Original source: Outshift by Cisco
This video from Outshift by Cisco covered a lot of ground. 19 segments stood out as worth your time. Everything below links directly to the timestamp in the original video.
Imagine asking your network system a question and getting an immediate, detailed answer, much like you would from a human expert. This new AI assistant aims to make complex network management tasks more accessible and efficient for IT professionals.
New Catalyst Center AI Assistant Enters Controlled Availability for Network Operations
Cisco has launched the Catalyst Center AI Assistant, an integration application designed to streamline network operations by answering configuration questions, providing compliance status, and offering operational insights. Built on Meraki's cloud backend, the AI Assistant leverages specialized large language models and is currently in controlled availability for early access customers, requiring the Catalyst Center to be cloud-registered.
This new tool aims to enhance network management efficiency by automating access to crucial information, from software upgrade procedures to device compliance statuses. Its phased rollout to early access customers indicates a strategic approach to scaling the technology, ultimately intending to support all Catalyst Center users and address the growing complexity of modern network infrastructures through intelligent automation.
"If you want to get access to the AI assistant, please talk with your account manager with the SE and ask them to give you early access. We are starting as controlled availability because we are starting to scale."
Catalyst Center AI Assistant Provides Detailed Troubleshooting Insights Through API Integration
The Catalyst Center AI Assistant demonstrates its capability to analyze extensive network data through API calls to deliver detailed troubleshooting information. During a live demonstration, the assistant rapidly provided insights into access point health, including air quality, power consumption, CPU utilization, and link utilization, crucial for identifying and resolving network issues efficiently.
By functioning as an integration app that communicates with the cloud infrastructure via APIs, the AI Assistant streamlines root cause analysis and proactive network monitoring. This approach mirrors existing integration methods, ensuring that the new AI capabilities fit seamlessly into current network management workflows, offering a rich and quick way to gather operational details.
"It will be very quick to actually get information and details about something that may happen in our infrastructure."
Programmatic Configuration Management Enhances Scalability and Consistency in Network Operations
Programmatic configuration management, leveraging CI/CD pipelines with tools like Jenkins and GitHub, significantly improves the scalability and consistency of network operations. This approach enables automated deployment of configuration templates to hundreds or thousands of devices, ensuring that changes are first tested in a lab environment before being rolled out to production, thereby reducing risks and maintaining infrastructure integrity. Integrated change control and maintenance scheduling further enhance operational efficiency.
By automating template deployments and incorporating version control, organizations can achieve a standardized and efficient process for managing network configurations. This method minimizes manual errors, ensures uniform application of policies, and provides detailed visibility into changes, marking a significant advancement over traditional, manual configuration methods for large-scale enterprise networks.
"With any pipelines it makes it so much easier to scale. We can do it programmatically across hundreds of devices. We can do it very consistent. It's very efficient."
▶ Watch this segment — 1:19:49
Terraform Automates Declarative Software Upgrades for Network Devices
Cisco's Catalyst Center Terraform provider facilitates declarative software upgrades, enabling network administrators to define desired image versions, and Terraform then identifies and updates non-compliant devices. The process involves terraform init to download the provider, terraform plan to outline changes, and terraform apply to execute the upgrade, which leverages built-in parallelism to update multiple devices simultaneously.
This method offers a more abstracted and stateful approach to software management compared to scripting with Ansible or Python SDKs. Terraform tracks the current state of devices, ensuring that only those requiring updates receive them, thereby enhancing efficiency and consistency across the network and reducing the need for manual intervention in large-scale deployments.
"If the device doesn't need to be updated, it will skip the device because hey, the device doesn't need a new software image because it already has that image."
▶ Watch this segment — 1:12:08
GitLab CI/CD Pipelines Integrate PyATS for Post-Configuration Operational Verification
GitLab's CI/CD pipelines offer robust capabilities for automating network configuration changes, with .gitlab-ci.yml files defining stages for executing Ansible playbooks, Terraform plans, or Python code. A crucial aspect of these pipelines is the integration of PyATS, which verifies operational data post-configuration. While automation tools like Ansible confirm configuration changes are applied, PyATS ensures that these changes have the desired functional impact, such as a new route being successfully advertised and learned by neighboring routers.
This comprehensive approach addresses the challenge of validating that configuration changes not only execute without error but also achieve the intended operational state within the network. By automating both the deployment and verification steps, network engineers can achieve higher levels of consistency and reliability, moving towards enterprise-grade automation solutions previously common only in software development.
"PyATS would go and will check operational data... making sure that the configuration was done as I would want it to be."
