ONUG recently hosted a webinar to highlight the progress of our Working Groups. The Orchestration and Automation (O&A) Working Group was the third to present what they’ve been working on, and what their goals are for the future. Group Co-Chairs Brian Silverman of McKesson, Michael Haugh of Gluware, and David Hegenbarth of Pliant led the discussion. You can watch the entire webinar here, or keep reading for a summary of the O&A Working Group’s objectives, priorities, and future plans. 

The group’s first whitepaper was a taxonomy document that focused on defining the problem. What does O&A mean to different organizations? Regardless of the size of your organization, the team found that O&A is something all companies struggle with. Their paper looked at an initial set of use cases and laid the fundamental pieces to drive the group forward. 

Laying the Groundwork

Haugh explained that conversations about O&A can go all over the place. So, they began work on their paper with the goal of creating base layer knowledge, defining domains, types, and comparing different automation approaches. They defined, described, and explained the following areas:

• Domains within the scope of O&A
• Users, types, and methods
• Automation comparison
• Best practices
• Solution requirements
• Getting started
• Example use-cases

Addressing Challenges

“It’s all about bringing software control,” explained Haugh. “We want to help organizations evolve their O&A approaches.” Organizations face the challenge of automating multiple domains, including legacy applications, non-programmatic infrastructure, WAN, and Software-Defined WAN. Evolving from a manual process to one that is policy-based and extends across domains is a challenge. The group narrowed down their focus to the following domains.

• Cloud
• Campus/LAN
• WAN
• SD-WAN
• Data Center
• Security with Proprietary Controller

Each domain has its challenges. For example, process-based automation may require integration into the management layer, including change management databases, ticketing systems, address management, as well as monitoring. While O&A encompasses a lot of layers, it’s “all about defining the software layer that gives you control over the network,” said Haugh. 

Details of the Paper

The O&A Working Group defined users of O&A as network engineers, operators, and architects, as well as DevOps teams and SRE. Types of O&A include network automation and orchestration, NCCM, policy-based automation, intent-based networking, and DevOps frameworks. “The goal is to automate from end to end,” explained Haugh. “That includes moving from manual to a model-driven, low-code/no-code infrastructure that can consume APIs.” 

Use Cases

The Working Group focused on three use case examples: 1) integrating with ticketing systems; 2) automating/troubleshooting, and 3) automating software updates. “There are a lot of entry points and a lot of pieces to the puzzle,” said Haugh. They explored three areas within these use cases, that Haugh categorized this way:

• Read Only: includes network inventory, config drift, config compliance, and capture state for troubleshooting.
• Read/Write: includes OS/patch upgrade, config management, ZTP, network security policy management, service provisioning, and auto-remediation. 
• Analytics: includes identifying patterns, network utilization, and predictive “what if” before change scenarios. 

Best Practices

Haugh next outlined best practices from the paper. As companies explore how to implement O&A, they should first perform a detailed assessment. Next, ensure their plan is aligned with corporate initiatives. Then, follow these practices: automate read-only first; start with compliance/audit; establish a process to test, verify and deploy; pre-check and post-check verify to be able to rollback, and start first with domains, then move across domains. 

Solutions Requirements

The ultimate goal is to be agile, minimize downtime, optimize performance, and secure the infrastructure. To implement a successful O&A strategy, companies must focus on basic requirements when evaluating solutions. The group’s paper breaks those requirements into three categories.

• Systems must allow for multi-vendor, role-based access, redundancy/failover, logging, scheduling, and database integration. 
• Device management solutions must include inventory, capture state information, and capture configuration. 
• Integration is essential, including API, programmatic interfaces, and support for standards. 

Moving Forward

The next paper the group is working on will focus on business case and ROI. However, first, the group is gathering data points by creating a framework for an internal assessment. That assessment is broken up in three ways. 

• People: What are my team’s skill sets with current and future technologies? Do we have a plan for professional development and new career paths for current employees? What about cross-functional teams, such as NetOps, SecOps, compliance, procure, etc.? Can we break down silos? What are our trends in hiring, outsourcing, or in-sourcing? 
• Technology: Assess your domains, both legacy and modern. What management systems do we need to integrate with? How are we monitoring those current systems? Then, assess your current O&A state. What is already in place? What is being performed manually? What network changes will affect O&A? 
• Process: You must understand how the process affects automation. Are there manual processes in place that slow down overall delivery? Look at what you are already measuring (availability, performance, change times, ticket open times), documentation, project management, bug/issue tracking, lifecycle management, and the approval process.

Silverman took over the discussion here to talk more about the upcoming business case and ROI paper. “The purpose of the business case framework is to help leaders drive automation within their organizations.”  It’s not meant to be a template or calculator. The framework will need to be customized by industry, scope, scale, and culture. It will cover important points, such as organizational maturity, resources, timeline, use cases, alignment with company strategy, and governance. 

Contribute to the Community

Currently, the Working Group meets weekly to discuss the paper’s outline and potential content. Their goal is to release the finished paper at the Fall ONUG live event. “We’d love your input,” Silverman emphasized. 

Hegenbarth concluded the session by highlighting that O&A is a new field, and ONUG’s Working Group is a great forum for people to bring their ideas. “Automation can be as simple as a single task that is repeated, or as complex as orchestration where you line up a lot of those tasks. When you move from a single task to orchestration, that’s the new part of this field.” O&A strategies are all over the place. This Working Group hopes that their collaboration and the resulting whitepapers will bring about a unified strategy to help companies make this transition. 

To get involved in any of ONUG’s Working Groups, click here. Learn more and see the results of this research at our next ONUG Spring, taking place May 6-7, 2021. . Reach out with any questions about the ONUG community.