4 OCNADD Benchmarking Testing

This section describes the performance testing scenarios and results for the Message Feed functionality provided by Oracle Communications Network Analytics Data Director. The message feed feature is tested with SCP, NRF, and SEPP as the source of the message feed.

• unresolvable-reference.html#GUID-A975082E-D02C-4523-9275-67D327B26667
• Performance Benchmarking for 270K MPS SCP Traffic with Synthetic Feed
• Performance Benchmarking for 9K BSF Traffic with HTTP2 Feed

4.1 Performance Benchmarking for 135K MPS Traffic with Synthetic Feed Replication

This performance benchmarking evaluates a centralized deployment with a single worker group, handling an aggregated feed of 135K MPS. The breakdown includes 90K MPS for SCP (45K TPS with 2 trigger points), 30K MPS for SEPP (15K TPS with 2 trigger points), and 15K MPS for NRF (7.5K TPS with 2 trigger points). The benchmark utilizes a replicated synthetic feed:

The benchmarking results are provided below:

Note:

One Ingress message from a NF is "1" MPS for OCNADD.

Benchmark Test Environment

• Ingress Traffic Rate: 135K MPS
• Observation Period: 3+ hours
• Deployment: OCNADD single-site, ASM disabled
• Kafka Replication Factor: 1
• Kafka PVC: 400GB
• Data Feed: Two TCP (Synthetic) Feeds (TLS enabled)
• Third-Party Applications: 1 endpoint per feed
• Software Versions:
  • cnDBTier: 24.3.0
  • SCP: 24.3.0
  • NRF: 24.3.0
  • SEPP: 24.3.0
  • CNCC: 24.3.0
• Security: SASL/SSL enabled between NRF, SCP, SEPP, and OCNADD
• Message Size: ~3500 Bytes
• Execution Time: 24 hours
• Call Mix: SCP, NRF, and SEPP NF CALL MIX models

Setup Details

• Environment: OCCNE (Bare metal) 23.3.4
• OCNADD: 25.1.100

Configuration

• NRF: Ingress Gateway and Egress Gateway messageCopy enabled
• SCP: Default configuration
• cSEPP: PLMN Ingress Gateway messageCopy enabled
• pSEPP: messageCopy disabled
• OCNADD:
  • Replication Factor: 1
  • Kafka PVC: 400GB. For more information about PVC requirements, see Kafka PVC Storage Requirements.

Benchmark Testcase Specifications

The testcase parameters are as follows:

• OCNADD
  • A feed is configured using the OCNADD Console with aggregation rules set for SCP, NRF and SEPP.
  • Message Ingestion Rate: 135K MPS
  Resource Specifications:

  Table 4-1 OCNADD Resource Specifications

  Services	CPU Request Per Pod	CPU Limit Per Pod	Memory Request Per Pod (Gi)	Memory Limit Per Pod (Gi)	min Replicas	max Replicas	Paritions/Retention	Topics
  ocnaddadapter	3	3	6	6	11	11	66	MAIN
  ocnaddadminservice	1	1	1	1	1	1	-	-
  ocnaddalarm	1	1	1	1	1	1	-	-
  ocnaddconfiguration	1	1	1	1	1	1	-	-
  ocnaddexport	2	4	4	64	1	2	-	-
  ocnaddgui	1	2	1	1	1	2	-	-
  ocnaddhealthmonitoring	1	1	1	1	1	1	-	-
  ocnaddkafka	7.5	7.5	96	96	6	6	-	-
  ocnaddnrfaggregation	2	2	2	2	1	1	6	NRF
  ocnaddscpaggregation	2	2	4	4	5	5	30	SCP
  ocnaddseppaggregation	2	2	4	4	2	2	12	SEPP
  ocnadduirouter	1	2	1	1	1	2	-	-
  zookeeper	1	1	1	2	3	3	-	-

• SCP: 90K MPS traffic (45K TPS with 2 trigger points)
• NRF: 15K MPS traffic (7.5K TPS with 2 trigger points)
• SEPP 30K MPS traffic (15K TPS with 2 trigger points)

