Bare-Metal Hardware Specifications
This document lists the hardware specifications for on-prem bare-metal deployment of version 2.8.0 of the Iguazio Data Science Platform (“the platform”).
The platform is available in two configurations, which differ in a variety of aspects, including the performance capacity, footprint, storage size, and scale capabilities:
- Development Kit
- A single data-node and single application-node cluster implementation. This configuration is designed mainly for evaluation trials and doesn’t include high availability (HA) or performance testing.
- Operational Cluster
- A scalable cluster implementation that is composed of multiple data and application nodes. This configuration was designed to achieve superior performance that enables real-time execution of analytics, machine-learning (ML), and artificial-intelligence (AI) applications in a production pipeline. The minimum requirement for HA support is three data nodes and three application nodes.
Development Kit Specifications
The Development Kit configuration is composed of a single data node with 10 TB of usable capacity and 4 application nodes in a 6U form factor.
The Development Kit configuration in bare-metal platform deployments must fulfill the following hardware specification requirements:
|Input/Output operations per second (IOPS )||Up to 100,000|
|Read/Write throughput||12 GB/sec|
|Form factor||6 RU|
|OS boot disk||1 TB|
|Usable storage capacity||10 TB NVMe|
|High availability (HA)||None|
|Data switch||1 switch
16 * 100 GbE
|Power consumption||1700W (typical)|
Operational Cluster Specifications
The Operational Cluster configuration is composed of three data nodes and four application nodes in a 13U form factor. The cluster’s high-availability (HA) scheme supports two replicas of the data so there’s no single point of failure (SPOF). (For example, node A can split its protected data 50%–50% between node B and node C.) The platform’s hardware is high-end and enterprise-grade in terms of reliability.
The platform is built on high-end NVMe solid-state drives (SSDs) with 2M hours of mean time between failure (MTBF) and 5 drive writes per day (DWPD) endurance.
The product architecture is designed for each cluster to scale up to 10 nodes with an overall usable capacity of up to 430 TB.
The Operational Cluster configuration in bare-metal platform deployments must fulfill the following hardware specification requirements:
|Input/Output operations per second (IOPS )||Up to 300,000|
|Read/Write throughput||12 GB/sec per node|
|Form factor||13 RU|
|OS boot disk||1 TB|
|Usable storage capacity||20–120 TB NVMe|
|High availability (HA)||1 + 1|
|Data switch||2 switches
16 * 100 GbE per switch
|Power consumption||3100W (typical)|