How Abstraction, Orchestration, and True Automation can make your Career Successful with Tim Fiola – ZNDP 090

ZNDP 090: How Abstraction, Orchestration and True Automation can make your career successful with Tim Fiola?

Zig Zsiga: [0:00] How abstraction orchestration and true automation can make your career successful with Tim Fiola.

Episode 90.

This is all about making you successful in your career with abstraction orchestration and what we call true automation what are they.

How can you learn them and immediately start leveraging them in your career right now welcome back my friends nerds geeks and ziglets out,

we have another episode of the zigbits network design podcast where zsiga bytes are faster than those gigabytes we strive to provide real-world context around technology

what's up everybody I hope everyone is doing great.

Zig zsiga here and welcome to episode 90 of the zigbits network design podcast once again my name is zsiga and I'm here to help you with network engineering

Network design and network architecture but today and today but today who cares we are going to do we're discussing abstraction orchestration.

[1:09] Automation helping me today is my good friend a frequent

zigbits podcast guest Tim Fiola Tim has been on a couple of the recent zigbits episodes episodes 83 and 89 and specifically speaking about episode 83.

The total that one's how to differentiate yourself as a network engineer it is currently.

Number one episode we've done over the last four years so if you haven't listened to it yet the community loves it and you will too hey if you want to listen to that episode go to zigbits dot text / 83.

Hey before we kick off the discussion with him today I have a couple of quick ish updates the first one the big one.

[1:55] I am officially writing a Cisco press book,

this has been legit this has been a career bucket list item and I am truly stoked about it now to be clear the book I am writing at this moment the title is.

Ççde so the Cisco certified design expert qualification exam specifically 400 - 2007 and it's the official certification guide so I'm writing

the official certification guide for the ççde qualification exam

it's just going to be truly awesome and I'm truly honored to write this book and I will keep everyone updated as I progress with the book and when it is completed.

It's really just an awesome experience so far it's going to be fun I don't know what I don't know yet so there's always that aspect but I'll be writing for the you know seeable future right so the second update right

because of my my combined workload between all the things I'm responsible for my day job zigbits Family Health

other responsibilities I had to order wrong Health was last house was supposed to be first I always mess that up Health was always supposed to be first and I always have it last.

[3:06] Because of all of these other responsibilities all these things going on I am moving the podcast I'm moving this show.

Back to a bi-weekly show just for a bit just for a bit and in the reason behind that is because I need,

to create some space sometime some some flexibility for me so I can get some of these large large extremely large projects in initiatives.

Like the course and the book and other things that I have on my massive kanban board in my office that I'm not making progress on.

So I need to I need to create the space and that way I can knock this stuff out.

And then I can get back to Bringing more content and more high-value content to you.

I hope all of this makes sense as always if you have any questions please feel free to email me

reach out to me directly Zig at zigbits dot Tech you can find me all the socials right Twitter LinkedIn whatever you can also reach out on Discord

right our Discord community and if you're not currently in our Discord Community you are missing out so go ahead and join its zigbits dot text / Discord quick and easy link and I will see you there.

Those are the updates that's it now for our conversation with Tim hey Tim how are you doing buddy how are you doing today.

Tim Fiola: [4:25] Zig I am doing awesome good to be here thank you.

Zig Zsiga: [4:29] Hey man hey this is great getting you on again I believe we recently did one and I'm going to I'm going to miss the number probably by heart I think it's actually episode 83 and we had a good show that,

maybe a month ago I'm going on time I'm going back in the future but but back in time I don't know how to do this we talked about how to differentiate yourself

as a network engineer and not be a commodity where most of us are probably Commodities these days doesn't work in years I think there was a great show so I'm glad to get you back here today to talk about

automation abstraction orchestration

kind of all buzzwords but again thanks for coming thanks for joining us why don't you tell us a little about yourself let everyone know who you are what you do and kind of why we're here today.

Tim Fiola: [5:14] Thanks Iggy bet my name is Tim Fiola I am a automation enthusiast.

I began my career as a network engineer by training then I became an Automation and coding enthusiasts by choice when I saw the power

that unleashes how it empowers the network engineer and how it empowers the companies they work for I'm also an avid tequila fan and,

man not even a recovering Lego holic anymore I'm just a full-on Lego holic again.

