From Building Network Automation Solutions
At a Glance
- High-intensity interactive online course;
- Jump-start your network automation career;
- Hands-on experience working on a solution to your own problem;
- 6 week course spread across ~2 months;
- Live discussion and guest speaker sessions;
- Design and coding assignments and group work;
- Final course completion certificate.
Data models are the crux of any successful network automation solution, and getting them just right is a mixture of science and art.
The guidelines discussed in this section include:
- Abstract everything - focus on data that describe your requirements, not data that the device need to be configured
- Avoid data duplication - never store the same bit of information in two places
- Separation of infrastructure and service data - define a common data model describing network infrastructure, and separate data models for individual services
We’ll also talk about data stores (from text files to relational databases), discuss their pros and cons, and figure out how to use Ansible with external data stores.
Hands-on assignments for this section include:
- Build a simple data model for VLAN service;
- Create a data model for VXLAN-based data center fabric;
- Model a simple Service Provider service (Internet access or L3VPN);
All assignments require you to build a data model and a sample configuration template to verify the data model adequately describes your service.
David Barroso the author of SDN Internet Router, co-author of NAPALM (a network automation abstraction library), and author of YANG (and OpenConfig) support in NAPALM.
In Autumn 2017 session he'll talk about OpenConfig support in NAPALM and demonstrate how you can use YANG data models to modify device configurations and get structured operational data in a multi-vendor environment.
Previous guest speakersMarcel Wiget started his professional as a UNIX administrator and software developer, and is now a Product Manager at Juniper Networks. Networking and Automation based on IETF standards and open source tools have always been at the center of his professional carrier, and he firmly believes in replacing traditional network management technologies like SNMP with user-designed vendor-neutral and model-driven approach to network management like OpenConfig.
His talk will introduce OpenConfig for configuration and telemetry and close with practical examples demonstrating the network capabilities of Ansible network modules in playbooks in combination with OpenConfig/YANG to provision Juniper Networks routers and switches.
This section focuses on the infrastructure and services data models. Watch:
- Abstract Everything section from the Network Automation Use Cases webinar (45 minutes)
- DMVPN Case Study parts of the Ansible for Networking Engineers webinar (1 hour)
- Automating Data Center Fabric deployments (1 hour 30 minutes) - focus on the Separate Data from Code part of the presentation (25 minutes)
- NETCONF, YANG and OpenConfig presentation by Marcel Wiget
- Data Models, Data Stores and Data Model Transformations presentation
Additional recommended materials
If you want to understand YANG and NETCONF better, watch the first part of the NETCONF and YANG webinar (~1 hour if you skip the hands-on parts). You might also want to watch the NETCONF Deep Dive section to understand the NETCONF protocol details.
You might also want to listen to Model Driven Networking podcast with Jeff Tantsura.
Recording of the January 2017 session
NETCONF, YANG and OpenConfig
- Vendor-Neutral YANG Data Models (PDF)
- Introduction to NETCONF and YANG
- OpenConfig and Custom YANG Data Models
- Using YANG Data Models on Junos