Anton Karneliuk

Author Archives: Anton Karneliuk

Automation 20. GNMI to Spreadsheet via Pandas or Intro to Data Analysis in Network Automation

Dear friend,

We all like spreadsheets, as this is an easy and intuitive way of representing data for us, humans. The possibility to analyze data with spreadsheets are endless, starting from simple aggregation till complicated pivots are endless. Even the people who say that they don’t like spreadsheets, use them so often. Network automation, for sure, is not an exclusion here. 



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Are Spreadsheets Valid In Automation Era?

They absolutely are. We haven’t seen so far any non-IT person, or even IT high-profile engineer or manager, which won’t like spreadsheets. To put it simple, spreadsheets can be an easy way to represent data collected from network devices; likewise, it can be a good way to provide some input, either as inventory or as configuration details for your network automation intent. So the questions becomes, provided you can read data from your spreadsheet in a programmatic way, what would you do with it?

The good news is that Continue reading

Automation 19. Enabling OCP SONiC To Be Managed Via GNMI With pyGNMI

Dear friend,

We hope you are doing great and had a nice time over the festive period to recharge your batteries towards the new year. We wish it to be successful, productive and prosperous. With this, let’s dive into the topic of today’s blog, which is network automation for OCP SONiC with GNMI and Python over pyGNMI.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Do I Need to Automate SONiC? How Can I Do It?

SONiC, which stands for Software for Open Networking in the Cloud, is a Network Operating System (NOS). SONiC’s main purpose is to run on the data center switches and provide simple and reliable connectivity between endpoints (Baremetal servers, virtual machines, containers, etc). As it is a piece of the network software, which runs on network hardware, it is for sure requires automation.

The good news is that it supports RESTCONF and GNMI with OpenConfig YANG modules, which we extensively cover in our Network Automation Trainings:

We offer the Continue reading

Kubernetes 004. Real-world Example of Deploying Kubernetes at Azure Kubernetes Services with Terrafom

Hello my friend,

In one of the past blogposts we have highlighted the journey for automation for engineers, who is willing to develop further outside of their core remit, such as networking, compute, or storage. In today’s blogpost we’ll provide some of examples how the different moving pieces are coming along together.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Do I Need to Learn Network Automation?

Yes, you do. Start today with our Network Automation Training programs:

We offer the following training programs for you:

During these trainings you will learn the following topics:

  • Success and failure strategies to build the automation tools.
  • Principles of software developments and the most useful and convenient tools.
  • Data encoding (free-text, XML, JSON, YAML, Protobuf).
  • Model-driven network automation with YANG, NETCONF, RESTCONF, GNMI.
  • Full configuration templating with Jinja2 based on the source Continue reading

Tips For Network Engineers Willing to Transit to Network Automation

Hello my friend,

quite often we are getting questions from various folks on LinkedIn or via private consultations about how start with network automation and, more importantly, whether to start with network automation at all. With this blogpost we’ll try to answer the most popular question.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

What to Do If You Already Want to Start?

The answer is simple: don’t wait. The amount of things you need to learn in order to succeed is huge and growing daily with new tools and platforms released. However, the good news is that we we provide you everything you need to start successfully in the role:

We offer the following training programs for you:

During these trainings you will learn the following topics:

  • Success and failure strategies to build the automation tools.
  • Principles of Continue reading

Automation 18. LibreNMS to NetBox Sync for Brownfield Automation Deployment

Dear friend,

Today we’ll look into a question, which is raising quite often on various meetups related to network automation or various threads that is the addition of any automation components (say, NetBox) to the existing networking managing suit, where you already have some inventory (say, LibreNMS). We’ll take a look into multiple approaches and, as usual, will share some code snippets how you can do that.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

What Is Integration Between Different Automation Tools?

Integrating different automation tools is typically the most labor-intensive work, as it requires understanding of the tools you are integrating, their internal hierarchy and APIs capabilities. That is not even saying that you shall be proficient with a variety of technologies, such as REST API or GRPC API, JSON/Protobuf/XML, Postman and many other tools. That’s a lot, and may be not easy to start with.