Falling off the wagon there I'm also have J & cie Service service provider number for 19 but I did let that expire several years ago and I also blog and,

now do podcast occasion.

Zig Zsiga: [6:00] Occasionally right yeah I mean right now I got you on once a month like that that's perfect right so.

And then I got to say like like the Lego stuff man like a hall like yeah you're totally you're not a recovering alcoholic anymore it's awesome I love Legos,

I can't get enough of them and my son loves them as well so that makes it even better for me because it's something we could do together you know.

Tim Fiola: [6:22] Yeah it's a good way it's a great way to spend time with your kids and bond with them for sure.

Zig Zsiga: [6:26] Yeah man so yeah we're going to talk about automation today I think and abstraction and kind of all of those things I would say it's probably why it's important and why we care what makes it.

Drive business success or how does it make a business successful and so I think that's kind of the presence presence that's not the right word preface previous summer,

yeah thank you premise thank you I'm like what's the right word that the premise of the show so let's start with abstraction if you want.

Tim Fiola: [6:57] Sure yeah abstraction orchestration and automation like you pointed out earlier are very they're very loaded terms they have different meanings in different contexts and.

Some people use them in different ways so as a network engineer who's trying to learn this and what

how these impact me my employer my culture what I'd like to do is to use your words is to demystify these terms.

Zig Zsiga: [7:27] Nice I like.

Tim Fiola: [7:27] In today's ship oh yeah I see what I did there so and before I before we really start talking about this these are all my opinions,

not those of any past or present employers and they're very open to discussion or any hate mail anyone cares to send me but.

Zig Zsiga: [7:46] No One's Gonna send you.

Tim Fiola: [7:48] Said you know you know yeah I might spike a bit Yeah and your daily yeah snow.

You're getting into it one common one among Network Engineers is what I call abstraction and we're going to discuss what that is and.

Turns out if you're doing any coding you're probably using abstraction at a minimum so abstraction is taken a it's an exercise in,

replacement so if you take a relatively simple task,

and scripted or make like something like an ansible Playbook to do it it's a replacement exercise because if I make a.

Python script that configures a VLAN on 50 switches now I've replaced directly touching 50 devices with a script that touches 50 devices I manage the script the script.

Interacts with the devices so it's an exercise in abstraction that is.

[8:57] That is that is the building block though of orchestration and Automation and we'll discover well,

discuss a bit about that and automation I'm sorry abstraction the scripting part of it tends to be implemented at the Grassroots level it typically starts with a network engineer to getting together saying hey we do this all the time,

we hate doing this all the time let's make a script and not do it all the time and move on to higher value stuff.

Zig Zsiga: [9:25] So I just say I just say I'm late.

Tim Fiola: [9:26] He describes abstraction.

Zig Zsiga: [9:27] I just got to say I'm just a lazy person I don't want to do the work I want to do some I want something that I can hit a button and it led does it for me and I know it's going to work like that's where I'm at.

Tim Fiola: [9:37] Yeah lazy whatever your motivation for doing it it adds value so.

Man just just run with it if that's what you're thinking so that's abstraction and it's hard it's encapsulating,

work in a script or Playbook perhaps and making that work repeatable by just pressing the button multiple times.

[10:00] So now we're going to talk about a little bit about excuse me orchestration so while abstraction

will involve a task that'll affect a single device at a time or a single type of device like

as in I'm provisioning VLAN on 50 switches one at a time the script goes out and Provisions each that veal on on each one of the 50 switches individually,

so orchestration starts to build on that a bit orchestration will start to involve coordinating tasks among multiple devices.

[10:39] For example let's say if you have a really complex topology and you're doing let's say a router upgrade.

So that router upgrade might involve a lot of steps with the network elements around it for instance you would probably run pre checks on the device

then you would Implement igp overload on the device then you might depend on your topology you might have to,

shut bgp sessions down on all the neighbors all the bgp neighbors around it so now you're not just touching one device you're starting to touch and,

orchestrate activity amongst several devices in one in one workflow

so then you know the device might get upgraded to continue the flow you'd run post checks on the device you would normalize your igp on the device that's being upgraded then you would normalize your bgp

on the on the neighbors of the device that's under upgrade.