Benchmark Test Results

Traffic Feed Details

Table 4-2 Traffic Feed Details

NF	NF Traffic Copy Trigger Points	Traffic Rate	Duration in hours	E2E Traffic Feed Average Latency	E2E Traffic Feed Success Rate
OCNADD	NA	135K MPS	24	Feed-1: 26.9 msec Feed-2: 26.5 msec	Feed-1: 99.98% Feed-2: 99.98%
SCP	Ingress Gateway (Request) Egress Gateway (Request)	45K TPS	24	NA	99.98%
NRF	Ingress Gateway (Request) Egress Gateway (Request)	7.5K TPS	24	NA	99.98%
SEPP	Ingress Gateway (Request) Egress Gateway (Request)	15K TPS	24	NA	99.98%

CPU and Memory Utilization

The following table describes the OCNADD CPU and memory utilization:

Table 4-3 OCNADD CPU and Memory Utilization

Micro-Service/ Container	Pod Count	CPU Utilization (%)	Memory Utilization (%)
adapter-1	11	60.6	19.0
adapter-2	11	63.0	19.4
kafka-broker	6	40.1	25.8
ocnaddadminservice	1	0.138	24.1
ocnaddalarm	1	0.217	27.5
ocnaddconfiguration	1	0.164	34.9
ocnaddexport	1	0.0115	2.33
ocnaddgui	1	0.00108	3.12
ocnaddhealthmonitoring	1	0.530	34.5
ocnaddnrfaggregation	1	70.4	32.4
ocnaddscpaggregation	5	58.0	18.6
ocnaddseppaggregation	2	70.0	20.3
ocnadduirouter	1	0.00758	28.7
zookeeper	3	0.125	31.0

4.2 Performance Benchmarking for 270K MPS SCP Traffic with Synthetic Feed

This performance benchmarking evaluates a centralized deployment with a single worker group, handling an SCP Model-C traffic of 270K MPS. The benchmark utilizes a synthetic feed:

The benchmarking results are provided below:

Note:

One Ingress message from a NF is "1" MPS for OCNADD.

Benchmark Test Environment

• Ingress Traffic Rate: 270K MPS
• Observation Period: 3+ hours
• Deployment: OCNADD single-site, ASM disabled
• Kafka Replication Factor: 1
• Kafka PVC: 300GB
• Data Feed: TCP (Synthetic)
• Third-Party Applications: 1 endpoint per feed
• Software Versions:
  • cnDBTier: 24.3.0
  • NRF: 24.3.0
  • CNCC: 24.3.0
• Security: SASL/SSL enabled between between SCP and OCNADD
• Message Size: ~3500 - 4000 Bytes
• Execution Time: 14 hours
• Call Mix: SCP Model-C

Setup Details

• Environment: OCCNE (Bare metal) 23.3.4
• OCNADD: 25.1.100

Configuration

• SCP: Default configuration, messageCopy enabled
• OCNADD:
  • Replication Factor: 1
  • Kafka PVC: 300GB. For more information about PVC requirements, see Kafka PVC Storage Requirements.

Benchmark Testcase Specifications

The testcase parameters are as follows:

• OCNADD
  • A feed is configured using the OCNADD Console with SCP.
  • Message Ingestion Rate: 270K MPS
  Resource Specifications:

  Table 4-4 OCNADD Resource Specifications

  Services	CPU Request Per Pod	CPU Limit Per Pod	Memory Request Per Pod (Gi)	Memory Limit Per Pod (Gi)	min Replicas	max Replicas	Paritions/Retention	Topics
  ocnaddadapter(TCPs)	3	3	6	6	21	21	126	MAIN
  ocnaddadminservice	1	1	1	1	1	1	-	-
  ocnaddalarm	1	1	1	1	1	1	-	-
  ocnaddconfiguration	1	1	1	1	1	1	-	-
  ocnaddexport	2	4	4	64	1	2	-	-
  ocnaddgui	1	2	1	1	1	2	-	-
  ocnaddhealthmonitoring	1	1	1	1	1	1	-	-
  ocnaddkafka	5	5	48	48	11	11	-	-
  ocnaddscpaggregation	2	2	4	4	11	11	66	SCP
  ocnadduirouter	1	2	1	1	1	2	-	-
  zookeeper	1	1	1	2	3	3	-	-

• SCP: SCP Traffic is 135K TPS using two trigger points.

