 Installation

EdgeSet is a multi-user web application. It’s designed to run on-premise or in a private cloud. You can install it on a physical server (“bare metal”) or a VM (such as from AWS, Azure, GCP, Digital Ocean, etc.).

Requirements

EdgeSet requires an 64-bit Intel-compatible (x86-64) or ARM (AArch64) CPU.

Table 2.1: Minimum system requirements
Concurrent query slots Memory CPU Cores Disk
2 8GB 2 16GB
5 12GB 6 20GB
15 24GB 16 30GB
50 64GB 64 60GB

EdgeSet also requires UEFI support.1

1 Most computers since 2006 have UEFI support.

Scaling up

If you save query results to “Internal Storage”, you will likely need more disk space. The amount of disk space depends on the size of the saved query results.

  • For VM deployments, you can increase the size of the virtual disk and EdgeSet will detect and use the additional space after a reboot.
  • For physical deployments, you can image the EdgeSet disk to a larger disk and EdgeSet will detect and use the additional space after a reboot.

RAM is allocated evenly between query slots. If you find that your queries are failing due to a lack of memory, you should increase the amount of system RAM.

AI Acceleration

EdgeSet’s embedded AI models are designed to run without any special hardware. However, you will notice increased performance on CPUs with vector instruction sets:

Table 2.2: CPUs supported by EdgeSet for accelerated AI inference
Make Series Example Model
AMD Piledriver FX-8350 (2012)
Zen 4 Ryzen 7 7800X3D (2023)
Amazon Graviton
Ampere Altra Q80-30 (2020)
Altra Max M128-30 (2021)
AmpereOne A192-32X (2025)
Google Axion
Intel Alder Lake Core i5-12400 (2022)
Cascade Lake Xeon Gold 6258R (2020)
Cooper Lake Xeon Gold 5300H (2020)
Haswell Intel Core i7-4790K (2014)
Ice Lake Xeon Gold 6330 (2021)
Ivy Bridge Core i5-3570K (2012)
Sandy Bridge Intel Core i5-2500K (2011)
Sapphire Rapids Xeon Gold 6430 (2023)
Skylake Core i5-6600K (2015) Microsoft Cobalt 200

Hyperscalers

Amazon Web Services (AWS)

Launch an EC2 instance using the EdgeSet AMI ID provided to you.

Google Cloud Platform (GCP)

Launch a GCE instance using the QCOW2 disk image (x86_64 or arm64).

Microsoft Azure

Launch an Azure Virtual Machine using the VHDX disk image (x86_64).

VPS Providers

Digital Ocean

Import the QCOW2 disk image (x86_64 or arm64) to your DigitalOcean account as a “Custom Image” (see https://www.digitalocean.com/blog/custom-images/). Launch a droplet with the custom image.

Hetzner Cloud

Import the raw disk image (x86_64 or arm64) as a snapshot (i.e. using hcloud-upload-image). Launch a new VM with the snapshot.

Physical Servers

Copy the raw disk image (x86_64 or arm64) to the target hard drive (e.g. using dd). The target drive should be bootable after imaging.

Hypervisors

Hyper-V

Create a virtual machine using the VHDX disk image (x86_64).