▶ Watch this segment — 1:28:10
Cisco Showcases Mature Catalyst Center API Ecosystem for Network Automation
Cisco has made the code used in the session, including Python, Ansible, and Terraform examples, publicly available on GitHub, providing developers with valuable resources for network automation. This initiative highlights the evolution and maturity of Catalyst Center's API platform, which has grown significantly since its launch in 2018. The Python SDK followed in 2019, Ansible collection in 2020, and the Terraform provider in 2021, culminating in an enterprise-grade suite of tools.
This continuous development demonstrates Cisco's commitment to enabling robust, scalable network management through infrastructure as code. The availability of these well-supported tools allows network engineers to confidently adopt various automation platforms, tailoring solutions to their preferred methods while leveraging a mature ecosystem.
"We have an enterprise-grade offering that we support, and that you can take it and use it with confidence."
▶ Watch this segment — 1:33:23
Integration Apps Enable Consistent and Scalable Infrastructure as Code Workflows for Catalyst Center
Integration apps are crucial for connecting Catalyst Center with third-party platforms, serving as a bridge for communication, notifications, and, critically, for facilitating infrastructure as code (IaC) workflows on CI/CD pipelines. This architecture ensures consistency, scalability, and enhanced security across network operations. By leveraging these apps, organizations can streamline configuration management and automate responses to network events, whether integrating with cloud services or on-premises data centers.
Running IaC workflows on platforms like Jenkins or GitLab via integration apps ensures that the entire team uses the same version of automation, leading to consistent outcomes and experiences. This standardized approach helps enforce security policies and allows for scalable network management, extending its utility to generative AI tools and other advanced applications.
"This is how you should run your infrastructure code workflows... on an integration app on a server like Jenkins, GitLab, any CI/CD pipelines because in that way the entire team can use exactly same integration, the same workflow."
Jenkins Automates Catalyst Center Template Deployments with Docker Containers for Consistency
Jenkins, when used with Catalyst Center's SDK, can automate template deployments within a Docker container, ensuring consistent execution across various users and environments. This pipeline code, which can also incorporate Ansible or Terraform, facilitates the installation of necessary libraries and provides transparent console output, allowing for easy troubleshooting and visibility into every step of the deployment process.
This approach eliminates inconsistencies that can arise from manual or laptop-based scripting, where different team members might use outdated versions of scripts. By centralizing automation on a pipeline platform, organizations can achieve enterprise-level scalability, security, and uniformity in network configuration, supporting complex workflows triggered by various tools, including AI platforms or ServiceNow.
"If Adrian wants to run the same script I give it to him today two days later or two weeks later I have done improvements it's a much better script. He's running the old version because nobody told him. We cannot scale at basically from an enterprise perspective."
▶ Watch this segment — 1:23:04
Catalyst Center Event Platform Delivers Real-Time Network Notifications to Multiple Destinations
The Catalyst Center event platform provides real-time notifications for network issues by collecting and publishing events from various sources, including switching, routing, wireless, and system integrations. Users can subscribe to an event catalog, enabling simultaneous delivery of alerts to multiple destinations such as webhooks, Webex, ServiceNow for incident creation, and PagerDuty.
This robust framework ensures that critical network events are communicated promptly to relevant systems and personnel, facilitating rapid response and mitigation. With 271 available events, the platform supports comprehensive logging, long-term data retention, and seamless integration with ITSM tools, enhancing overall network operational efficiency and incident management.
"I could subscribe to multiple events. Maybe I want two webhooks. I want a Webex notification, send something to ServiceNow to create an incident, and PagerDuty all at the same time."
AI Code Generation Requires Specific Library Versioning to Overcome Knowledge Cut-off
When using AI models like ChatGPT for code generation, users must specify the exact library and version for Catalyst Center SDK, Ansible, or Terraform. This is crucial because AI models have knowledge cut-off dates, meaning they may provide outdated code if not explicitly instructed to use a current library version. Failing to do so can result in generated code that relies on old APIs or practices, leading to errors and inefficiencies.
By providing precise instructions regarding libraries and their versions, developers can leverage AI to generate highly accurate and up-to-date code, saving significant development time. While AI-generated code may not be perfect, it can achieve 90-95% accuracy, significantly streamlining the initial coding phase for tasks like Ansible playbooks or Terraform configurations, making it essential to test the output in a lab environment.
"What you need to do is you need to ask the large language model of your choice, hey, I want you to help me, but I want you to use this library... And I want you to do it for this version of the library."
▶ Watch this segment — 1:35:30
Catalyst Center Offers Robust Out-of-the-Box Integrations for Comprehensive IT Workflows
Cisco's Catalyst Center provides a suite of out-of-the-box integrations, including options for ITSM, PagerDuty, Webex, and Splunk, designed to enable seamless communication with external platforms. These integrations facilitate diverse workflows such as CMDB inventory synchronization, automated software upgrades, and incident creation in ITSM systems like ServiceNow. By reformatting notification messages to suit specific platform requirements, Catalyst Center ensures effective interoperability and comprehensive IT automation.