Zig Zsiga: [11:41] So we just we just went through a whole migration right there in a nutshell those are steps right there to migrate that that's awesome,

and I just want to make it clear so we're comparing the abstraction orchestration abstractions kind of a single task on multiple devices Maybe,

is that right and in orchestrations multiple tasks on multiple devices part of a workflow.

Tim Fiola: [12:06] Yeah involves start more of a workflow involving most likely in multiple devices but if you look at each one of those individual steps we just walked through each one of those can be done with an encapsulated.

Work item for instance running pre checks on the device to be upgraded that is a script implementing igp overload that's a separate script the.

Excuse me the bgp session on the neighbors that's a stat is a distinct script so it goes on and on we're what we're doing right now is we're encapsulating.

Excuse me we're orchestrating different tasks among multiple devices.

Zig Zsiga: [12:50] Okay that makes sense that makes perfect sense I'm tracking thank you.

Tim Fiola: [12:55] All right so so.

Orchestration is building on the abstraction you're taking the abstracted tasks and orchestrating them amongst,

you know in the workflow amongst what is likely multiple devices so why isn't this automation this isn't what I call true automation because.

To initiate this someone still has to press a button.

You're still initiated manually saying now is the time let's upgrade this device let's do the work.

Zig Zsiga: [13:28] Don't press the red button don't press the red.

Tim Fiola: [13:31] Right.

Zig Zsiga: [13:33] That's what's going through my head.

Tim Fiola: [13:34] Press button yeah right you don't want to do it.

Before its time right so another thing that orchestration will require as you look at this workflow is more feedback.

About the status of each task each step must return specific feedback so the next step nose.

That it's okay to execute,

or to gracefully exit if it's some there's something unexpected and another interesting thing that is its own show,

and is multiple podcast is it starts to require a good source of Truth for the data in orchestration you're using data and State.

Amongst about many Network elements so you need a good source of Truth to draw state from and to compare on the network elements you're touching.

My current employer is Network to code and we have several blog posts and podcasts that cover the source of Truth it is truly it becomes the heart of any automated infrastructure Source the truth.

Zig Zsiga: [14:47] Yeah so even like architecture speaking like source of Truth is such a critical component like even we start talking about like zero trust architecture and future State architectures they all require a source of Truth like that single point that has the truth of the environment in real time

and that way it can make a real good valid decisions on the data because it's analyzing that data,

real-time analytics and then making suggestions and or making decisions for you so it's interesting that you bring that up right so we're kind of merging multiple

I would say subfields because there's the the programming side the automation orchestration abstraction side and there's a single there's a source of truth but then if you start talking about the architecture side where we're talking about like

again zero trust I mentioned but like software-defined networking and controller base items

that source of truth it's kind of overlapping and all those places I think that might be something that we want to talk about in the future for sure.

Tim Fiola: [15:45] Yeah I'm happy to talk about it is truly a it is truly its own.

It's own thing to discuss further sometime it's really interesting and so with orchestration,

I'm depending on you know and these are

column checkpoints on a spectrum abstraction orchestration automation these are not hard checkpoints what we're talking about here is just things to look for characteristics to look for to understand

is this orchestration is it true automation or is it just really sophisticated abstraction,

this what I'd like the network engineers in the audience to understand is how the Spectrum.

Develops from abstraction and true automation there's all these points in between and these are these are some of the characteristics look for on that Spectrum when you're trying to figure out maybe where you're at.

A couple more things about orchestration before we moved on to what I call true automation.

[16:50] Organizations cooperating becomes a bigger deal if you start orchestrating with more.

More elements and more elements that aren't necessarily controlled by your group you need to start coordinating amongst groups.

To get them to agree to the process and the source of Truth and whatnot the other part of it is.

You might depend on where you're at you might be investing in a platform to use for orchestration.

So maybe you're not just running scripts you're running scripts that are executed by some controller.

That is taking feedback from the process and knowing when to initiate the next step so orchestration and I'll just say group cooperation amongst groups at at least a low-level start to become.

Involved in the process.

