How can we help?

Auvik collector deployment strategies for MSPs

Follow

As an MSP, there are a few key steps to bringing a customer network onboard with Auvik. But before we dive into customer onboarding, let’s figure out the best way of deploying Auvik based on your existing network architecture and resources available within your own and your customers’ facilities. 

First, do you operate your own data centre or co-location facility, and do you have a Layer 3 tunnel into some or all of your customers?

Note: You can modify active collectors, and switch between single and multi-tenant collectors at any time. This doesn’t affect your Auvik subscription in any way as Auvik’s pricing isn’t based on collectors.

Single-tenant collector

The Auvik collector is a lightweight, Ubuntu-based distribution that acts as a data collector and remote management proxy for your customers’ networks.

A single collector can be deployed for each group of inter-reachable subnets. If you have network segments that aren’t interconnected, you can deploy a collector for each siloed network segment. There’s no limit to the number of collectors you can run in parallel.

Regardless of the platform you deploy the collector on, the requirements for the single-tenant collector are:

  • 1 vCPU
  • 2GB RAM
  • 8GB disk space

You can deploy the Auvik single-tenant collector on a hypervisor, on bare metal, or in a container.

If you’re deploying the collector on:

  • A hypervisor
    • Running VMware
      • Deploy using the OVA-based installer
    • Running Hyper-V
      • Deploy using the scripted installer
    • Running something else (e.g., Proxmox)
      • Deploy using the scripted installer
  • Bare metal
    • Deploy using the scripted installer
  • A container (e.g., Docker, Linux)
    • Deploy using the scripted installer

Deploying dedicated hardware

For smaller customers or those with strict security requirements, deploying Auvik alongside existing customer infrastructure may not be possible.

As a result, some of our MSP customers want a box they can plug into for initial network discovery and ongoing monitoring. Since the Auvik collector is Linux-based, it can be deployed on low-powered hardware with an x86 / x64 processor. Typically we see MSPs use:

  • Low-powered home theatre PC (HTPC) hardware
  • Micro-PCs, like the Intel NUC

The Auvik collector firmware can’t currently be deployed on ARM-based platforms like the Raspberry Pi.

Multi-tenant collector

As an MSP, you may have a data centre or co-location facility that provides a central point of Layer 3 reachability into your customers’ sites. In such an environment, deploying a dedicated instance of the Auvik collector for every customer leads to high resource utilization on your hypervisor, or increases the number of bare metal boxes you’re shipping out to customer sites.

In this case, you have the option to deploy a shared collector, a single instance of Auvik that maps and monitors multiple client networks.

The amount of resources required for the multi-tenant collector are proportional to the number of shared tenants. But at a minimum, we recommend dedicating the following resources to a shared collector:

  • 2 vCPUs
  • 4GB RAM
  • 16GB disk space

Just like the single-tenant collector, the Auvik multi-tenant collector can be deployed on a hypervisor, on bare metal, or in a container.

If you’re deploying the collector on:

  • A hypervisor
    • Running VMware
      • Deploy using the OVA-based installer
    • Running Hyper-V
      • Deploy using the scripted installer
    • Running something else (e.g., Proxmox)
      • Deploy using the scripted installer
  • Bare metal
    • Deploy using the scripted installer
  • Container (e.g., Docker, Linux)
    • Deploy using the scripted installer

If you have any questions about your deployment strategy, please contact Auvik support.

Have more questions? Submit a request

Comments

Powered by Zendesk