HCL CDP Hardware Sizing Details For AWS
This section outlines the minimum hardware requirements to guide AWS infrastructure sizing for the deployment of HCL CDP in support of application deployment.
The objective of this guide is to provide an estimated baseline for the AWS infrastructure components—CPU, memory, storage, and network resources—necessary to achieve expected performance levels and ensure a reliable, efficient operational environment.
The hardware sizing details include:
- Number of servers per component
- Per-server specifications (vCPU, RAM, and Storage)
- Total resource footprint
These recommendations serve as a starting point for capacity planning and should be adapted based on workload characteristics, usage patterns, and scaling expectations.
Component-wise Hardware Summary
The following table provides a detailed breakdown of AWS infrastructure requirements for each system component. For every listed component, it outlines the number of servers allocated, the per-server hardware specifications (vCPU, RAM, and HDD), and the resulting total resource consumption. This granular view helps assess the compute, memory, and storage needs across the architecture and supports capacity planning, resource optimization, and scaling decisions.
| Component | No. of Servers | vCPU (per server) | RAM (GiB) | HDD (GB) | Total vCPU | Total RAM | Total HDD |
|---|---|---|---|---|---|---|---|
| Pixel Listeners | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| SST HS - Zookeeper | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| SST HS - Aerospike | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| SST HS - Processor | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| SST HS - Offline Processor | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| Mongo | 2 | 2 | 8 | 100 | 4 | 16 | 200 |
| Mongo Connector | 1 | 2 | 4 | 50 | 2 | 4 | 50 |
| RTS | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| RTS - Offline | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| Athena | - | - | - | - | - | - | - |
| Druid | 1 | 8 | 64 | 200 | 8 | 64 | 200 |
| UI/Core API | 1 | 1 | 4 | 50 | 1 | 4 | 50 |
| SFTP | 1 | 1 | 4 | 200 | 1 | 4 | 200 |
| DMP-Producer | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| PII Cache | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| MySQL Master | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| MySQL Slave | 1 | 4 | 16 | 100 | 4 | 16 | 100 |
| Trigger | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| Scheduler | 1 | 4 | 16 | 100 | 4 | 16 | 100 |
| Scheduler Core | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| Trigger Core | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| Scheduler Sender | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| Trigger Sender | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| Webhooks | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| TC Redis | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| Flip | 1 | 1 | 4 | 100 | 1 | 4 | 100 |
| Flip AS | 1 | 1 | 8 | 100 | 1 | 8 | 100 |
| Celery Worker | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| Celery Redis | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| Celery RMQ | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| DI API | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| Prometheus | 1 | 2 | 4 | 200 | 2 | 4 | 200 |
| Grafana | 1 | 1 | 4 | 50 | 1 | 4 | 50 |
| Kafka Connectors | 1 | 4 | 8 | 100 | 4 | 8 | 100 |
| Kafka Zookeeper | 1 | 4 | 8 | 100 | 4 | 8 | 100 |
| Kafka Brokers | 1 | 4 | 8 | 200 | 4 | 8 | 200 |
| Kafka ControlCenter | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| Kafka SchemaRegistry | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| Neo4j Graph | 1 | 4 | 8 | 100 | 4 | 8 | 100 |
| Offline Segmentation | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| Audience Uploader | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| ES - Master Nodes | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| ES - Data Nodes | 1 | 2 | 16 | 100 | 2 | 16 | 100 |
| ES - Ingest Nodes | 1 | 2 | 4 | 100 | 2 | 4 | 100 |
| ES - Coordinating Nodes | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| ES - Kibana | 1 | 2 | 8 | 100 | 2 | 8 | 100 |
| FluentBit | 2 | 2 | 4 | 100 | 4 | 8 | 200 |
| S3Log Uploader | 2 | 4 | 8 | 200 | 8 | 16 | 400 |
Disclaimer - This guide illustrates the minimal hardware sizing requirement, based on various assumptions. This guide does not replace the need for specific Infrastructure details before the deployment. The mentioned consumptions are based on a specific scenario and hardware. Proper estimation should be done before buying the hardware. On top of the suggested hardware, appropriate buffer should be added for any expected peak/future load.
Category-wise Hardware Summary
| Category | Components | Total vCPU | Total RAM (GiB) | Total HDD (GB) |
|---|---|---|---|---|
| Data Layer | Mongo, MySQL Master, MySQL Slave, Aerospike, TC Redis, Celery Redis, Kafka (all), Neo4j | 30 | 80 | 1000 |
| Processing Layer | Processor, Offline Processor, Celery Worker, Celery RMQ, Trigger, Scheduler, Core/Sender | 27 | 68 | 900 |
| Monitoring | Prometheus, Grafana | 3 | 8 | 250 |
| UI/API/Access | UI/Core API, SFTP, DI API, Athena (not specified), Flip, Flip AS | 9 | 28 | 650 |
| Logging & Search | Logging - ES (all), FluentBit | 18 | 52 | 800 |
| Uploaders | S3Log Uploader, Audience Uploader | 10 | 20 | 500 |
| Other | Pixel Listeners, PII Cache, RTS, RTS Offline, Webhooks, DMP Producer, Zookeeper, Druid | 23 | 76 | 1150 |
Total Infrastructure Summary
| Resource | Total |
|---|---|
| Total Servers | 54 |
| Total vCPUs | 110 |
| Total RAM (GiB) | 372 |
| Total HDD (GB) | 5450 |