Smart Traffic Director

Load Balancing Reinvented: Make your infrastructure Autonomous

Powered by Reinforcement Learning (RL)

End State is Infrastructure becoming autonomous

Works on requests ingressing into a cluster as well as requests going out

Smart Traffic Director

Load Balancing Reinvented: Make your infrastructure Autonomous

Powered by Reinforcement Learning (RL)

End State is Infrastructure becoming autonomous

Works on requests ingressing into a cluster as well as requests going out

Ecosystems & Integrations

What the Smart Global Load Balancer solves for you?

Avesha Smart Load Balancer can be trained to react to changes in server and network state and take actions to schedule requests for server resources in order to optimize the utilization.

Optimized Resource Utilization

Smart GLB uses Reinforcement Learning for intelligent request assignment, preventing resource waste and promoting efficient server utilization over traditional load balancing.

Autonomous Adjustment

Smart GLB autonomously adjusts to server and network changes, efficiently managing dynamic workloads and eliminating the need for manual tuning.

Improved Response to Workload Variability

Smart GLB accounts for request duration, overcoming traditional load balancing inefficiencies in mixed workloads, preventing server overload and under-utilization.

Reduced Latency

Smart GLB autonomously assigns latency-sensitive requests to suitable servers, enhancing performance and user experience by considering request type and server latencies.

Avesha Smart Global Load Balancer Architecture

Avesha Smart Global Load Balancer uses several data points such as service dependency maps, service loading data, bandwidth, and latency metrics to optimize path selection with RL.

How to create a Smart Global Load Balancer?

Here are the steps on how to enable Avesha Smart Load Balancer for intelligent path selection for service to service communication across a cluster and optimize and rightsize resource utilization of POD in a cluster.

Discover service endpoint across cluster / Multi-cloud and within a cluster. Configure telemetry servers or your favorite Application Performace tool (creds) in the Avesha portal. Avesha system provides you with required data sets to be enabled.

Identify optimization objectives: Latency, bandwidth, cost, QoE, etc.

Define tuning periodicity based on endpoints and WAN links for services and edge.

Identify/configure your GitOps tools in the Avesha portal and visualize impact in KubeSlice metrics visualization.

Discover service endpoint across cluster / Multi-cloud and within a cluster. Configure telemetry servers or your favorite Application Performace tool (creds) in the Avesha portal. Avesha system provides you with required data sets to be enabled.

Identify optimization objectives: Latency, bandwidth, cost, QoE, etc.

Define tuning periodicity based on endpoints and WAN links for services and edge.

Identify/configure your GitOps tools in the Avesha portal and visualize impact in KubeSlice metrics visualization.