Zig Zsiga: [17:44] Yeah so so I think there's a huge call out there that you mentioned I think it's important that like you can't do these things in a silo right like can't you can't just do it on your own little organ little Department Ruben your sub like,

team like the network team if you want to go I was I had a show I did recently with another colleague friend of mine and we talked about Shadow,

automation

like Shadow automation like it's a real thing man like it's a like we talked about Shadow it and I'm going on making up this term right so Shadow it's--but a thing for years people would different Cloud because they you know their it Department couldn't couldn't support their needs well

if you think about that

shadow shadow automations the same thing it's hey your automation Dev team application team Our IT team can't support your needs so now you got these pockets of people doing coding in Automation and

abstraction orchestration like we're talking about here and that's a real thing like it's a real true true thing,

to bring our own issues into all of this I think it's extremely important to identify that too but you said a number of great things so I think this is this is going to be a great conversation that's I'm happy I'm excited I'm pumped up so let's keep going man.

Tim Fiola: [19:00] Yeah so now we get to it starts to become like the Crown Jewel and that's why.

What I'll call Define as closed-loop automation so closed-loop automation is similar to orchestration but the process that you're kicking off the workflow,

it's executed in response to an event not to event not without scuse me.

In response to an event not a human intervention so you know you look at Skynet as the ultimate you know closed-loop automation.

Zig Zsiga: [19:37] As Skynet.

Tim Fiola: [19:40] Have to use that as an example I mean well you know I don't agree with what Skynet did but that was a true example of closed-loop automation.

It identified something and did just awful stuff but.

It also you know brought us the Terminator so closely Pop animation just called Skynet Skynet is the crown jewel of closed-loop automation so as long as the event is not sewn explicitly,

you know manually initiating the process it's automation because it's a reaction to some type of event,

so automation again it requires even more detailed feedback so something.

Your automation platform in this example knows when to execute for example if a bgp session on a pier goes down.

Troubleshooting a bgp session humans when they're doing it

they all do the first five or six things maybe in a slightly different order but you do five or six things like right off the bat you check the config you check any rate you know static routes you might have configured you check the interface everyone does the same like four five.

[20:55] Yeah boilerplate stuff so you know shut no shut interfaces Okay so.

When I when I think of an example of true closed-loop automation let's say that bgp session goes down an automated,

process sees that session go down and let's say it just automatically creates a ticket with a ticketing system it,

pulls certain details from the device logs them in the ticket and then kicks off basic a basic basic troubleshooting script where you're checking the bgp status checking interfaces doing a shot no shot.

All that all that same boilerplate stuff that humans do now if after all that the bgp session comes up awesome note that.

Close the ticket and just hey,

you have a close ticket and a human never touched it the system just fixed it by itself if that session does not come up after the boilerplate stuff then you can kick it up to a human being to take a look at and even though humans involved.

It is still true Automation and we can talk a bit about that did you have a question.

Zig Zsiga: [22:06] So so I have a number of thoughts so.

Um first off Skynet like this is telling me robots taking over the world and we're done for right that's where I'm going first right because we said Skynet.

But that that's that's one spectrum of the equation right wouldn't potential reality with Automation in the long run is the hey robots could take over the world yeah,

but like like I think there's another aspect here.

[22:37] Some of the steps that we think our automation are pre-scripted tasks like like you mentioned like troubleshooting a beach be neighbor ship right there are there are subsets there is a subset of tasks that we all do,

when The Neighbor ship goes down and so we can write those down and say okay mr. automation or mr. you know system or control or whatever it is here are those steps

right and then that controller can intuitively check those steps for us and validate hey these are good to go

one of the key things I think is imperative to this entire closed-loop automation cycle and it sounds like I might be on this point I might be maybe I'm wrong and you can tell me to him

is it you have to have that feedback mechanism whatever you're calling it The automation system needs to know what's happening in real time or as close to real time as possible,

so that neighbor ship goes down how fast is that automation system know that memberships down so it can react accordingly and if we don't have that feedback mechanism,

then this automation system is not going to be able to do these things in real time.

Tim Fiola: [23:38] Yep there is no automation without without feedback that is certain.

