Driving Industry 4.0 at Distributed Edges with Cloud Orchestration
• Offline capabilities ( store and forward buffering , connection resiliency , etc .)
• Elasticity / Scalability
• Extensibility
• High availability and reliability
• Cloud orchestrated edge : lifecycle of distributed edge nodes based on its individual business configurations
• Automated , and continuous deployment , orchestration , management , replacement , recovery of edge nodes ( CI / CD )
APPROACH AND ARCHITECTURE FOR BUSINESS APPLICATIONS ON DISTRIBUTED EDGES WITH CLOUD ORCHESTRATION
Addressing the requirements discussed above , this section introduces an effective approach and architecture for edge computing solutions enabling mission critical execution of distributed manufacturing processes based on [ 14 ] [ 15 ]. One of the main elements of this approach is moving business applications as well as business data ( e . g ., master data , manufacturing orders ) closer to the manufacturing processes and combine them with OT data from multiple systems .
This approach aims at achieving faster decision-making closer where data is generated . Also , we elaborate on how an efficient edge computing architecture addresses the main limitations of cloud computing such as latency , limited bandwidth and intermittent connectivity that may result to costly disruptions within manufacturing processes . Finally , we introduce how continuous edge lifecycle management and centralized business configuration ensures low TCO as edge nodes and its workloads distributed in manufacturing processes are orchestrated from the cloud .
Edge-Cloud continuum for mission critical execution of distributed manufacturing processes . According to [ 12 ], the edge-cloud continuum is a physical infrastructure comprising of the internet , from discrete , decentralized devices to a centralized data center . Figure 2 shows an edgecloud continuum focusing on enabling mission critical execution of distributed manufacturing processes on the edge . It considers the characteristics from Table 1 and is based on the following layers .
IoT devices layer consists of any device with sensing capabilities , connected with a network , ingesting data and to be controlled / orchestrated from the overlying layers .
Edge layer consists of a set of distributed physical edge devices ( hardware such as gateways , hubs , PLCs , etc .), software components and the corresponding deployed business applications and data that forms an edge node . Physical edge devices are classified with different sizes ( S , M , L , XL ) depending on its computing and storage capabilities . Business applications are deployed on these devices and brings the necessary functionalities fulfilling a set mission critical and latency sensitive scenario . Example of applications are SAP Digital Manufacturing Cloud for edge computing [ 14 ] or AAS Service [ 15 ]. This layer is characterized ( see dimensions of Figure 2 ) by fast response times ( low latency ) and context awareness based on ingested data from assets that
IIC Journal of Innovation - 23 -