Srinivasa Chaganti

Network Data Link (NetDL) streaming network data for advanced functionality A central piece of Arista’s strategy is the Network Data Link (NetDL) architecture. Chief Technology Officer Ken Duda explained during the call that data is the key to understanding and improving network operations. He emphasized the importance of gathering and storing large data sets for machine learning and AI. That’s where Arista’s NetDL technology comes into play. “With NetDL we continually monitor every device, not simply taking snapshots, but rather streaming every network event, every counter, every piece of data in real time, archiving a full history in NetDL,” Duda said. Duda added that NetDL also collects flow data and inbound network telemetry data gathered by switches. All that data is then enriched with user, service and application layer data from external sources outside the network. All that data helps Arista and its users to understand not just how each part of the network is performing but also which users are using the network for what purposes. TOM SHAW

2

2 Comments

Upgrading to a #Nocode OSS is necessary to achieve GenAI-powered #autonomous #networks. In order to make (or recommend) incremental improvements to processes and connectors, you need your AI agent to understand what's going on. Symphonica, the world's first AI-native #nocode OSS, purposefully stores all workflows and connectors in JSON files that a GenAI agent can understand and improve. Download this easy-to-read whitepaper to learn more: https://lnkd.in/eFKaET8f Welcome to the Agile Telco generation! Get in touch if your OSS has become a bottleneck to innovation. #ai #genai #autonomousnetworks #AN #nocode #oss #telco #agile

55

6 Comments

O-RAN network operators who aren’t using the Qualcomm X100 5G RAN Accelerator are leaving money on the table. Learn about the massive savings you can expect with the use of this network card in our new whitepaper: #TeamQualcomm 

36

Why managing conflict between rApps is vital for RAN automation at scale Intelligent automation using O-RAN Alliance rApps has the potential to revolutionize the world of network management and orchestration. O-RAN Alliance rApps can automate radio access networks at scale, leveraging rich data and AI/ML to automate use cases such as network healing, network optimization, outage compensation, performance anomaly detection, energy saving and many more. As CSPs deploy rApps at scale, the need to co-ordinate between them to ensure optimal behavior increases. In this blog we consider strategies for intelligent co-ordination between different rApps to avoid conflicts between them. Read below article to understand more. https://lnkd.in/dBrZetvu

9

1 Comment

Scaling the real time critical data path of the 5G NR RAN is very crucial as we scale the RAN to support higher throughputs and scheduled users per TTI. In this video I discuss how data decomposition can be used to scale the 5G NR real time critical data part

62

2 Comments

“The network is not an afterthought, it’s critical” in delivering the capacity for AI – as my colleague Amit Malik says. Every new AI application generates more data, escalating demands on the network. While data usage is surging and data centers are expanding rapidly, it is the network that is the most important element here. Amit Malik emphasizes the need for networks to swiftly adapt to meet user requirements. As I’ve said before, the networks of today are insufficient for tomorrow’s needs. To quote Amit, networks need to be adaptive: scalable, secure, sustainable, self-optimizing and self-configurable. https://ow.ly/5yTf50RJV7R #Network #Networks #Telco #Telecoms #AI

108

3 Comments

The surging demand for AI data centers will put a strain on underlying infrastructure. Infinera’s Russ Esmacher spoke to Next Gen Infrastructure about the huge need for solutions and tools to stay ahead of this curve of massive bandwidth growth. Watch now: #DataCenter #AI #ArtificialIntelligence

6

Private 5G needs to be super simple to use. There needs to be a single pane of glass to orchestrate & manage all the P5G components. Moreover, MSPs and large enterprises need a single pane of glass to manage all their P5G networks. Aarna Networks Multi Cluster Orchestration Platform (AMCOP) provides both of the above functionalities. Check this out using our experience kit.

18

On the Infinera blog, Jon Baldry covers his top four takeaways from this year’s series of Global PartnerPlus Xchanges, from the amazing breadth of application coverage to the amazing range of partners. Read the post now: #PartnerPlusProgram #InfineraXchange #OpticalNetworking

12

The surging demand for AI data centers will put a strain on underlying infrastructure. Infinera’s Russ Esmacher spoke to Next Gen Infrastructure about the huge need for solutions and tools to stay ahead of this curve of massive bandwidth growth. Watch now: #DataCenter #AI #ArtificialIntelligence

7

Once market leader BSNL, now not only progressing towards indeginious 4G but also strongly working towards indeginious 5G…BSNL walking past all the odds in the competitive telecom market…. Switch to BSNL…. #BSNL #CONNECTING #INDIA #CONNECTING #HEARTS https://lnkd.in/g2D4XCRv

2

3 types of isolation: 1. one workload fron the other 2. workload to host 3. but harder is host to workload 3. cannot be done without diving in silicon / chip level protection. #Attestation is core and need to be verifiable by 3rd party #Confidential #Conputing #Identiverse #TEE

1

1 Comment

Please check out my two talks below.

