2 Deployment Environment
This section provides information about the cloud native infrastructure used for SEPP benchmarking.
The following diagram represents the deployment configurations of SEPP:
Figure 2-1 Deployment Diagram
The details about the deployment configurations are as follows:
- cnDBTier, cSEPP, pSEPP are deployed in same cluster
- perfgo-client and perfgo-server are deployed in another cluster
- DNS server is running on a standalone VM in a different cluster
- cSEPP to pSEPP n32 TLS connection is on External IP (metalLB)
- perfgo-client is sending traffic on SEPP external IP
- SEPP to perfgo-server connection is using External IP (metalLB)
- Both SEPP have 30 local PLMNs
- Network delay of 50 ms on server side
2.1 Overall Summary of Benchmarking Configurations
The following table lists the overall summary of benchmarking configurations.
Overall Summary
| S. No. | Execution Time | SEPP Release | CNCC Release | CNDBTier Release | ASM Enabled /Disabled | OSO Installed | HA Proxy Installed | Cluster | Site-1 MPS | Site-2 MPS | Overall MPS | Network Delay on server side | Features Enabled | Status | Remarks |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Run1 | 24hr | csepp-23.3.1 & psepp-23.3.1 | 23.3.1 | 23.3.0 | Disabled | Yes | NA | groove2 | 22.4K | 22.4K | 44.8K | NA | NA | PASS | NA |
2.2 Deployed Components
Deployment Platform
Oracle Communications Cloud Native Environment (OCCNE) 22.3.0 and BareMetal are used for performing benchmark tests.
Observability Services
The following table lists observability services and their versions that are used for SEPP benchmark tests. These services fetch SEPP metrics, alerts, logs, and traces.
Table 2-1 Observability Services
| Service | Version |
|---|---|
| Elasticsearch | 7.9.3 |
| Fluentd | 3.1.0 |
| Kibana | 7.9.3 |
| Prometheus | 1.2.2 |
| Grafana | 7.5.11 |
| Jaeger | 1.22.0 |
Cloud Native Orchestrator
Kubernetes 1.20.x is used for managing application pods across the cluster.
cnDBTier
cnDBTier 22.4.0 is used for performing benchmark tests.
For more information about above mentioned components, see Oracle Communications Cloud Native Core, Security Edge Protection Proxy Installation, Upgrade, and Fault Recovery Guide.
2.3 Deployment Resources
The performance and capacity of SEPP can vary based on the chosen environment and how SEPP is deployed. This section provides information about CNE resources used to perform benchmark tests.
2.3.1 Resource Requirements for CNE Observability Services
The following table provides information about the number of pods required by each CNE service.
Table 2-2 Resource Requirements for CNE Observability Services
| Service Name | Number of Pods |
|---|---|
| Prometheus Server | 1 |
| Prometheus-pushgateway | 1 |
| Alert Manager | 2 |
| Fluentd | 1 per worker node |
| Prom-node-exporter | 1 per worker node |
| MetalLB speaker | 1 per worker node |
| ES Data/Master | 3/3 |
| ES Curator | 1 |
| ES-exporter | 1 |
| Grafana | 1 |
| Kibana | 1 |
| kube-state-metrics | 1 |
| jaeger-agent | 1 per worker node |
| jaeger-collector | 1 |
| jaeger-query | 1 |
| rook-ceph-osd | 1 for each raw disk available to OS on all worker nodes |
| rook-ceph-mgr | 1 |
| rook-ceph-mon | 3 |
| rook-ceph-osd | 1 |