Installation Options
This article outlines the various installation methods and modes that are available for customer selection.
| Item | Embedded Cluster | Existing Cluster |
|---|---|---|
| Description | Establishes a functional Kubernetes cluster on Linux servers. Subsequently, iceDQ is integrated into the Kubernetes cluster, hence the name “embedded cluster”. | iceDQ is installed on a customer-managed dedicated Kubernetes cluster, such as AWS EKS, AKS, or GKE, hence the name “existing cluster”. |
| When to choose | Your organization either doesn't have Kubernetes services or, if it does have them, the team managing them cannot provide a dedicated cluster for installation. | You organization has Kubernetes services, an expert team for managing Kubernetes and can readily provide a dedicated cluster for installation. |
| Online Install | The servers have internet access so the customer can use one-line command that downloads packages in real-time and does the installation. | The servers have internet access so the customer can use one-line command that downloads packages in real-time and does the installation. |
| Offline Install | The servers lack internet access available so the customer must manually download the installation package and transfer it to the servers where the installation will take place. | The servers lack internet access available so the customer must manually download the installation package and transfer it to the servers where the installation will take place. |
Proof Of Concept Sizing
Below are the different embedded and existing cluster sizing recommendations for running a proof of concept.
- Functional: You are interested in only verifying the functionality.
- Performance: You are interested in verifying the functionality as well as performance & scalability.
Embedded Cluster
Functional
- One machine with 16 vCPUs
- 32GB of RAM on the machine
- 850GB+ of SSD space for the machine
Performance
- One machine (a) with 64 vCPUs for handling engine workload
- 64GB of RAM for machine (a)
- Two machines (b & c) with 32 vCPUs for handling application workload
- 32GB of RAM for machine (b & c)
- 1TB of SSD space for every machine
important
Please check Embedded Cluster System Requirements for additional pre-requisites
Existing Cluster
Amazon Kubernetes Service (EKS)
If you want to deploy in a dedicated Amazon Kubernetes Service Cluster.
| Item | Functional | Performance |
|---|---|---|
| Cluster Size | Single Node = 1 Worker Node | Multi Node = 2 Worker Nodes |
| Instance Size | m5.4xlarge | m5.8xlarge (each node) |
| RWX Class Storage | 750GB EFS | 1TB EFS (each node) |
| Load Balancer | Network Load Balancer | Network Load Balancer |
| Database (Optional) | RDS PostgresSQL v15+ | RDS/ Aurora PostgresSQL v15+ |
| Backup Store (Optional) | S3 | S3 |
Azure Kubernetes Service (AKS)
If you want to deploy in a dedicated Azure Kubernetes Service Cluster.
| Item | Functional | Performance |
|---|---|---|
| Cluster Size | Single Node = 1 Worker Node | Multi Node = 2 Worker Nodes |
| Instance Size | Standard_D16ds_v4 | Standard_D32ds_v4 (each node) |
| RWX Class Storage | 750GB CSI | 1TB CSI (each node) |
| Load Balancer | Azure Load Balancer | Azure Load Balancer |
| Database (Optional) | Azure PostgreSQL v15+ | Azure PostgreSQL v15+ |
| Backup Store (Optional) | Azure Blob | Azure Blob |
Google Kubernetes Service (GKE)
If you want to deploy in a dedicated Azure Kubernetes Service Cluster.
| Item | Functional | Performance |
|---|---|---|
| Cluster Size | Single Node = 1 Worker Node | Multi Node = 2 Worker Nodes |
| Instance Size | c3d-standard-16 | c3d-standard-30 (each node) |
| RWX Class Storage | 750GB | 1TB (each node) |
| Load Balancer | Google Load Balancer | Google Load Balancer |
| Database (Optional) | PostgreSQL v15+ | PostgreSQL v15+ |
| Backup Store (Optional) | Google Cloud Storage | Google Cloud Storage |
important
Please check Existing Cluster System Requirements for additional pre-requisites