25

1 Comment

It was great to talk to Brandon Larson about Mavenir approach to automation and autonomous networks, and how it relates to the transition to open networks (including Open RAN, but way more than that). https://lnkd.in/e2pgNU55

4

J.D. Power’s survey found that 5G home internet was roughly on par with fiber for performance and reliability. That’s a pretty shocking finding on its own. Fiber’s primary selling point is its superior speed and consistency, but evidently customers aren’t seeing much difference between the two. Fixed wireless particularly stood out on cost of service. 5G Home internet led fiber in cost-of-service by more than 60 points -- roughly the same differential as the first- and last-placed providers in J.D. Power’s study of internet providers in the North Central region.

1

Expanding the vendor ecosystem is one of the big goals of Open RAN. Here is a blog I wrote at Award Solutions about how the RAN is being disaggregated in 5G. But disaggregation is not without its challenges for operators. Feel free to share your thoughts on the challenges that RAN disaggregation poses in comments below. #openran #5G

5

Are you prepared for the wireless tech revolution that's reshaping our cities? With the advent of 5G and the horizon of 6G looming, we're not just stepping, but leaping into a new era where connectivity is faster, smarter, and more ubiquitous than ever. Imagine smart grids that respond in real-time to energy demands or IoT networks so intricate they manage everything from traffic flows to public safety—all enabled by next-gen wireless technology. But here's where it gets really interesting: the potential for these technologies to democratize access to information and services, bridging digital divides and fostering inclusivity. However, this future isn't guaranteed; it requires deliberate strategy, investment, and collaboration across sectors. Let's harness the transformative power of wireless innovation to create cities that aren't just connected but are truly inclusive and sustainable. The future is calling—how will we answer?

3

1 Comment

Synergy in Private 5G and AI Cloud The RAN-in-the-Cloud solution offers an answer to the challenge of underutilization of vRAN computing resources by introducing dynamic resource allocation, and sharing the computing resources with AI workloads. It uses the flexible architecture of cloud computing to dynamically scale network functions up or down based on real-time demand, optimizing server utilization and reducing costs. At last week's Linux Foundation ONE Summit 2024, which I attended, NVIDIA and Aarna Networks jointly introduced a great solution for the Private '5G and AI' Cloud scenario. Of course, their ambitions don't stop at private networks and clouds, but for this post, I cover only the private 5G network and AI cloud environment. According to the architecture they created, here is how it works: ◼ Orchestrate Core, RAN Functions, Prometheus Agents: The central orchestrator (Aarna Networks' AMCOP's BPMN Workflow Engine in this case) manages the deployment and orchestration of Core and RAN functions. It also ensures that Prometheus agents are set up for network monitoring. The Core and RAN (CU1, CU2, DU1, DU2) functions are set up on two separate servers, labeled gNB S1 and gNB S2, along with the Prometheus agents. ◼ Telemetry Agents Report RAN Workload Level Metrics: Admin User configures the Threshold values on OPA (Open Policy Agent), for closed loop to be activated. The Threshold Checking Application (TCA) receives telemetry data via the telemetry agents in 5G Core. The telemetry data includes metrics collected from pods (isolated containers within Kubernetes clusters), providing insights into the performance and resource usage of each RAN function. The Prometheus agents deployed on each gNB S1 and gNB S2 monitor RAN functions like DU1 and DU2, sending metrics back to TCA for threshold checking. ◼ Workflow Engine Scales DU and Edge Application: Based on thresholds set, for example 20% utilization, in the OPA, the Workflow Engine (WFE) scales Distributed Units (DUs) and edge applications as needed. The scaling action takes place within the edge cloud servers. Edge Multi-Cluster Orchestrator (EMCO) and CDS (Controller Design Studio) migrates/reconfigures DU2 to Server#1 (nondisruptive), and the CU on Server#2 deleted. The Lifecycle Manager (LCM) reconfigures the fronthaul switch. This ensures that RAN traffic is routed appropriately to other RAN components after CU2 is deleted. Finally, AI/ML application is deployed on MEC Server#1, and Collectors/Kafka/OPA continue to monitor the threshold metrics for when to reverse the situation. Overall, the combination of cloud architecture and dynamic resource allocation by empowers private 5G customers to achieve high-quality, cost-effective private 5G RAN networks, and run AI workloads simultaneously. The next step for nVidia/AarnaNetworks is to quickly evolve their orchestrator to Nephio. #5GandAI #Private5G #DynamicScaling #DynamicResourceManagement #nVidia #AarnaNetworks #ONAP #Nephio

120

5 Comments

Thrilled about the latest advancements in network optimization! 🔧 The team at Ericsson has developed HYDRA, a cutting-edge framework for intent-based RAN optimization using reinforcement learning (RL). HYDRA tackles the challenges of dynamic intents and environmental drift head-on, ensuring faster, scalable, and safer RL training and deployment. This innovation allows for quicker adaptation to operator preferences and more efficient utilization of network resources. HYDRA's data-centric approach and offline RL capabilities are game-changers, enabling proactive policy training and seamless deployment. What excites me the most is HYDRA's ability to automate the entire RL workflow, from data collection to model deployment, with robust risk mitigation and visualization tools. The future of autonomous network management looks bright with HYDRA leading the way, ensuring that RL models remain effective and adaptable over time. 🌐💡 How do you think HYDRA's intent-based RAN optimization will impact the future of network management? #AI #NetworkOptimization #RL #Ericsson #Innovation #TeamEricsson

15

1 Comment