Home > Storage > PowerScale (Isilon) > Industry Solutions and Verticals > Analytics > Dell Streaming Data Platform: Architecture, Configuration, and Considerations > SDP Edge
The Edge is a small computer cluster that aggregates data from multiple sensors in the field. In a highly available configuration, three nodes form a Kubernetes cluster, and a distributed file system is used to store Pravega streams. A single server can be used if high availability is not required.
SDP Edge is running on a Kubernetes cluster deployed with Kubespray. Kubespray is an open-source project developed to facilitate Kubernetes deployment on Cloud platforms or Bare-metal servers. In this solution, we are using Kubespray to deploy Kubernetes cluster on Bare-metal servers at the Edge. By using Kubespray, customers can deploy SDP at the edge with a small footprint that requires fewer physical resources. Kubespray also offers the possibility to deploy Kubernetes on a single node for edge sites that do not require High Availability (HA) or on three nodes for edge sites that require HA. Ubuntu 18.04.x and Red Hat Enterprise Linux 8.4.x are supported.
From a storage point of view, there are two deployment options based on the number of nodes within the Kubernetes cluster:
From a processing point of view, there are a few options within SDP: Apache Flink, Spark, Pravega Search, and Video Analytics.
For the networking configuration, Kubespray is using calico. There is nothing to do from a user point of view as the installer package will take care of everything. The important thing to know is that for each edge deployment, end users will have to provide a range of available and unused IPs. See Table 2 for details.
Deployment option | Number of IPs | Consumers |
Single node | 3 floating IPs, within the same subnet as the node | 1 IP for service pod 1 IP for ingress service 1 IP for Segment Store |
3 nodes | 5 floating IPs, within the same subnet as the nodes | 1 IP for service pod 1 IP for ingress service 3 IPs for Segment Store |