Zero Touch Lifecycle Management

Most vendors in the networking business offer some form of Zero Touch Provisioning (ZTP).  And most enterprise network engineers run ZTP to install equipment and automate day 0 configuration.  Some use software-defined networking for day 1 operations.  Nobody really does Zero Touch Lifecycle Management (ZTLM) except for hyper-scalers.  Until now.  Hedgehog offers a turnkey software appliance so enterprise customers can network like hyper-scalers with completely automated ZTLM, including:  

  • Day 0: rack, stack, wire, install and configure with ZTP
  • Day 1: provision multiple tenants with Virtual Private Clouds and Transit Gateway services using GitOps 
  • Day 2: monitor AI workload performance, identify anomalies, and inspect the network to determine root cause
  • Day 3: upgrade software, add, remove and replace network equipment with Zero Touch Lifecycle Management 
Get Hedgehog
ztlm-icon

Turnkey Software Appliance

Hedgehog offers a turnkey software appliance that includes everything you need to run hybrid AI cloud infrastructure.  The appliance combines over a dozen open source projects.  The list includes ONIE, Flatcar Linux, SONiC, FRR, DPDK, K3S, Grafana Alloy, and Hedgehog to name a few.   Working from the ground up, here's how the software appliance bundles it all together:

  • SONiC is the open source network operating system that runs your switches.
  • Your infrastructure as code describes the configuration of your network.  We encourage you to treat it as source code in a version control system like Git.  
  • The Hedgehog ZTLM (aka "fabricator") installs, updates and removes all software components in your Hedgehog cloud.
  • The Hedgehog control plane models your network as Kubernetes custom resource definitions and uses K8s to enforce configuration state in your network.
  • The Hedgehog data plane receives instructions from the control plane to route packets throughout your Hedgehog hybrid AI cloud network.
  • The Hedgehog Virtual Private Cloud is an API you use to provision tenants with their own private infrastructure
  • The Hedgehog Transit Gateway peers Hedgehog VPCs with external VPCs on AWS, Azure, GCP or other Hedgehog locations with the same cloud native API.
hedgehog-software-appliance-1

Experience Hedgehog in a Virtual Lab

Hedgehog is free to download and evaluate.  Just fill out a short form to download Hedgehog, then install our software appliance in virtual lab mode.  Hedgehog will then install and configure a virtual data center with Linux virtual hosts, SONiC virtual switches, and Hedgehog virtual controllers.  You can then use the Hedgehog API to create VPCs, test multi-tenant isolation, and peer VPCs.  You can also test open observability with an export to Grafana Cloud.  

  1. Download the Hedgehog software appliance to a host that meets system requirements
  2. Install the appliance as a virtual lab
  3. Run a virtual data center
  4. Use Hedgehog API to create VPCs
  5. Use Hedgehog API to peer VPCs
Group 208

Day 0 - Zero Touch Provisioning

Day 0 is really the only day that has serious physical touch.  A human or two are still required to rack, stack and wire your servers and switches.  Ask your hardware solution provider to provide you with help here if you need it.  From there, you simply write some YAML that represents how you have wired your rack, then build the Hedgehog software appliance with your infrastructure as code.  When you run the appliance, it will install and configure software on your network switches and hosts so you have a functional hybrid cloud.  

 

  1. Rack, Stack & Wire
  2. Build & Deploy Hedgehog Controllers
  3. Observe as the fabric self-provisions
Group 161

Day 1 - Cloud Native Operations


Once you are up and running, you are free to run your cloud operations as you wish. Hedgehog is built on Kubernetes, so you have an inherently cloud native API you can use to automate your cloud operations with best practices.  Provision multiple tenants with Hedgehog VPCs using Ansible, Terraform or a self-service customer console.  Implement GitOps with version control of your Hedgehog infrastructure as code.  Deploy your network configuration with Kubernetes the same way you manage CI/CD for your AI workloads.  Likewise for observability.  Monitor your network logs, metrics and traces with Grafana in the same dashboards you use for your applications, compute and storage.  

Use Hedgehog API for multi-tenant cloud operations:

  1. Provision tenant VPCs
  2. Peer external VPCs through the Hedgehog Transit Gateway
Group

Day 2 - Open Observability

Now your cloud is humming.  You have lots of tenants running their AI workloads on shared infrastructure.  What could possibly go wrong?  Murphy's Law always applies.  Monitor everything in one place to get a clear performance picture.  When you identify anomalies, use the Hedgehog inspect command to figure out what's going wrong .   

  1. Monitor workloads, network, compute, storage all in the same Grafana dashboards

  2. Identify anomalies

  3. Inspect to determine root cause

Group (1)

Day 3 - Zero Touch Lifecycle Management

 

More than a dozen different open source projects is a LOT to maintain and upgrade. Thankfully, Hedgehog automatically upgrades your open network software.  As your business expands, your AI workloads will need more infrastructure.  Hedgehog automates the process of adding switches and hosts to your network.  You can remove and replace switches and hosts, too.  

  1. Upgrade software

  2. Add switches

  3. Remove switches

  4. Replace switches

Group 213