Zig Zsiga: [23:41] Hey there we go that's awesome yeah so okay that was my big my big heart bright Sky Net and then there has to be like a feedback mechanism maybe that was the issue with Skynet there was no feedback mechanism I don't know maybe.

Tim Fiola: [23:54] That's a whole nother episode I'd love to talk about Skynet.

Zig Zsiga: [23:58] Let's do that let's do that in the future cheers.

Tim Fiola: [24:01] Let's schedule a troubleshoot Skynet yeah how could it have really taken over the world yeah yeah exactly.

Zig Zsiga: [24:06] Fast going to take over the world.

Tim Fiola: [24:08] So continuing our workflow here let's say the bgp session doesn't come up after the automated,

boilerplate troubleshooting well then you you kick the ticket up to human being,

and say I can't you know the system says I can't fix this fix it so a human being does what they do they can get involved and actually fix it if it takes more than the system can

it can deal with and then

then then close the ticket themselves and then again the the system will see that clothes ticket and then do something like hey.

Let me let me notify the customer that we open the ticket for them and now it's fixing me close a ticket for them and depending on how long this whole thing took.

Customer might not even know that they were that they were down for a bit so it.

That is an example of one workflow that is automated.

Zig Zsiga: [25:05] So I have a question right I'm going to jump in right questions is always okay so let's say

the automation system went to these Ford Ford steps whatever those steps are for whatever problem right it's pre-scripted there's four steps hey I went to these four steps it's still not up

whatever the problem is so down I ticked I push this up to the human person right that human interactive person they do some additional work and they figure out the issue they solved the issue

so is there any way to backfill that process with the additional information

the additional troubleshooting steps that human person took so that in the future when that same issue happens well now instead of four steps the automation / that automation system now can do five six seven or eight steps and then maybe.

Automatically solve the issue now and now those human people don't even have to be a part of it anymore.

Tim Fiola: [25:57] Yeah in a if you are in a sophisticated environment that can do what we just discussed hopefully you.

You lean into it and do exactly what you just described Zig is is look at why how many times did that system have to kick this kind of ticket up to a person how did they end up fixing it.

Is that something we can have the system do automatically or check for in the future.

Zig Zsiga: [26:23] That's cool so that you can just kind of update your scripts update your automation once you have a kind of a good metric system or good

view of hey this things happen a hundred times or 15 times whatever the metric value is hey what did we do each of those times when we did the exact same thing right we did these five other steps for for these two other steps well let's go ahead and add those to the script here's the

here's the trouble issue.

Situation some something you get to know the state right that's all about State tracking you had to know when the when it's an outage State and what specific outage it is what's the actual variables and what's that state,

okay that's the state we validated it and then now we have to actually fix that state here are the steps to fix it and then we three validated that it's fixed and then we can close the ticket man this is

this is true automation that's what this is this is like.

Tim Fiola: [27:14] Yeah and.

And one one more way to think about it now that we've gone through Automation and orchestration is earlier we mentioned source of Truth and talked about that,

the delta in your network between the source of Truth which reflects the desire state of the network and the operational steady state of the network there's going to be a Delta,

between those.

That's your work for the day or if you're in automation that's your automation systems work is bringing the network.

Operational State back in line with the desired state of the network and the desired state of the network is.

Reflected in your source of Truth.

And that is why source of Truth starts to become a huge deal because you're taking automated actions to align the two.

Zig Zsiga: [28:11] Yes that goes into that sources truth caught topic and conversation I be concerned about,

security of my source of truth right like what happens if my sources truth is compromised like what's what's the impact will the impacts very large right we don't need to go down the weeds but if my source of Truth is not actually truthful well there's a problem

right and maybe I'm opening up that I don't know onion I'm peeling back the onion or whatever multiple layers but I think that that's a good point to make is that

we have the source of Truth system we have to make sure that its Integrity is checked and validated all every step of the way.

Tim Fiola: [28:49] That is why source of Truth is its own I mean it's really it becomes the heart of an automated platform it's just it's everything.

Zig Zsiga: [29:01] That's really cool okay yeah so we're going to I'm definitely gonna tee that up man just so you know Tim you're we're gonna we're gonna draft up another show on source of truth because I think that's that's hyper critical that's my new buzzword the last couple months is hypercritical but.

