ITC3, LLC announced Network-Operator – a revolutionary new network appliance and infrastructure orchestrator built entirely upon the principles of Chaos Engineering, with a focus on microservice architecture integration with traditional and legacy networks. The containerized appliance will be available in a free BETA release from the company’s website (www.itc3.io) in the coming weeks. Network-Operator – written completely in Go (by Google) runs on top of the lightweight Alpine Linux operating system and was created with the highest priorities placed on automation, pipelines, high availability and API controls.
Chief Architect on the project, Jeremy Pogue, added: “From day 1 the entire project has been built on Preemptible instances, that delete themselves randomly with a maximum lifespan of 24 hours. So for the life of this project, the environment (except for the git repos) is destroyed every single day at a random time. That was extremely difficult to deal with early on, but it has provided us with solid automation pipelines, unit and integration testing, auto-recoverability, and powerful high availability mechanisms at the network layer. To my knowledge, this has never been done for a true router / network appliance.”
Another high priority focus is CI/CD.
Pogue added, “As a traditional network architect, I really wanted to challenge myself and apply the things I had learned over my time embedded with numerous software development teams building network connectivity for Kubernetes clusters. While common practice today for software releases, we have not seen this with networking. I believe this is due to the challenges in being able to “test” your network as you would with a software release. The whole reason for Network-Operator was to embed network engineer level knowledge into a software program. For this to work we had to be able to create an environment, fail it, write tests for that failure, tear it down, and start all over again… this would have been impossible to do without complete CI/CD pipelines.”
Time to connect
Creating secure and stable site-to-site and cloud connections is hard, but the benefits are undeniable. Routers, Firewalls, VPNs, mobile connectivity, Intrusion Prevention Appliances, Intrusion Detection Appliances… with Network-Operator you can throw out all of that, and replace it with a simple container.
Network-Operator provides a zero-touch deployment, which creates a full-service, interconnect orchestrator and provides a modern, secure overlay built with microservices into and between any environment, anywhere you have Internet connectivity. For example:
- Two (or unlimited) physical locations
- One (or unlimited) physical location + a colocation facility
- AWS + Google Cloud (or any number of cloud providers and / or regions)
- Physical location + Azure (or any number of cloud providers)
Simply said, Network-Operator creates secure links between any location, and you can forget about networking, and focus on running your business.
How it works
Network-Operator is completely containerized. The only dependency is that it requires Docker… which means Network-Operator can run virtually anywhere. Cloud, Bare metal, On-premise, Kubernetes deployment, Google Cloud, AWS, Azure, GovCloud, just to name a few. If Docker runs, Network-Operator runs.
Network-Operator removes the complexity of networking between locations. Install in a Docker environment and begin connecting. The result is stable connections with the following features:
- Layer 2 Connectivity – full mesh environments via IPSec tunnels built with Tinc
- Layer 3 Connectivity – via GoBGP
- Log Delivery to Elasticsearch – shipped via FluentD
- Monitoring – using Prometheus and SNMP
- Visualization and Intuition – custom built exporters integrated with Netflix Vizceral
- Charts & Graphs – from NetData
- BGP Traffic Policies – by Flowspec
- Security – IPSec tunneling with AES Encryption in active state; KMS based certificates for encrypting data at rest
- Self-healing
Why it works (all the time)
One word… Chaos. We bring together a lot of really powerful open source projects. What we really wanted to do was build an operator that put the power of these projects in the hands of the average network owner. Chaos Engineering is a method of experimenting on infrastructure in order to build confidence in the system’s capability to withstand turbulent real-world conditions. Complex distributed systems – as managed by Network-Operator – are hard to manage. Everything eventually fails. Network-Operator will manage and repairs those failures for the average administrator just like it has for us thousands of times over the past few years.
Network-Operator was designed, built, and tested on the on the most unpredictable environment we could create, that self destructs randomly all day while we are working with it. It automatically heals, redeploys, configures, manages, destroys, rebuilds, and cleans up on its own. While we do allow for configuring and customizing the environment via yaml based config file, you can literally deploy a global hybrid interconnect directly out of the box and touch nothing. The auto deployment and self-healing features of Network-Operator work so seamlessly, users will only need to sit back and admire the gorgeous monitoring charts and dashboards.
Additional information will be available @Twitter, LinkedIn, GitHub, or at itc3.guru for updates and the official BETA release date.
