How autonomous network operations are enabled with Google Cloud and Ericsson (demo)

Key Takeaways

[Music] Hello, my name is uh Mammadul Kamar. I'm a solution architect in the uh Google cloud industry team out of uh Texas. Uh and today I'll be uh showcasing our Ericson intentbased uh service management. Uh this is a technology exploration that uh we're doing with the Ericson team. So Google and uh Ericson have uh really uh long-standing relationship uh where we actually uh have solution from Ericson that are deployed onto uh on top of Google but in addition to that we do also technology exploration in different area. So today we'll be uh talking about our collaboration in terms of uh managing uh automatically network uh services. So here you can see uh an architecture for this uh technology collaboration where we have uh multiple agents that are actually uh communicating each other and then we'll be focusing on the uh service design but also on the cognitive AI solution which is deployed on top of GCP. So the uh cognitive AI solution from Ericson um is going to be leveraging our technologies uh AI technologies like Gemini and also Vert.ex AI. Uh so we'll uh dive deeper into the uh use case over here. So we have a scenario where the uh we want to actually uh create connectivity right so we have a hospital here that has emergency services and they actually need to uh request connectivity. So what they have to do is just to state their business requirements. uh once we uh once we actually have those business requirements our Gemini is actually able to recognize the intent and then map the vision for the intent. So that means technically based on the use case if it's a streaming use case if it's voice communication it will know what type of uh settings that needs to be actually created on the network. So in this case could be uh anything like uh the latency requirements, the bandwidth requirements um etc. So we can see here we're actually processing the require the requirements from the operator and then based on that we're defining the uh configuration that needs to be implemented in the network. So once we have all of these settings the uh Ericson service orchestration is going to kick in and then make the actual change into the network. So it has some standard interfaces TM forum based interfaces open interfaces that are actually used to communicate with the different layers uh in the uh network site. So a little bit more so you can see here we have once the u network slice is deployed we also keep monitoring it. So that means if there is some any changing condition in the network, we will actually have our measuring agents that will actually uh gather those uh metrics observability and then use that to actually suggest some changes. So this could be done in an open loop or it could be done in closed loop automation as well. So going back to the uh interface for this use case, we could see here that we actually have TM forum u compliant interface. So that means as long as the uh on the network side whatever network subsystem you have as long as they are compliant with DTM forum uh standard APIs we'll be able to do that integration. So in addition to that we could actually look at the uh overall loop which actually shows the different agents that actually free your operations team from manually managing those network dynamically. So you can see the different agents that actually leverage our agent technology and then they communicate to actually achieve the end use case. Uh and then thank you for your attention and uh and feel free to reach out to our Google team for more on this uh use case. [Music]