ArmoniK Newsletter - April 2026

🚀 ArmoniK 2.23.0 — New LTS Release Now Available

We’re excited to announce ArmoniK 2.23.0, the new Long Term Support (LTS) release of ArmoniK.

This version represents a major milestone, delivering improved stability, higher performance for data-intensive workloads, and a cleaner, more maintainable operational model. 2.23.0 is now the recommended version for all production deployments.

⏳ Support & Lifecycle

• ArmoniK 2.23.0 is the new LTS
• ArmoniK 2.22.x remains supported until June 3rd, 2026

🧩 Official SDK Support

This release officially supports the following SDKs (exact versions defined in versions.tfvars.json):

• C#
• Java
• C++

All SDKs benefit from tighter integration, better performance, and long-term compatibility guarantees.

🌐 Official Load Balancer Support

ArmoniK 2.23.0 introduces official support for the ArmoniK Load Balancer, featuring:

• Native Kubernetes deployment
• Seamless integration with ingress, TLS/mTLS, and the GUI
• First-class support in reference Kubernetes deployments

This significantly simplifies secure, production-grade cluster exposure.

⚠️ Breaking Changes You Should Know About

Declarative Initialization (Authentication & Partitions)

Initialization is now fully declarative and idempotent:

• Authentication (roles, users, certificates)
• Partitions

All configuration is applied at startup using structured environment variables—no more manual MongoDB edits. This enables true infrastructure-as-code workflows with Helm or Terraform.

Distroless Container Images

• All orchestrator images are now distroless
• Important: agent and worker containers must explicitly use the same user ID to avoid filesystem permission issues

This change reduces CVEs, image size, and operational risk.

⚡ Major Performance Improvements

Node-Level Cache for Agents

A new node-level cache dramatically improves performance for data-intensive workloads:

• Faster task startup
• Reduced object storage traffic and costs

Faster Streaming Transfers

• Fewer gRPC messages and lower CPU overhead

C++ SDK Enhancements

• Parallelized upload/download operations
• Reduced network latency through batching and concurrency (Note: client handlers must now be thread-safe)

🧠 Core Platform Improvements

• Upgrade to .NET 10 (LTS) for orchestration services
• New NATS JetStream adapter
• Removal of legacy RabbitMQ AMQP 0.9.1 adapter

🏗 Infrastructure Updates

• Unified Ingress + GUI module, including Load Balancer deployment
• New Percona MongoDB support via the PSMDB Operator

✨ Redesigned C# SDK

The C# SDK has been fully redesigned for clarity and long-term maintainability:

• Explicit task-graph definitions
• Unified blob data abstraction
• Modern async workflows
• Cleaner worker model with dedicated static and dynamic workers

Compared to the archived extension, the new SDK offers better debuggability, predictability, and control, making it the preferred choice for future development.

🛣 Roadmap & Upcoming Activities

ArmoniK is driving the next wave of distributed computing with high-impact projects and collaborations:

ArmoniK User Group — Inaugural Session (April 17th, 2026)

Join us for the first-ever ArmoniK User Group session, created to bring together users, developers, and partners around the ArmoniK ecosystem.

This inaugural session hosted in Paris by Natixis, aims to:

• Share best practices for deploying and operating ArmoniK in production (sessions from CA-CIB and Natixis under preparation)
• Present the newly established roadmap for late 2026 and early 2027
• Collect direct user feedback, expectations, and feature requests
• Launch a space for collaboration, knowledge exchange, and community building

Several POCs with European Financial Services Institutions are underway

Use cases are tested to demonstrate ArmoniK’s value in mission-critical environments and will lead to potential additional clients in 2026 :

• Scalable, reliable workflows for complex computations
• Faster, cost-efficient analytics and risk simulations
• Improved throughput and operational resilience

Formal Verification for Distributed Workflows

Strengthening correctness and predictability:

• Guarantees task dependencies, execution order, and results
• Reduces risk of errors in large-scale deployments
• Enables auditable, production-grade workflows

Expanding into AI Inference Orchestration

Enabling distributed, high-performance AI workloads:

• Scale and manage multiple AI models concurrently
• Utilize heterogeneous computing resources
• Integrate seamlessly with ArmoniK SDKs for automated scheduling and aggregation
• Target AI-heavy industries: finance, imaging, geosciences

Internship: Accelerated Seismic Imaging with Generative AI

Hands-on project for students and early-career researchers:

• Boost seismic imaging using distributed workflows + generative AI
• Gain experience in Kubernetes, C# SDK, and task orchestration
• Collaborate directly with ArmoniK engineers and researchers

✅ Ready for Production

With declarative initialization, official load balancer support, major performance gains, and long-term platform alignment, ArmoniK 2.23.0 is the most robust and future-proof release to date.

👉 We strongly encourage all users to plan their upgrade and adopt 2.23.0 for production workloads.