For more complex scenarios where third-party platforms in the cloud might not directly reach Catalyst Center, or when unique authentication and API structures are involved, custom integration apps can be developed. This flexibility allows organizations to tailor their network management solutions, ensuring that Catalyst Center can effectively participate in broader IT ecosystems and policy-driven automation.
"We have the option like with ITSM to have an app that we publish on ServiceNow and adapters that run on Catalyst Center. This will create change request, create incidents. I can run a software upgrade. I can have CMDB inventory synchronization."
Ansible Collections Automate Catalyst Center Device Provisioning in Under a Minute
Cisco's Catalyst Center utilizes Ansible collections to streamline network configuration, packaging modules that enable precise control over various settings. A demonstration showcased how a playbook, employing the provision_workflow_manager module, can automatically provision a device to a specific site in less than a minute. This capability covers all configuration options available in Catalyst Center and simplifies complex tasks into single, manageable modules.
This automation solution, written in YAML, accelerates deployment and reduces manual effort, promoting efficiency in network operations. The Ansible collection, which has been available since December 2020 and supports the latest Catalyst Center versions, represents a significant step towards encouraging network engineers to adopt infrastructure as code methodologies for scalable and consistent management.
"Provisioning of the device completed successfully. I don't even know it took less than a minute for this."
Automated Software Upgrades with Ansible Workflows Enhance Network Scalability
Automating software upgrades using APIs, particularly through Ansible workflows, significantly enhances the scalability and scheduling ease of network management. These TAC-supported modules streamline the entire upgrade process, encompassing image updating, downloading from Cisco, distribution to devices, and activation. This eliminates the need for manual intervention during off-hours, allowing network engineers to schedule upgrades for optimal times without needing to be physically present or online.
The adoption of such automated workflows by Cisco engineering ensures high-quality, comprehensive modules that manage all necessary steps. This declarative approach, also seen in tools like Terraform, identifies non-compliant devices and applies desired software images, ensuring that only necessary updates are performed, thus improving efficiency and consistency across the network.
"For me I schedule my jobs. They are done I have the service on integration. They're scheduled for when they are scheduled it they're going to be completed. I can do the same with APIs."
▶ Watch this segment — 1:09:52
Catalyst Center Simplifies Software Image Management with Golden Image Assignment
Catalyst Center facilitates software image management by allowing users to import images from a local server, a critical feature for air-gapped environments. Once imported, administrators can assign a "golden image"—a recommended software version—to specific device types and network sites. This assignment automatically identifies devices that are non-compliant with the designated golden image, simplifying the process of ensuring software consistency across the infrastructure.
The AI Assistant further aids this process by providing information related to Software Image Management (SWIM) and guiding users through upgrade steps, including links to relevant documentation tailored to the specific Catalyst Center version. This streamlined approach eliminates manual steps, such as setting up TFTP servers or direct SSH downloads, enhancing efficiency and compliance management for network engineers.
"The compliance is the perfect tool because it tells us hey these devices are non-compliant from software perspective. This is the recommended image. Let's do a software distribution and activation."
▶ Watch this segment — 1:04:02
Infrastructure as Code Centralizes Management of Multiple Network Controllers
Managing multiple network controllers, such as Catalyst Center clusters, Meraki organizations, SD-WAN, and data center security systems, presents significant challenges for network administrators. Infrastructure as Code (IaC) offers a solution by automating workflows across the entire infrastructure, ensuring consistency, speed, and scalability. This approach allows engineers to define and manage network configurations programmatically, replacing manual, error-prone processes.
By embracing IaC, organizations can achieve consistent outcomes for various configurations, from applying firewall rules to provisioning new services across campus and branch networks. This method facilitates easier scaling to production environments and enables multiple team members to utilize standardized pipelines, overcoming the complexities of traditional, UI-driven network management across disparate systems.
"Infrastructure as code is really helping us to automate workflows and that's why we are showing you from a UI but also from APIs perspective. It helps us to automate workflows across the entire infrastructure."
Catalyst Center Reports Offer Flexible Data Generation and Automated Consumption
Catalyst Center provides robust reporting capabilities, featuring 66 predefined and flexible reports designed to generate large datasets that may be difficult to acquire efficiently through direct API calls. These reports can be configured to run daily or on a custom schedule, outputting data in various formats such as CSV, JSON, or Tableau, depending on consumption needs. For instance, reports can detail client sessions over the last 24 hours, including up to 51 different parameters.