The good news is that all these things you can learn at out Network Automation Training programs:

Continue reading

Automation 17. Using Operational Commands via NETCONF at Nokia SR OS in Sequential Mode

Dear friend,

After a bit of break caused by preparation to Kubernetes exams (we will continue blogs about Kubernetes as well) we are getting back to network and network automation topics. One of the interesting things, which is gradually emerging these days, is the possibility to manage multiple aspects of network devices (not only configuration or collection of operational data), such us issuing ping/traceroute checks, copying file, etc in a model-drive way (i.e., NETCONF, RESTCONF, GNMI with YANG). Today we are going to look into such a topic.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Is that About Model-Driven Automation?

It is exactly that. NETCONF/YANG all the things, my friend! Usage of model-driven network automation significantly improves the stability and manageability of the network due to much simpler way to perform all the operations remotely. You don’t need to scrape and parse CLI anymore; instead, you interact with network devices via programmable API, what makes it possible to integrate them Continue reading

Kubernetes 003. Developing and Publishing Cloud-Native Application: Overview of Key Kubernetes Primitives

Hello my friend,

It took a bit since our previous blogpost about the setup of the highly-available Kubernetes cluster with multiple control plane and worker nodes. We aimed to write the blogpost about the upgrade, but we will park it for now for two reasons:

  • Despite this topic is needed for the exam, it is not very often performed within the cluster.
  • We want to provide a broader overview of Kubernetes from the perspective of building and using applications on top of it.

Therefore, we decided to walk you through the main components used to build and publish your application in a cloud native way on Kubernetes. Let’s dive into that.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Can Kubernetes Help to Automate Networks?

These days there are interesting projects emerging with Kubernetes acting as a management plane for network devices. Kubernetes in such projects has a role of the entity, which distributes configuration to the worker nodes, which are either proxies for Continue reading

Kubernetes 002. Scaling Out Cluster and Turning It in Highly Available One

Hello my friend,

As mentioned in the previous blogpost, we continue looking into Kubernetes. In the previous blogpost we have built a simple cluster consisting of one control plane node, which is the one ruling the cluster, and two worker nodes, which are the ones hosting the customers’ workloads. Today we will add a few more nodes, both workers and control plane, to the cluster to convert it into a high available one.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Are You Still Teaching Network Automation?

We absolutely are. One of the important things we have figured out is that automation is for sure the cross-platform technology. Therefore, whatever area you are coming from or heading two (networking, system administration, or cloud engineering), knowledge of automation frameworks and components, such as Ansible, Bash, Python, YAML, JSON, REST API, GRPC/GNMI is very beneficial and, in fact, is almost mandatory these days.

And in our Network Automation Trainings we have put it together in Continue reading

Kubernetes 001. Building Cluster on Ubuntu Linux with Docker and Calico in 2022

Hello my friend,

The last Black Friday (in 2021) I’ve purchased some Kubernetes trainings from Linux Foundation with quite a huge discount. I’ve purchased and then has been doing nothing in their regard for quite a bit until I’ve received a notification in the beginning of the September that I have only 3 months to use the training and pass exam or it will expire. Moreover, I have 2 exams to pass, so… All my blogs for a bit will be dedicated to Kubernetes only, as I don’t have time to look into anything else for the time being. I will be using these blogs as a brain dump for me during preparation for CKA (Certified Kubernetes Administrator) and CKAD (Certified Kubernetes Application Developer). Whether I will manage it in 2,5 months, the time will show. Let’s get started.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

How Is Kubernetes Related to Network Automation?

Kubernetes is one of the most popular platforms to build Continue reading

Practical OpenStack #006. What is the step-by-step deployment plan

Hello my friend,

Once we are sorted with the theoretical part of how we are to deploy OpenStack, we are moving to a practical bit. Today we are going to explain the steps we are going to take in order to deploy OpenStack, the supporting storage and how they are to be integrated.

Don’t Forget to Learn How to Automate All the Things

Video to the Topic

Need Help? Contract Us

If you need a trusted and experienced partner to automate your network and IT infrastructure, get in touch with us.

P.S.

If you have further questions or you need help with your networks, we are happy to assist you, just send us a message. Also don’t forget to share the article on your social media, if you like it.

