Advances, Systems and Applications
Table 1 The dimensions of the cloud-native light-cone model
From: A multi-dimensional extensible cloud-native service stack for enterprises
Dimension | Example of scale | ||
|---|---|---|---|
| Â | Index | Object | Instance |
x: application | A1 | process (task) | a mapper task of a Spark MR job |
| Â | A2 | process group (job) | a Spark MR job |
| Â | A3 | application framework | a group of Spark runtime services |
| Â | A4 | application cluster | a Kubernetes cluster with Spark-operator |
y: infrastructure | I1 | device | a GPU |
| Â | I2 | node | a VM with a GPU |
| Â | I3 | IaaS stack instance | an OpenStack cluster |
| Â | I4 | IaaS source | a public cloud |
z: tenant | T1 | user | a personal (sub)account on a public cloud |
| Â | T2 | user group | a user group on a public cloud |
| Â | T3 | virtual organization | a project on a public cloud |
| Â | T4 | organization | an enterprise account on a public cloud |
t: workflow | W1 | intra-job | an Allreduce-based AI training |
| Â | W2 | inter-job | an ensemble-learning-based AI training |
| Â | W3 | inter-service | an AI pipeline from training to inference |
| Â | W4 | inter-cluster | an AI pipeline for CI/CD |