ZEISS Demonstrates the Power of Scalable Workflows with Ampere Altra and SpinKube

This case study details how ZEISS Group significantly reduced its cloud computing costs by leveraging WebAssembly (WASM) and Ampere Altra processors. The challenge was the high cost of maintaining an order processing system designed to handle massive spikes in demand (up to 10,000 near-simultaneous requests) while remaining mostly idle. Traditional Kubernetes deployments, even with paused pods, resulted in substantial overprovisioning and wasted resources.

The solution involved adopting Fermyon's SpinKube platform, which utilizes WASM to efficiently manage workloads in Kubernetes. WASM's lightweight nature allows for just-in-time pod creation, eliminating the latency and cost associated with pre-provisioned resources. Furthermore, WASM's pre-compiled bytecode minimizes overhead on Ampere Altra processors, further enhancing efficiency.

ZEISS implemented three Spin apps (a distributor and two receivers) written in WASM, handling order processing with remarkable results. According to Kai Walter, Distinguished Architect at ZEISS, the SpinKube system processed the same number of orders as their previous Node.js system at 60% lower cost using Ampere processor VMs compared to x86 VMs.

The migration was seamless, running Spin apps alongside existing Kubernetes workloads without requiring a complete system overhaul. This iterative approach minimized disruption and allowed ZEISS to leverage its existing Arm node pools. The efficiency of WASM applications allowed ZEISS to utilize smaller, less expensive Azure instances (Standard DS2 and D2pds v5), resulting in substantial cost savings.

Key benefits included:

• Reduced Costs: Achieved a 60% reduction in costs compared to the previous x86-based system.
• Improved Efficiency: WASM's lightweight nature and Ampere Altra's optimized architecture minimized resource consumption.
• Simplified Deployment: SpinKube enabled a smooth, iterative migration without requiring major infrastructure changes.
• Scalability: The system efficiently handles fluctuating demand, scaling up and down as needed.

The case study highlights the potential of WASM and Ampere Altra processors to significantly reduce cloud costs for organizations with highly scalable, yet often idle, workloads. It showcases a practical, cost-effective solution to the overprovisioning conundrum commonly faced in cloud infrastructure management. For more details, refer to the linked video and resources.

The above is the detailed content of ZEISS Demonstrates the Power of Scalable Workflows with Ampere Altra and SpinKube. For more information, please follow other related articles on the PHP Chinese website!