Benchmark Test Results

Traffic Feed Details

Table 4-5 Traffic Feed Details

NF	NF Traffic Copy Trigger Points	Traffic Rate	Duration in hours	E2E Traffic Feed Average Latency	E2E Traffic Feed Success Rate
OCNADD	NA	270K MPS	14	51.1 msec	99.99%
SCP	Ingress Gateway (Request) Egress Gateway (Request)	135K MPS	14	NA	99.98%

CPU and Memory Utilization

The following table describes the OCNADD CPU and memory utilization:

Table 4-6 OCNADD CPU and Memory Utilization

Micro-Service/ Container	Pod Count	CPU Utilization (%)	Memory Utilization (%)
ocnaddscpaggregation	11	65.5	29.8
ocnaddadapter(TCPs)	21	58.1	29.0
kafka-broker	11	44.1	60.5
zookeeper	3	0.152	33.7
ocnaddadminservice	1	0.134	27.9
ocnaddalarm	1	0.399	18.4
ocnaddconfiguration	1	0.232	33.9
ocnaddexport	1	0.0183	2.21
ocnaddgui	1	0.00110	3.12
ocnaddhealthmonitoring	1	0.710	33.2
ocnadduirouter	1	0.00812	25.3

4.3 Performance Benchmarking for 270K MPS SCP Traffic with Kafka Feed

This performance benchmarking evaluates a centralized deployment with a single worker group, handling an SCP traffic of 270K MPS. The benchmark utilizes a Kafka feed:

The benchmarking results are provided below:

Note:

One Ingress message from a NF is "1" MPS for OCNADD.

Benchmark Test Environment

• Ingress Traffic Rate: 270K MPS
• Observation Period: 3+ hours
• Deployment: OCNADD single-site, ASM disabled
• Kafka Replication Factor: 1
• Kafka PVC: 300GB
• Data Feed: Kafka-feed
• Third-Party Applications: 1 endpoint per feed
• Software Versions:
  • cnDBTier: 24.3.0
  • SCP: 24.3.0
  • CNCC: 24.3.0
• Security: SASL/SSL enabled between between SCP and OCNADD
• Message Size: ~3500 - 4000 Bytes
• Execution Time: 12 hours
• Call Mix: SCP Model-C

Setup Details

• Environment: OCCNE (Bare metal) 23.3.4
• OCNADD: 25.1.100

Configuration

• SCP: Default configuration, messageCopy enabled
• OCNADD:
  • Replication Factor: 1
  • Kafka PVC: 300GB. For more information about PVC requirements, see Kafka PVC Storage Requirements.

Benchmark Testcase Specifications

The testcase parameters are as follows:

• OCNADD
  • A feed is configured using the OCNADD Console with SCP.
  • Message Ingestion Rate: 270 K MPS
  Resource Specifications:

  Services	CPU Request Per Pod	CPU Limit Per Pod	Memory Request Per Pod (Gi)	Memory Limit Per Pod (Gi)	min Replicas	max Replicas	Paritions/Retention	Topics
  ocnaddadapter	-	-	-	-	-	-	261	MAIN
  ocnaddadminservice	1	1	1	1	1	1	-	-
  ocnaddalarm	1	1	1	1	1	1	-	-
  ocnaddconfiguration	1	1	1	1	1	1	-	-
  ocnaddexport	2	4	4	64	1	2	-	-
  ocnaddgui	1	2	1	1	1	2	-	-
  ocnaddhealthmonitoring	1	1	1	1	1	1	-	-
  ocnaddkafka	5	5	48	48	11	11	-	-
  ocnaddscpaggregation	2	2	4	4	11	11	66	SCP
  ocnadduirouter	1	2	1	1	1	2	-	-
  zookeeper	1	1	1	2	3	3	-	-

• SCP: SCP Traffic is 135K TPS using two trigger points.

Benchmark Test Results

Traffic Feed Details

Table 4-7 Traffic Feed Details

NF	NF Traffic Copy Trigger Points	Traffic Rate	Duration in hours	E2E Traffic Feed Average Latency	E2E Traffic Feed Success Rate
OCNADD	NA	270K MPS	12	Feed-1: 26.9 msec Feed-2: 26.5 msec	Feed-1: 99.98% Feed-2: 99.98%
SCP	Ingress Gateway (Request) Egress Gateway (Request)	135K TPS	12	NA	99.98%

CPU and Memory Utilization

The following table describes the OCNADD CPU and memory utilization:

