KubeSlice

Term 2 · Jun-Aug · 2 projects

• KubeSlice Controller HA (Active/Standby) Support

  LFX Mentorship · Listed in the official cohort

  Source IssueLFX portal

  Add Active/Standby HA support for the KubeSlice Controller. Leader election: Only one controller cluster (the “Active” node) holds a distributed lock (e.g., a Lease) and is permitted to write updates to worker clusters or manage Slice configurations. Constant state sync: The Standby cluster continuously mirrors the Active cluster’s state, including relevant KubeSlice CRDs (Slices, ServiceExports, Clusters, etc.). Heartbeating: The Standby cluster monitors the health of the Active cluster. Failover: If the Active cluster fails to renew its lease, the Standby cluster detects the timeout, acquires the lock, promotes itself to Active, and worker clusters are updated to use the new Active controller.

  • Go
  • Kubernetes (controllers
  • CRDs
  • client-go)

• Partial Mesh Support (MVP: Hub-and-Spoke)

  LFX Mentorship · Listed in the official cohort

  Source IssueLFX portal

  Add partial mesh support to KubeSlice using a small-scope MVP topology: Hub-and-Spoke. Users can define one (or two) hub clusters for a slice; the controller computes desired connections so only hub↔spoke links are established (no spoke↔spoke), and workers reconcile connectivity accordingly. Includes status reporting for convergence and failure reasons.

  • Go
  • Kubernetes (CRDs
  • controllers
  • client-go)

Term 3 · Sep-Nov · 4 projects

• Enhance and Automate End-to-End (E2E) Testing Across the KubeSlice Ecosystem

  LFX Mentorship · Listed in the official cohort

  @gourishbiradarSource IssueLFX portal

  The KubeSlice project consists of multiple repositories that work together to provide application connectivity and network services across Kubernetes clusters. Currently, our E2E tests are outdated and need significant improvement. This project aims to automate the E2E testing process by improving the current test suite and implementing new tests where necessary. This will ensure that the entire KubeSlice ecosystem works seamlessly together.

  • Go (Golang)
  • Kubernetes
  • E2E testing (kind
  • Ginkgo)
  • CI/CD (GitHub Actions).

• Implement Comprehensive Unit & Integration Testing for kubeslice-cli

  LFX Mentorship · Listed in the official cohort

  @gourishbiradarSource IssueLFX portal

  The kubeslice-cli repository currently lacks comprehensive unit and integration tests making it hard to test the changes. This project aims to implement a robust testing framework to ensure the reliability and stability of the CLI tool. The Mentee will write unit tests for existing functions & integration tests for the CLI commands & also set up a continuous integration pipeline to run these tests automatically on every commit and pull request.

  • Go (Golang)
  • Go testing frameworks (testify)
  • Kubernetes
  • CLI
  • CI/CD (GitHub Actions).

• Implement Custom Topology Definition for a Slice

  LFX Mentorship · Listed in the official cohort

  @gourishbiradarSource IssueLFX portal

  KubeSlice currently uses a full-mesh topology for the slice overlay network, where every cluster connects to every other cluster in the slice. This results in unnecessary tunnel creation and resource consumption. This project proposes a topology-aware design, allowing users to define custom connectivity matrices for slices. Users can specify partial meshes. Additionally, the project will support configuring each cluster's VPN deployment type (client/server) to accommodate network constraints such as firewalls or NAT.

  • Go (Golang)
  • Kubernetes CRDs and controllers
  • VPN networking concepts (client/server)
  • overlay networking.

• Implement Dynamic IPAM for the Slice Overlay Network

  LFX Mentorship · Listed in the official cohort

  @gourishbiradarSource IssueLFX portal

  In the current KubeSlice design, IP address management (IPAM) for slice overlay networks is static and inefficient. A predefined CIDR block (e.g., 10.1.0.0/16) is divided into a fixed number of subnets regardless of how many clusters participate in the slice, leading to significant IP space wastage. This project aims to implement a dynamic IPAM system that allocates IP subnets to clusters on demand and reclaims unused ranges when clusters leave the slice. It will ensure efficient address utilization, synchronization across clusters, and integration with the KubeSlice control plane.

  • Go (Golang)
  • Kubernetes controllers and CRDs
  • IPAM concepts
  • networking (CIDRs
  • subnets).