In unlocking the potential of IoT and its rich data streams of data for heavy industry, GE Digital has pioneered digital transformation across long-established sectors of the global economy. Its scalable, asset-centric data foundation provides a comprehensive and secure multi-tenant application platform that can run, scale, and extend digital industrial solutions. It leverages SaaS design principles to span edge technologies, analytics and machine learning, big data, and asset-centric digital twins. Real-time stream processing unifies edge sensors and centralized information systems in an integrated cloud fabric.
An innovative business model for monetizing IoT across industries and devices promised to open new pathways to data leverage. However, established data integration approaches, even using containers, quickly exposed gaps in manageability and scalability. Neither big data warehousing nor microservices enabled sufficient agility to meet the rapidly evolving landscape of signals, sensors, and analytics across the company’s diverse customer base. Application deployment hit DevOps hard, as the number of moving parts and the dynamic activities accelerated. Moreover, keeping up with the most innovative of its customers risked GE needing to manage an expanding network of tenant islands.
- Software manageability and agility became the prime directive. To make the transition to a more future-proof technology strategy, the organization turned to the experts at the CloudGeometry. We worked together to design and deploy a SaaS architecture that could reach the next level of resilience and robustness. Kubernetes as service in Amazon (EKS) was used to deploy, scale, and operate all system environments. It’s a virtuous combination of containerization and work orchestration.
- Ingestion of IIoT event payloads from both physical and digital devices are piped through Kafka streaming logic, accommodating different customers with different operational needs. This approach helps to quickly deploy the full-scale system for any new tenant; Docker, Jenkins, and Ansible drive a CI/CD toolchain. It also simplifies resizing different instances to fit workload needs dynamically.
- The company’s platform is designed to offer a mix and match of IIoT workloads. One customer/tenant running on the platform in turn hosts multiple child entities as end-users and collections of IoT Device networks. This sort of multi-tier federation is also a key benefit of Kubernetes design principles. A separate ELK cluster is deployed per each tenant; it supports fine-grained billing, monitoring, and analytics.
Migrating 12 microservices from EC2 to AWS EKS translated a SaaS business model into an operationally sustainable multi-tenancy IIoT workload portfolio offering.
<div class="case__txt--cols"><div><h4>Secure multi-tier multi-tenancy</h4><p>Kubernetes runs multiple isolated applications on elastic infrastructure, leveraging namespaces and IAM security roles / policies.</p></div><div><h4>IIoT workload deployment automation</h4><p>Automation of workload launch, scaling, rollbacks, etc. makes on-boarding new industrial customers a matter only a few human hours.</p></div><div><h4>Full-stack fine-grained billing</h4><p>Customized CostToServe module calculates compute and storage resource utilization per tenant, per device and per user.</p></div></div>