Table 4-8 OCNADD CPU and Memory Utilization

Micro-Service/ Container	Pod Count	CPU Utilization (%)	Memory Utilization (%)
kafka-broker	11	46.1	16.3
ocnaddadminservice	1	0.0855	25.9
ocnaddalarm	1	0.80	26.8
ocnaddconfiguration	1	0.139	33.2
ocnaddexport	1	0.0115	2.18
ocnaddgui	1	0.000999	3.10
ocnaddhealthmonitoring	1	0.339	35.0
ocnaddscpaggregation	11	68.2	27.9
ocnadduirouter	1	0.00654	54.0
zookeeper	3	0.121	28.8

4.4 Performance Benchmarking for 9K BSF Traffic with HTTP2 Feed

The performance benchmarking test is performed on OCNADD with 9K traffic with HTTPS feed. The traffic includes:

• BSF Traffic: 9K MPS
• Enabled BSF Ingress Gateway and Egress Gateway Message Copy Feature.

The benchmarking results are provided below:

Note:

One Ingress message from a NF is "1" MPS for OCNADD.

Benchmark Test Environment

• Ingress Traffic Rate: 9K MPS
• Observation Period: 3+ hours
• Deployment: OCNADD single-site, ASM disabled
• Kafka Replication Factor: 1
• Kafka PVC: 30GB
• Data Feed: HTTP2
• Third-Party Applications: 1 endpoint per feed
• Software Versions:
  • cnDBTier: 24.3.0
  • BSF: 24.3.0
  • CNCC: 24.3.0
• Security: SASL/SSL enabled between between BSF and OCNADD
• Message Size: ~1200 Bytes
• Execution Time: 12 hours
• Call Mix: Binding-Create/Delete

Setup Details

• Environment: OCCNE (Bare metal) 23.3.4
• OCNADD: 25.1.100

Configuration

• BSF: Ingress Gateway and Egress Gateway messageCopy enabled
• OCNADD:
  • Replication Factor: 1
  • Kafka PVC: 30GB. For more information about PVC requirements, see Kafka PVC Storage Requirements.

Benchmark Testcase Specifications

The testcase parameters are as follows:

• OCNADD
  • A feed is configured using the OCNADD Console with BSF.
  • Message Ingestion Rate: 9K MPS
  Resource Specifications:

  Table 4-9 OCNADD Resource Specifications

  Services	CPU Request Per Pod	CPU Limit Per Pod	Memory Request Per Pod (Gi)	Memory Limit Per Pod (Gi)	min Replicas	max Replicas	Paritions/Retention	Topics
  ocnaddadapter	3	3	4	4	1	1	9	MAIN
  ocnaddadminservice	1	1	1	1	1	1	-	-
  ocnaddalarm	1	1	1	1	1	1	-	-
  ocnaddbsfaggregation	3	3	4	4	1	1	6	BSF
  ocnaddconfiguration	1	1	1	1	1	1	-	-
  ocnaddexport	2	4	4	64	1	2	-	-
  ocnaddfilter	2	2	3	3	1	1	-	-
  ocnaddgui	1	2	1	1	1	2	-	-
  ocnaddhealthmonitoring	1	1	1	1	1	1	-	-
  ocnaddkafka	2	2	24	24	4	4	-	-
  ocnadduirouter	1	2	1	1	1	2	-	-
  zookeeper	1	1	1	2	3	3	-	-

• BSF : BSF Traffic: 4.5K TPS using two trigger points.

Benchmark Test Results

Traffic Feed Details

Table 4-10 Traffic Feed Details

NF	NF Traffic Copy Trigger Points	Traffic Rate	Duration in hours	E2E Traffic Feed Average Latency	E2E Traffic Feed Success Rate
OCNADD	NA	9K MPS	12	20.20ms	Feed: 100%
BSF	Ingress Gateway (Request) Egress Gateway (Request)	4.5K TPS	12	NA	100%

CPU and Memory Utilization

The following table describes the OCNADD CPU and memory utilization:

Table 4-11 OCNADD CPU and Memory Utilization

Micro-Service/ Container	Pod Count	CPU Utilization (%)	Memory Utilization (%)
adapter	1	33.5	32.1
kafka-broker	4	12.3	11.4
ocnaddadminservice	1	0.287	27.5
ocnaddalarm	1	0.0963	37.7
ocnaddbsfaggregation	1	31.0	22.1
ocnaddconfiguration	1	0.110	36.7
ocnaddexport	1	0.00686	2.32
ocnaddgui	1	0.00107	3.13
ocnaddhealthmonitoring	1	0.0898	35.1
ocnadduirouter	1	0.00664	29.5
zookeeper	3	0.122	27.0

4.5 Performance Benchmarking for 30K MPS PCF Traffic with HTTP2 Feed

The performance benchmarking test is performed on OCNADD with 31K MPS traffic traffic with HTTP2 feed. The traffic includes:

• PCF Traffic: 30K MPS
• Enabled PCF Ingress Gateway and Egress Gateway Message Copy Feature.

