cloud platform used? 09:39:54 From Rogier Noldus (Ericsson & TUD) : Replying to "eCPRI does not all..." Ok, but then we still have the situation that there is no standard for the interface between BBU-DU and RRH. That may hamper multi-vendor RAN. 09:41:06 From Navid Nikaein : Replying to "In the pre-Open-RAN ..." I understand your point. Maybe what I could add is the market segment that has been evolved from just public network to public/private network, which calls for more diversity of the vendors and more openness. Agreed. 09:41:41 From Mutua : How do I get my vanilla k8s cluster to accept this deployment? Iā€™m thinking networking 09:43:34 From Navid Nikaein : Replying to "what is the cloud pl..." Telco-optimized Kubernetes providing synchronization, BGP-optimized data-plane, and auto-device discovery such as RU 09:44:32 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Reacted to "Telco-optimized Kube..." with šŸ‘ 09:45:33 From Navid Nikaein : Replying to "How do I get my vani..." Note that Kubernetes is just a control plane, the SMO (Mx-Operator) is the one aware of the 5G entities and E2E network service able to manage the whole lifecycle 09:50:27 From Navid Nikaein : Here is the relation between K8S control plane and Operator. Also have a look at https://kubernetes.io/docs/concepts/extend-kubernetes/operator/ 09:51:07 From Mutua : thanks 09:51:07 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : this look like a centralized NearRT RIC implementation 09:51:27 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : is it possible to also implement the Distributed NearRT RIC implementation?? 09:52:12 From Navid Nikaein : Replying to "is it possible to al..." Could you elaborate more on what do you mean by distributed Near-RT RIC? 09:53:00 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Replying to "is it possible to al..." Distributed nRT RIC = I mean separate NEarRT RIC instance for each E2 node, so DU(nRT1), CU-UP(nRT2), CU-CP(nRT3)ā€¦ā€¦ā€¦. 09:55:22 From Navid Nikaein : Replying to "is it possible to al..." FlexRIC is developed in a form of software development kit (SDK) consisting of a server library and an agent library with two optional extensions: controller-internal Applications (iApps) and communication interfaces (refer to figure below). The SDK facilitates the realization of specialized SD-RAN controllers to target specific use cases, while being simple to use. You can realize a distributed near-RT RIC with FlexRIC, but I imagine that the logic in xApps may become more complex 09:56:01 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Reacted to "FlexRIC is developed..." with šŸ‘ 09:56:04 From Alireza : Reacted to "FlexRIC is developed..." with šŸ‘ 09:57:04 From Carlo Centofanti : Where will you share the video recording of this talk? 09:57:36 From Julian Jimenez : Reacted to "Where will you share..." with šŸ‘ 09:57:49 From Rogier Noldus (Ericsson & TUD) : Is the sound working? 09:58:10 From slin : Is FlexRIC open sourced? 09:58:10 From Rogier Noldus (Ericsson & TUD) : Silence... 09:58:39 From Navid Nikaein : Replying to "Is the sound working..." it seems so 09:58:40 From Ilias Chatzistefanidis (EURECOM) : Is the sound ok? 09:58:44 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : able to hear loud and clear 09:58:49 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Reacted to "Is the sound ok?" with šŸ‘ 09:58:49 From Navid Nikaein : Reacted to "able to hear loud an..." with šŸ‘ 09:59:13 From Alireza : Reacted to "Is the sound ok?" with šŸ‘ 10:01:21 From Navid Nikaein : Replying to "is it possible to al..." Yes, it is. https://gitlab.eurecom.fr/mosaic5g/flexric 10:02:39 From Rogier Noldus (Ericsson & TUD) : Is there an API defined for developing xApps? 10:04:42 From Alireza MOHAMMADI (EURECOM) : In the platform you're able to scale your RIC to have more instances and scope them into a specific set of E2 nodes. In that sense while remining complient to O-RAN you could also explore more on the distributed RIC ideas. 10:05:43 From slin : It seems for applicatoin developers, go or python are more popular choices, here you write your xApp in C, is it purely for performance consideration or are there other reasons? 