Tim Fiola: [29:15] It certainly is and I'll see if I can bring some friends on who who know a lot about it because there's there's so much there it is it is it is I think becoming its own area of expertise in the industry.

Zig Zsiga: [29:29] Yeah for sure so we'll definitely keep that up I think it would be beneficial to all the listeners all those Network Engineers are trying to become an automation specialist and abstraction specialist and orchestration Masters all right so I cut you off so I think we were at the

obstacles to move from abstraction kind of to orchestration and automation.

Tim Fiola: [29:50] Yeah so we're what we're going to do is we're going to look at each,

at the different factors that that will affect how far your you and your organization can go on the the spectrum of this automation Spectrum are you going to stop it.

Are you going to stop at abstraction can you make it to orchestration can you get to True automation we're going to we're going to talk about what will start to impede you as you progress on this automation Spectrum here.

Zig Zsiga: [30:17] Yeah that's good that's good so was that some of that stuff is going to be really about probably the business to like in how the business is structured to really Embrace automation or not embrace.

Tim Fiola: [30:30] Yeah and you know it a lot of it will start with

with culture and this is really at the Grassroots level getting getting your your line Engineers to really Buy in and there's there's a few things in a culture that can.

Work against.

Starting on this journey we're talking about the first one is what I call the cult of the CLI no one can work the CLI like I can I can you know look how fast I can type their that is real.

And it's real and to a point I get it.

But there's a larger Force at play here that it's going to be hard to to to Really swim against the tide.

With that you know another another just cultural thing is just fear you know automation is BS it.

We'll mess stuff up it will you know I'll be automating failure if I make a mistake well you know no doubt.

Everyone makes mistakes and automation.

Zig Zsiga: [31:34] Yeah we all do we're human.

Tim Fiola: [31:36] Can amplify it can amplify it on balance though as a business and as an engineer you have to look at what am I going to what's going to do the most good,

in the longer term it's starting this this journey and moving along this spectrum.

As far as you can that is what it will ultimately add the most value so getting this automation is BS fear and.

Just animosity it that does maybe nut does need to be addressed and then along the lines of what we just talked about another one is just I don't trust machines and that's Skynet syndrome.

And Skynet you know is probably doing.

As much as anyone to make people not trust machines but on balance you're going to you're going to need to.

But it's not just machines it's people putting in the code and if it's done smartly by knowledgeable people your odds of success go way up.

Zig Zsiga: [32:34] Well I mean if you,

all dacian built-in right I think that helps a lot if you don't have validation that's probably the issue but if you have validation in there to make sure your codes accurate and then you're checking the state before we make a change all those devices and then you're making the change.

And you have an expected success state that you're going to kind of great everything with it's almost like I'm

I think this outcome this is what I should see and if I don't see it will fall back

type of model right and that's that's not what we can do as a human would do that normally but we do it for one device at a time and we don't scale out we can't scale out the way that automation can

so so a couple things right you mentioned the cult of CLI,

which I totally resonate with because I like to see a lie I am a CLI jockey I love the CLI and yeah I'm pretty fast at typing and configuring things alive man like even if I don't touch the sea alive for months I'm like back on it going away and type it away

I do have this this internal thought that no one's going to know what a CLI is in the future.

[33:41] Like if we go ten years from now they're still be a steel I and they'll be that,

commodity that that with the differentiation research so we talked about last time right that that 1% or less

add a note the CLI is and they use it they know how to use it

but the rest of the commodity Engineers those the rest of the network engineers and designers out there all they're going to know is the GUI and how to hit buttons I can click the buttons on the gooey and then rely on automation,

so I do have that concern and then yeah I don't trust machines right like I see that and being fully serious.

[34:18] Excuse me being fully serious I see that more often in the government world,

then anywhere else like we're not going to trust a machine nope there's going to be user interaction user.

Some sort of human is going to be in the process no matter what. Hundred.

I think that's a huge cultural issue governance issue and I don't know if that's ever actually going to change in the government world.

Tim Fiola: [34:44] That is an interesting perspective how culturally it's so ingrained yeah time will tell with that.