The benchmarking results are provided below:

Note:

One Ingress message from a NF is "1" MPS for OCNADD.

Benchmark Test Environment

• Ingress Traffic Rate: 30K MPS
• Observation Period: 3+ hours
• Deployment: OCNADD single-site, ASM disabled
• Kafka Replication Factor: 1
• Kafka PVC: 50GB
• Data Feed: HTTP2
• Third-Party Applications: 1 endpoint per feed
• Software Versions:
  • cnDBTier: 24.3.0
  • PCF: 24.3.0
  • CNCC: 24.3.0
• Security: SASL/SSL enabled between between PCF and OCNADD
• Message Size: ~2079 Bytes
• Execution Time: 12 hours
• Call Mix: Session Management (SM) traffic

Setup Details

• Environment: OCCNE (Bare metal) 23.3.4
• OCNADD: 25.1.100

Configuration

• PCF: Ingress Gateway and Egress Gateway messageCopy enabled
• OCNADD:
  • Replication Factor: 1
  • Kafka PVC: 50GB. For more information about PVC requirements, see Kafka PVC Storage Requirements.

Benchmark Testcase Specifications

The testcase parameters are as follows:

• OCNADD
  • A feed is configured using the OCNADD Console with PCF.
  • Message Ingestion Rate: 30K MPS
  Resource Specifications:

  Table 4-12 OCNADD Resource Specifications

  Services	CPU Request Per Pod	CPU Limit Per Pod	Memory Request Per Pod (Gi)	Memory Limit Per Pod (Gi)	min Replicas	max Replicas	Paritions/Retention	Topics
  ocnaddadapter	3	3	4	4	3	3	27	MAIN
  ocnaddadminservice	1	1	1	1	1	1	-	-
  ocnaddalarm	1	1	1	1	1	1	-	-
  ocnaddpcfaggregation	2	2	2	2	2	2	12	PCF
  ocnaddconfiguration	1	1	1	1	1	1	-	-
  ocnaddexport	2	4	4	64	1	2	-	-
  ocnaddfilter	2	2	3	3	1	1	-	-
  ocnaddgui	1	2	1	1	1	2	-	-
  ocnaddhealthmonitoring	1	1	1	1	1	1	-	-
  ocnaddkafka	2	2	24	24	4	4	-	-
  ocnadduirouter	1	2	1	1	1	2	-	-
  zookeeper	1	1	1	2	3	3	-	-

• PCF : PCF Traffic: 15K TPS using two trigger point.

Benchmark Test Results

Traffic Feed Details

Table 4-13 Traffic Feed Details

NF	NF Traffic Copy Trigger Points	Traffic Rate	Duration in hours	E2E Traffic Feed Average Latency	E2E Traffic Feed Success Rate
OCNADD	NA	30K MPS	12	33.6ms	Feed: 99.99%
PCF	Ingress Gateway (Request) Egress Gateway (Request)	15K TPS	12	NA	99.99%

CPU and Memory Utilization

The following table describes the OCNADD CPU and memory utilization:

Table 4-14 OCNADD CPU and Memory Utilization

Micro-Service/ Container	Pod Count	CPU Utilization (%)	Memory Utilization (%)
httpsfeed-adapter	3	51.6	62.10
kafka-broker	4	50.8	18.10
ocnaddadminservice	1	0.08	33.90
ocnaddalarm	1	0.12	19.30
ocnaddconfiguration	1	0.09	40.70
ocnaddexport	1	0.03	2.05
ocnaddgui	1	0.00	3.09
ocnaddhealthmonitoring	1	0.13	35.70
ocnaddpcfaggregation	2	61.4	32.90
ocnadduirouter	1	0.01	32.50
zookeeper	3	0.14	48.20