10:05:47 From Bhavin Gajjar RAN Solution Architect-Freelance Consultant : Replying to "is it possible to al..." can we see the amount of resources it utilizes per Cell? wrt pcores used, memory used...etc 10:07:10 From ChiehChun Chen (Eurecom) : Replying to "Is there an API de..." yes, there are four main APIs used to develop xApp in c: https://gitlab.eurecom.fr/mosaic5g/flexric/-/blob/master/src/xApp/e42_xapp_api.h 1. init -> init_xapp_api() 2. get stats -> e2_nodes_xapp_api() 3. subscribe to SM -> report_sm_xapp_api() 4. stop -> try_stop_xapp_api() 10:12:07 From ChiehChun Chen (Eurecom) : Replying to "It seems for appli..." The baseline of xApp is developed in C, which is called C SDK in the FlexRIC, and the swig is devleoped in C SDK to support developing xApps in different languages depening on the xApp developr. Based on the C SDK, the api provide for the other language are used the same api as C, which reduces the complexity to develop different languges xApp 10:12:48 From Alireza MOHAMMADI (EURECOM) : Open RAN Studio installs the SMO into your cluster which seamlessly integrates to K8s. Then you would be able to declare your deployments like how you normally interact with K8s or via our CLI. The SMO (Mx-Operator) handles any other thing needed including the networking setup. 10:14:34 From Alireza MOHAMMADI (EURECOM) : Yes. The platform produces all the required metrics including the compute, RAN or CN stats, cost and energy of the deployment. The metrics are provided in the premimum versions that work base on the same foundation as the ORS. 10:15:34 From Rogier Noldus (Ericsson & TUD) : Is Open RAN 2.0 a formal term? 10:15:52 From Rogier Noldus (Ericsson & TUD) : ( i.e. defined by O-RAN organisation ) 10:20:06 From Hubert Djuitcheu (RPTU) : how important should be the security of Open RAN regarding the near RT RIC, and do you think this topic has good Futur? 10:21:55 From Alireza MOHAMMADI (EURECOM) : The O-RAN specifications have not developed much on the security. Given our cloud-native system and the xApps that are independent running containers, one could imagine the usual challenges in the cloud-native as well as the access control. 10:22:34 From Alireza MOHAMMADI (EURECOM) : If you want to investigate more on this please feel free to contact us to hear more about the efforts we made in this direction. 10:24:04 From Hubert Djuitcheu (RPTU) : Replying to "how important should..." thank you I will contact you, then I am actually working on this topic and want some support from experts 10:24:38 From Ali Musab : If we use OAI then the current Flexric Patch is locked to Week-41 branch. When will the new release will come? as we cannot rebase it to the branch of OAI we want due to the binaries in the patch. 10:24:39 From Alireza MOHAMMADI (EURECOM) : Looking forward to hearing from you. 10:25:09 From Hubert Djuitcheu (RPTU) : Reacted to "Looking forward to h..." with šŸ‘šŸ¾ 10:25:55 From NJONGSSI KOUAM : We value your feedback Feedback form: https://forms.gle/bKuGTFPoLByZCsPP6 Follow us: https://www.linkedin.com/company/bubbleran Docs: https://bubbleran.com/docs/ 10:28:32 From Ali Musab : My question was not answered 10:28:48 From NJONGSSI KOUAM : I'm working for load balancing in cloudRAN , please can I have some advice? 10:28:52 From Ali Musab : If we use OAI then the current Flexric Patch is locked to Week-41 branch. When will the new release will come? as we cannot rebase it to the branch of OAI we want due to the binaries in the patch. 10:28:59 From ChiehChun Chen (Eurecom) : Replying to "If we use OAI then..." We are looking for the next versino of OAI which is merged the E2 agent to port with Thrid party RIC platform. https://gitlab.eurecom.fr/oai/openairinterface5g/-/merge_requests/1902 For the patch currently provided by the FlexRIC, we will update it when we find a good version of latest OAI. Do you have any speicifc version of OAI nned to rebase? 10:29:53 From Ali Musab : Replying to "If we use OAI then t..." latest stable ver of OAI is needed as the UE causes problems in the week-41 10:30:10 From Ali Musab : Replying to "If we use OAI then t..." multiple UEs have isseus 10:30:24 From Ilias Chatzistefanidis (EURECOM) : https://forms.gle/bKuGTFPoLByZCsPP6 10:30:33 From Amirreza Ghafoori : yes 10:30:34 From Ali Musab : Replying to "If we use OAI then t..." the link you provided is 4 months old 10:30:57 From Hubert Djuitcheu (RPTU) : thank you 10:31:00 From Muhammad Mohtasim Sajjad (QUT) : Thank you