South Korean telecom giant SK Telecom worked with Intel on a new network core technology that alters current cloud-native principles and uses artificial intelligence (AI) to help slash network latency for 6G networks.

The technology is dubbed Inline Service Mesh and is designed to work in a still-to-be-developed 6G network core architecture. The technology uses Intel Xeon processors and AI to reduce latency between unit function and increase the communication speed within the 6G core network architecture.

In a white paper tied to the work, the firms explain that current cloud-native architectures used in the construction of 5G networks rely on existing container sidecar proxies to manage network functions. While these allow for greater service flexibility due to the ability to use this containerized architecture to support multiple functions, this also adds latency to control plane processing.

“This latency overhead is caused by the current industry standard approach of using a service mesh sidecar Proxy container in microservice pods. This has become a pain point that needs to be addressed in the deployment of 5G core network functions (NFs) on cloud-native platforms,” the white paper explains. “If this latency problem can be solved, the service mesh has the potential to provide greater benefits to 5G core deployments, especially as we move forward to 6G and AI/machine learning (ML) [ML] integrated core network.”

The duo’s fix for this latency issue is an inline service mesh instead of a sidecar Proxy like Envoy, which they claim can cut latency for transactions between the session management function (SMF) gateway and Packet data unit (PDU) session microservices by 70% and gateway CPU usage by 33%. This latency reduction is expected to become more important as network functions continue to be disaggregated into smaller microservices.

AI is also central to this work as a higher level of automation will be needed to manage and direct a growing number of network functions running through a 6G network core that are expected to fill this latency advantage.

“As core network technology continues to develop, the various systems that make up the network and the detailed functions that provide various services are also explosively increasing,” the firms noted. “As network complexity continues to increase, the process of sending and receiving messages is frequently recreated. Therefore, communication delays will increase compared to before. It is difficult to address these limitations with communication standard technologies (service communication proxies) for interconnection between unit functions within the existing core network.”

SK Telecom had previously tapped Intel on work targeted at reducing latency with 5G technology. That work was followed up by a partnership to develop an network functions virtualization (nfv) platform targeted at 5G deployments.

SK Telecom said it plans to use results from this latest study with Intel in commercial equipment beginning next year. That’s well ahead of the timeline that standards bodies have in mind for developing specific 6G standards.

SK Telecom and Intel’s efforts are somewhat similar to what the Istio community is working on with its Ambient Mesh project. That project was initially unveiled in late 2022, and is focused on simplifying what the Istio service mesh looks like by allowing users to “forgo sidecar proxies in favor of a mesh data plane that’s integrated into their infrastructure.”

“It is significantly easier to adopt and maintain and represents a huge cost reduction,” Louis Ryan, CTO at Solo.io and co-creator of Istio, told SDxCentral last year. “We called the effort ‘ambient’ because we think mesh should be part of your infrastructure furniture, something that’s always there giving value.”