2 Deployment Environment
This section provides information about the cloud native platform used for SCP benchmarking.
2.1 Deployed Components
Deployment Platform
Oracle Communications Cloud Native Core, Cloud Native Environment (CNE) 24.1.0 and CNE on Bare Metal 24.1.0 is used for performing benchmark tests.
Observability Services
The following table lists services that are part of CNE and used for fetching SCP metrics.
Table 2-1 Observability Services
| Service Name | Version |
|---|---|
| Oracle OpenSearch | 2.3.0 |
| Oracle OpenSearch Dashboard | 2.3.0 |
| Fluentd | 1.16.2 |
| Prometheus | 2.51.1 |
| Grafana | 9.5.3 |
| Jaeger | 1.52.0 |
Cloud Native Orchestrator
Kubernetes 1.28.6 is used to manage application pods across the cluster.
cnDBTier
cnDBTier 25.1.100 is used to perform benchmark tests.
For more information about above mentioned software, see Oracle Communications Cloud Native Core, cnDBTier Installation, Upgrade, and Fault Recovery Guide.
2.2 Deployment Resources
The performance and capacity of SCP can vary based on the chosen environment and how SCP is deployed. This section provides information about CNE and cnDBTier resources used to perform benchmark tests.
2.2.1 Cluster Details
The following table provides information about the types of servers and the number of servers used in the test environment:
Table 2-2 Test Bed 1 - CNE on Bare Metal
| Nodes | Type | Count |
|---|---|---|
| Primary Nodes | HP Gen10 RMS | 3 |
| Worker Nodes | HP Gen10 Blades | 29 |
| HP Gen8 Blades | 7 | |
| Top of Rack Switch | Cisco Nexus9000 93180YC-EX | 2 |
| Enclosure Switch | HP 6120 | 2 |
The following table provides information about the number of pods required by each CNE service.
Table 2-3 CNE Common Services Observability Resources
| Service Name | Number of Pods | RAM Request/Limit | vCPU Request/Limit | PVC Size Recommendation |
|---|---|---|---|---|
| Prometheus Server | 2 | 50Gi/50Gi | 12/12 | 150GB to 200GB |
| Prometheus-pushgateway | 1 | 32Mi/32Mi | 10m/10m | NA |
| Alert Manager | 2 | 164Mi/164Mi | 40m/40m | NA |
| Fluentd | 1 per Worker Node | 200Mi/500Mi | 100m/100m | NA |
| Prom-node-exporter | 1 per Worker Node | 512Mi/512Mi | 800m/800m | NA |
| MetalLB speaker | 1 per Worker Node | 100Mi/100Mi | 100m/100m | NA |
| OpenSearch Data | 3/3 | 32Gi/32Gi (JVM 16) | 2/2 | 300GB |
| OpenSearch Master | 3/3 | 16Gi/16Gi(JVM 8) | 1/1 | 300GB |
| ISM Policy | 3/3 | 128Mi/128Mi | 100m/100m | NA |
| OpenSearch Client | 1 | 128Mi/128Mi | 100m/100m | NA |
| Grafana | 1 | 500Mi/500Mi | 500m/500m | NA |
| Kibana | 1 | 500Mi/1Gi | 100m/1 | NA |
| kube-state-metrics | 1 | 200Mi/200Mi | 50m/50m | NA |
| jaeger-agent | 1 per Worker Node | 128Mi/512Mi | 256m/500m | NA |
| jaeger-collector | 1 | 512Mi/1Gi | 500m/1250m | NA |
| jaeger-query | 1 | 128Mi/512Mi | 256m/500m | NA |
| rook-ceph-osd | 1 for each raw disk available to OS on all Worker Node | 1Gi/8Gi | 500m/1 | NA |
| rook-ceph-mgr | 1 | 1Gi/1Gi | 500m/500m | NA |
| rook-ceph-mon | 3 | 1Gi/1Gi | 500m/500m | NA |
| rook-ceph-operator | 1 | 2Gi/2Gi | 100m/500m | NA |
Table 2-4 Test Bed 2 - VMware Tanzu
| Nodes | Type | Count |
|---|---|---|
| Primary Nodes | VM (8 CPU and 64 GB Memory) | 3 |
| Worker Nodes | VM(32 CPU and 128 GB Memory) | 51 |
| Underlying Hardware | Cisco Nexus9000 93180YC-EX | 19 |
Table 2-5 Test Bed 3 - CNE on Bare Metal
| Nodes | Type | Count |
|---|---|---|
| Primary Nodes | X9 Server and NVME | 3 |
| Worker Nodes | X9 Server and NVME | 17 |
Table 2-6 Test Bed 4 - CNE on Bare Metal
| Nodes | Type | Count |
|---|---|---|
| Primary Nodes | ORACLE SERVER X8-2 | 3 |
| Worker Nodes | ORACLE SERVER X8-2 | 45 |
| Top of Rack Switch | Cisco 93108tc-ex | 2 |
The following table provides information about the number of pods required by each CNE service.