Report generation can be automated and consumed via webhooks or APIs, allowing for integration with other applications, such as Python environments, for dashboard creation or data retention. Users receive a download link via email rather than the report itself, accommodating large file sizes. This system ensures that network administrators have access to comprehensive, timely data for analysis and operational decision-making without manual intervention.
"Reports allow us to run any of these 66 reports plus flexible reports and give you the data that you want in a format that is important for you."
Catalyst Center Python SDK Automates Device Onboarding with Robust API Support
The Catalyst Center Python SDK enables the automation of device onboarding, replicating manual UI steps through simple Python scripts. Supporting versions as far back as 1.2.10 and up to the latest 3.1.6.0, the SDK incorporates all Catalyst Center REST APIs, ensuring compatibility with new features. With an average of 50,000 monthly downloads, the SDK demonstrates significant adoption among network professionals seeking to automate their workflows.
This SDK allows for asynchronous execution of tasks, such as adding devices, by returning a task ID, which can then be monitored for completion. By abstracting complex API calls into user-friendly Python functions, the SDK empowers engineers to streamline device management, providing a consistent and scalable approach to network provisioning.
"All the functionality that you can do in the in the graphical user interface, you can also do it within Python code."
Network Engineers Encouraged to Adopt CI/CD Pipelines for Configuration Changes
Network engineers are strongly encouraged to adopt CI/CD pipelines for managing configuration changes, moving beyond manual or laptop-based scripting. This shift is presented as essential for achieving scalability and consistency in enterprise-level network solutions. While performing tasks via the UI or individual scripts might work for small-scale operations, pipelines offer a standardized and automated approach that is crucial for complex, large-scale environments.
Embracing APIs and CI/CD methodologies ensures that configuration changes are applied uniformly, reducing human error and improving operational efficiency. This foundational change in workflow allows for better version control, collaborative development, and reliable deployment, transforming network management into a more robust and predictable process.
"We hope by the end of this session you're going to appreciate it's much better to do it with APIs."
▶ Watch this segment — 1:27:39
DevNet Offers Free Catalyst Center Sandboxes and Extensive GitHub Resources for Automation
Cisco's DevNet provides free Catalyst Center sandboxes with full administrative access, allowing network engineers to test automation scripts and build SDA fabrics in a risk-free environment. These sandboxes, which integrate with Cisco Identity Services Engine (ICE), offer a hands-on platform for experimenting with infrastructure as code and generative AI solutions. Additionally, the EM Programmability GitHub organization hosts an extensive collection of code repositories, including Python SDK, Ansible modules, and Terraform providers, supporting a wide range of network automation use cases.
These resources underscore Cisco's commitment to promoting scalable, consistent, and secure network management through automation. By offering free access to lab environments and a wealth of open-source code, DevNet empowers network professionals to transition from manual GUI-based operations to programmatic control, fostering innovation and efficiency in network management.
"You build these playbooks once, right? And then you can just run them as many times as you want."
▶ Watch this segment — 1:37:32
Also mentioned in this video
- The packed agenda, starting with fundamental Catalyst Center workflows, noting… (2:59)
- Gabi details Catalyst Center's API capabilities, including 1,245 APIs in… (3:51)
- Catalyst Center's REST APIs, emphasizing continuous improvement in… (7:38)
- Gabi clarifies the difference between synchronous and asynchronous API calls,… (9:37)
- Catalyst Center APIs are rate-limited at the API level, not the platform level,… (11:42)
- Gabi details the 271 available events in Catalyst Center, many of which are… (14:50)
- Regarding the availability of reports and events, confirming their… (28:48)
- Various methods for device onboarding into Catalyst Center, including discovery… (30:28)
- Essential pre-onboarding steps, including creating a network hierarchy with… (32:23)
- During device discovery, Catalyst Center validates credentials, checks for… (35:02)
- Explaining device provisioning, which enables Catalyst Center to actively… (43:00)
- Various provisioning options (45:00)
- Adrian details CLI templates for device configurations, projects as groupings… (46:27)
- Template simulation for pre-testing configurations without applying them,… (49:22)
- Device provisioning pushes startup configuration, applies CLI templates, can… (52:05)
- Post-onboarding and provisioning steps, including software upgrades and CLI… (59:30)
- Catalyst Center's Software Image Management (SWIM) for managing device images,… (1:00:57)
- Gabi details the two-step image upgrade process (1:07:03)
- Applying additional configurations via CLI templates, highlighting how the AI… (1:17:37)
Summarised from Outshift by Cisco · 1:41:40. All credit belongs to the original creators. Streamed.News summarises publicly available video content.