BR,

Anton Karneliuk

Building High-Available Web Services: Open Source Load Balancing Based on HAProxy + FRR and Origin Web Server Based on NGINX Connected to Arista EVPN/VXLAN. Part 2 – Configuration and Validation.

Hello my friend,

In the previous blogpost we have started discussion about the Open Source Load Balancing solution, which leverages HAProxy and FRR, which is connected to the data centre fabric running EVPN/VXLAN on Arista EOS switches and serving content from NGINX-based origin servers. In that blogpost we covered the architectural guidelines and design principles. Today we will cover the configuration and the validation of the solution. 



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

What Is More Important: Network Technologies or Network Automation?

The truth is that both disciplines are equally important. Knowledge and skills in network technologies will allow you to build great connectivity solutions to empower businesses across the globe and spread its capabilities as nowadays, perhaps, 99% of all business operations leverage IT and network technologies either entirely or at least partially. In its turn, network automation allows to make operational activities (e.g., configuration, troubleshooting, analysis, etc) much more precise and predictable, decreasing the probability of outages or Continue reading

Building High-Available Web Services: Open Source Load Balancing Based on HAProxy + FRR and Origin Web Server Based on NGINX Connected to Arista EVPN/VXLAN. Part 1.

Hello my friend,

Recently we’ve been working on an interesting (at least for me) project, which is an MVP of the highly available infrastructure for web services. There are multiple approaches existing to create such a solution including “simply” putting everything in Kubernetes. However, in our case we are building a solution for a telco cloud, which is traditionally not the best candidate for a cloud native world. Moreover, putting it to Kubernetes will require to build a Kubernetes cluster first, which is completely separate magnitude of the problem. Originally we were planning to write this blogpost the last weekend, but it took us a little bit longer to put everything together properly. Let’s see, what we are to share with you.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

You Are Not Talking about Automation Today, Aren’t You?

Yes, today’s blogpost is dedicated to the network technologies (to a huge mix of different network and infrastructure technologies, to be honest). That’s why there Continue reading

pygnmi 15. Overview of nornir_pygnmi

Hello my friend,

Are you looking for building network automation at scale leveraging the future-proof model-driven network automation? Besides attending our zero-to-hero network automation training and network automation with nornir, we suggest you to take a look at nornir_pygnmi, the new plugin we have created for Nornir to simplify management of network devices with gNMI.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Is GNMI a Good Interface for Network Automation?

Yes, it is. GNMI is one of the most recent interfaces created for the management plane, which allows you to manage the network devices (i.e., retrieve configuration and operational data, modify configuration) and collect the streaming or event-driven telemetry. Sounds like one-size-fits-all, isn’t it? On top of that, GNMI supports also different transport channels (i.e., encrypted and non-encrypted), which makes it suitable both for lab testing and for production environment. You may feel that we are biased to gNMI, and you are right. Actually, that is a Continue reading

Tools 11. Running Prometheus Exporters on Dis-aggregated Data Center Switches (Cisco NX-OS, NVIDIA Cumulus Linux, Arista EOS)

Hello my friend,

Monitoring is a key component of any IT and network systems, as it gives us possibility to know what is going on in those systems. That is important at least for two reasons:

  • Observing actual status of the system counters
  • Analyze trends over the time period

Let’s see how we can attempt to normalize collection of such a data with Prometheus from dis-aggregated network switches empowered by Linux.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

What Is the Relation Between Monitoring and Automation?

The answer would be “closed-loop automation” which is also often sold as IBN (Intent-Based Networking). It works as the collected data is automatically compared against certain patterns or thresholds and in case of threshold violation, some automated action can be triggered. It can be as simple as a notification via a mail, a Slack channel or a Telegram chat; however, it can be also a more complicated, where some automated workflow is triggered to perform corrective configuration Continue reading

Automation 16. How to Prepare Cisco Nexus 9000 (Cisco NX-OS) to Be Automated with NETCONF and GNMI. Complete Guide.

Hello my friend,

When I booted up my laptop in the morning I was aiming to write something simple and easy: it is a summer and I don’t want to spend the whole Sunday on writing a blogpost. However, the topic, which is supposed to be an easy ride, appeared to be a bit more complicated. Well, it is even better then, as the official documentation to the topic is unfortunately vague.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Shall I use NETCONF or GNMI for automation?