Table 2-7 CNE Common Services Observability Resources
| Service Name | Number of Pods | RAM Request/Limit | vCPU Request/Limit | PVC Size Recommendation |
|---|---|---|---|---|
| Prometheus Server | 2 | 50Gi/50Gi | 16/16 | 150GB to 800GB |
| Prometheus-pushgateway | 1 | 2Gi/3Gi | 2/4 | NA |
| Alert Manager | 2 | 164Mi/164Mi | 40m/40m | NA |
| Fluentd | 1 per Worker Node | 200Mi/500Mi | 100m/100m | NA |
| Prom-node-exporter | 1 per Worker Node | 512Mi/512Mi | 800m/800m | NA |
| MetalLB speaker | 1 per Worker Node | 100Mi/100Mi | 100m/100m | NA |
| OpenSearch Data | 3/3 | 164Gi/100Mi | 1/8 | 300GB |
| OpenSearch Master | 3/3 | 16Gi/16Gi(JVM 8) | 1/1 | 300GB |
| ISM Policy | 3/3 | 128Mi/128Mi | 100m/100m | NA |
| OpenSearch Client | 1 | 128Mi/128Mi | 100m/100m | NA |
| Grafana | 1 | 500Mi/500Mi | 500m/500m | NA |
| Kibana | 1 | 500Mi/1Gi | 100m/1 | NA |
| kube-state-metrics | 1 | 200Mi/200Mi | 50m/50m | NA |
| jaeger-agent | 1 per Worker Node | 128Mi/512Mi | 256m/500m | NA |
| jaeger-collector | 1 | 512Mi/1Gi | 500m/1250m | NA |
| jaeger-query | 1 | 128Mi/512Mi | 256m/500m | NA |
| rook-ceph-osd | 1 for each raw disk available to OS on all Worker Node | 1Gi/8Gi | 500m/1 | NA |
| rook-ceph-mgr | 1 | 1Gi/1Gi | 500m/500m | NA |
| rook-ceph-mon | 3 | 1Gi/1Gi | 500m/500m | NA |
| rook-ceph-operator | 1 | 2Gi/2Gi | 100m/500m | NA |
2.2.2 cnDBTier Resources
The following table describes resources required by cnDBTier pods to perform SCP benchmark tests.
Table 2-8 Test Bed 1 - cnDBTier Resources
| cnDBTier Pods | Replica | vCPU | RAM (GB) | ASM Sidecar | cnDBTier Sidecar | Storage | Ephemeral Storage | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Request | Limit | Request | Limit | vCPU | RAM (GB) | RAM (GB) | RAM(GB) | PVC(GB) | Count | Req(M) | Limit(G) | ||
| SQL - Replication (ndbmysqld) StatefulSet | 2 | 2 | 3 | 2 | 4 | 1 | 1 | 1 | 1 | 30 | 1 | 90 | 1 |
| MGMT (ndbmgmd) StatefulSet | 3 | 2 | 3 | 2 | 4 | 1 | 1 | NA | NA | 30 | 1 | 90 | 1 |
| DB (ndbmtd) StatefulSet | 4 | 3 | 4 | 4 | 4 | 1 | 1 | 1 | 1 | 30 | 2 | 90 | 1 |
| db-backup-manager-svc | 1 | 1 | 1 | 1 | 1 | 1 | 1 | NA | NA | NA | NA | 90 | 1 |
| db-replication-svc | 1 | 1 | 2 | 1 | 2 | 1 | 1 | NA | NA | NA | NA | 90 | 1 |
| db-monitor-svc | 1 | 1 | 2 | 1 | 2 | 1 | 1 | NA | NA | NA | NA | 90 | 1 |
| db-connitivity-service | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | NA | NA | NA | NA |
| SQL - Access (ndbappmysqld) StatefulSet | 2 | 3 | 4 | 4 | 4 | 1 | 1 | NA | NA | 20 | 2 | 90 | 1 |
Table 2-9 Test Bed 2 - cnDBTier Resources
| cnDBTier Pods | Replica | vCPU | RAM (GB) | ||
|---|---|---|---|---|---|
| Request | Limit | Request | Limit | ||
| SQL - Replication(ndbmysqld) | 2 | 4 | 4 | 10 | 10 |
| MGMT (ndbmgmd) StatefulSet | 3 | 2 | 2 | 7 | 7 |
| DB (ndbmtd) StatefulSet | 4 | 3 | 3 | 7 | 7 |
| SQL - Access (ndbappmysqld) | 2 | 4 | 4 | 8 | 8 |
| db-backup-manager-svc | 1 | 0.1 | 0.1 | 0.128 | 0.128 |
| db-monitor-svc | 1 | 0.2 | 0.2 | 0.5 | 0.5 |
2.2.3 SCP Resources
The following table provides information about resource requirements to perform SCP benchmark tests:
Table 2-10 SCP Resources
| Microservice Name | SCP Service Pods | |||
|---|---|---|---|---|
| vCPU/Pod | Memory/Pod (in Gi) | |||
| Min | Max | Min | Max | |
| Helm test | 3 | 3 | 3 | 3 |
| Helm Hook | 3 | 3 | 3 | 3 |
| scpc-subscription | 1 | 1 | 1 | 1 |
| scpc-notification | 4 | 4 | 4 | 4 |
| scpc-audit | 3 | 3 | 4 | 4 |
| scpc-configuration | 2 | 2 | 2 | 2 |
| scp-cache | 8 | 8 | 8 | 8 |
| scp-loadmanager | 8 | 8 | 8 | 8 |
| scp-nrfproxy | 8 | 8 | 8 | 8 |
| scp-worker (Profile 1) | 4 | 4 | 8 | 8 |
| scp-worker (Profile 2) | 8 | 8 | 12 | 12 |
| scp-worker (Profile 3) | 12 | 12 | 16 | 16 |
| scp-mediation | 8 | 8 | 8 | 8 |
| scp-nrfproxy-oauth | 8 | 8 | 8 | 8 |
| scpc-alternate-resolution | 2 | 2 | 2 | 2 |