Recommended Hardware Configuration
for SQream DB

While SQream DB can run on any NVIDIA GPU-enabled hardware, we have some recommendations to achieve the best performance

Learn what it takes to deploy a SQream DB cluster

Our reference architecture document is intended for influencers and decision makers, IT and system architects, system administrators, and experienced users who are interested in a comprehensive reference for a SQream DB installation

SQream DB reference architecture for 50-100 users

On-premise Distributed

This configuration is the most flexible one. In this multi-node solution, scaling up for more concurrency and storage is easiest

Dell C4130 or C4140 with NVIDIA Tesla P100 cards

Dell C4140 based solution

IBM Power AC922 with 4 NVIDIA Tesla V100 cards

IBM AC922 Power9 based solution
Datasheet: SQream DB for Power systems

Chassis

Chassis capable of housing 2-4 NVIDIA Tesla GPUs

Examples: IBM AC922 Power9, Dell R730/R740, Cisco UCS C240 M4, HP DL380g10, Lenovo System x3650 M5, Supermicro SYS-2028U / 6028U

CPU

Dual IBM Power9 or Dual Intel Xeon Gold processor

Recommended: IBM Power9 with (16C) or Intel Xeon Gold 6150 (18C)

RAM

256 GB RAM per logical GPU

Recommended: 512GB

NVIDIA Tesla GPU

At least 2 NVIDIA Tesla series GPUs

Recommended: 4x NVIDIA Tesla V100 with NVLink

Storage configuration

LUNRecommended amountSizeConfiguration
OS2x100GB SSDRAID1
Fast spool (optional)2x400GB SSDRAID1
DataNetwork attached storage, vendor dependent

Superior Scaling with a Distributed Solution

This is the preferred method for larger installations of SQream DB, as it allows for superior scaling in both compute power and storage. It can be paired with any SAN/NAS, from a variety of vendors.

Multi-node SQream DB scaling

Contact your SQream representative to have this solution tailored for your needs. Our experienced engineers can match the right hardware and software stack for your current and future needs.