The answer to this question, as usual, it depends. The first fact to think about is whether your platform supports any of these protocols. For example, the platform we will talk about today started supporting GNMI relateively recently and, to my surprise, it supports quite a legacy version of GNMI. At the same time, NETCONF, being an older technology compared to GNMI, are supported more robust. On the other hand, there are platforms, such as Nokia Continue reading

Automation 15. The Good, The Bad and the Ugly of Model-Driven Network Automation Featuring Cisco, Nokia, and OpenConfig YANG

Hello my friend,

All of us (definitely me, at least) are always thrilled hearing news from network vendors on their implementation of model-driven interfaces for network management. Having spent years automating network devices in a text-based paradigm (i.e., from CLI-based automation to full fledged configuration rendering with a replacement), I’m a firm believer that model-driven approach based on YANG modules and protocols such as GNMI, NETCONF, and RESTCONF, is a proper way to go. Recently we disclosed the development we are doing in terms of network topology visualization with DANT. And today we’d like to share lessons learned based on that experience.

We planned to write this blogpost for a few weeks if not months, but due to various reasons it was delayed. We are delighted to finally post it, so that you can get some useful ideas how you can build your own CI/CD pipeline with GitHub, probably the most popular platform for collaborative software development.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission  Continue reading

pygnmi 14. Using skip-verify for self-signed certificates

Hello my friend,

lab testing is one of the crucial parts of any new technology introduction. At the same time, we know that in labs we typically have a lower security requirements. For example, we use SSL certificates, but we don’t have proper PKI and, therefore, certificates are self-signed.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Is GNMI a Good Interface for Network Automation?

Yes, it is. GNMI is one of the most recent interfaces created for the management plane, which allows you to manage the network devices (i.e., retrieve configuration and operational data, modify configuration) and collect the streaming or event-driven telemetry. Sounds like one-size-fits-all, isn’t it? On top of that, GNMI supports also different transport channels (i.e., encrypted and non-encrypted), which makes it suitable both for lab testing and for production environment. You may feel that we are biased to gNMI, and you are right. Actually, that is a reason why we created pyGNMI library, Open Continue reading

Random. How to Choose Your Desktop Operating System for Network Automation Development

Hello my friend,

After writing quite long and complicated previous blogpost about CI/CD with GitHub, I need some therapy to write something light and chill. I decided to choose the setup of the working space for development and utilisation of the network automation and, in general, network design and operations. Though I don’t pretend to be absolutely objective and unbiased, as it is simply not possible, I intend to share some observations I did from my own experience and discussions with our network automation students, which I hope will be interesting for you.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Why Is It Important?

During our Zero-to-Hero Network Automation Trainings, and other trainings as well, we talk a lot about choice of tools to build automation solutions: they shall be fit for purpose and easy to use. However, in addition to that, you should also feel a fun, when you utilise them. It may sound odd, as we are Continue reading

Automation 14. Deep dive into Building CI/CD for Network Automation and Software Development with GitHub Actions

Hello my friend,

We planned to write this blogpost for a few weeks if not months, but due to various reasons it was delayed. We are delighted to finally post it, so that you can get some useful ideas how you can build your own CI/CD pipeline with GitHub, probably the most popular platform for collaborative software development.



1
2
3
4
5
No part of this blogpost could be reproduced, stored in a 

retrieval system, or transmitted in any form or by any 

means, electronic, mechanical or photocopying, recording, 

or otherwise, for commercial purposes without the 

prior permission of the author.

Regards

A lot of lessons about building the CI/CD pipelines and importance of unit testing and linting checks I learned from a colleague of mine, Leigh Anderson, whom I’m very grateful for that.

CI/CD Overview

CI/CD is an approach, which is very often used in software development, and discussed outside of that area. It stands for:

  • CI (Continuous Integration) is a process, where the created software (for sake of simplicity, any piece of code) is getting ready to be deployed.
  • CD (Continuous Deployment) is a process, where the software, which is ready for deployment, is actually deployed Continue reading
1 2 3 4 … 8