Zig Zsiga: [34:51] Yeah yeah I mean it's like about like

we talked about business success right that's what we're focusing on but like in the government rules about Mission success and if we lose control over something and we're letting a machine or a computer or whatever you want to call it do that task well now what happens if they do something wrong.

Right there's there's an implication that there's an impact what could happen right and that that I think that's probably the biggest concern that most of the leaders in the government world have so.

Tim Fiola: [35:18] That's that's really good perspective I don't have any visibility as to how the government really operates when you look under the hood like you do so that's really interesting perspective.

Zig Zsiga: [35:28] Yeah yeah so I think it's a great I think there's a great culture items so I think I kind of cut you off but I wanted to make sure I said a couple things so let's I'll turn it back to you man go ahead Tim.

Tim Fiola: [35:40] Sure so the you know another we touched on it a little bit earlier but another obstacle to automation is you know moving.

Further to the right on that spectrum is cooperation between groups,

just a simple device that you control or simple group advices that you control you know abstraction and maybe simple orchestration you can move there without,

perhaps without any other cooperation from other groups

but as you start to orchestrate more actions and more we're going to start talking about workflows here workflows that cross-organizational lines those organizations have to.

Start cooperating with each other they have to agree on workflow they have to agree on source of truth there's a there's a lot of coordination that has to happen.

[36:33] To allow you to get out of abstraction and then to you know kind of simple orchestration or maybe Advanced orchestration if the group's don't cooperate doesn't work that's just that's just it.

So because of this you start to see.

Orchestration and for sure automation is management up to.

The the sea level needs to get involved to 1,

coordinate work between the groups to make them cooperate to get there the groups to buy in and to to properly incent the groups to work together now,

if the incentives have to be properly spelled out you know for instance the objective should State explicitly,

you need to support this specific automation effort,

did you get the objective shouldn't say something like increase productivity by 20 or 30% with without additional headcount what happens there is.

You know management May read that as I need to make my people work twenty to thirty percent hard.

Zig Zsiga: [37:46] Yeah yeah make them work more and harder and faster.

Tim Fiola: [37:50] Yeah yeah and I use the example there was an episode of The Simpsons where Homer was.

He was working for a super villain I think his name is Scorpio

and Homer was in charge of the nuclear power plant in Scorpios townie controlled and Homer was sitting there in the control room with his engineers and the engineers were busy pressing buttons and beeping and all this stuff was happening and

Homer says hey are you guys working and they said yes and he said her can you work harder and they said oh yeah and then they just started pressing the buttons quicker that,

and he was like wow what a great manager you know wow I'm so good at managing that is not now clear that's not how the real world world works so,

the objectives from the sea level have to be specific to the automation orchestration platform and effort at hand if not you can go off the rails,

like like Homer was starting to do there.

Zig Zsiga: [38:46] It sounds like that to be connected right then to be they only have to understand what's going on they'd be part of the process they have to know automation tool would also automation workflow they're working on and that way they can talk the talk be part of it

understand the implicit implications across the board so I think I think there's some learning that has to occur occur at that sea level as well it's not as easy to say it's just it's just Network engineers

they have to learn this stuff that the sea level staff also have to understand automation

not the ones and zeros or the syntax or the tools per se

but the implications and how they could manage the teams and the workflow is that they're managing from a people perspective.

Tim Fiola: [39:27] That's and that's that is so important because we're now we're talking about.

Workflow orchestration and and and workflow automation that,

if you think about it what the system's you're touching you're probably touching CRM you're touching a ticketing system you're touching the network you're touching probably some databases you're touching the source of Truth these are all.

Systems that get it that.

That interact with the automation platform via CLI the net I mean sorry via API

and the network is just another the network is just another

piece of that workflow the network might not even be the central piece it you know this the central job might be something way different but the network is a is a part of that workflow.

And so automating your network is important if you want to get true end and strategic company-wide workflow automation.

Zig Zsiga: [40:27] That's awesome that's awesome so I think I think at this point I want to get into like if we can identify some values.

Impact values or or.

I want to I want to try to put it in business success terms as much as possible and I would like to try to compare I don't know I don't know if we actually outline this so I'm putting on the spot to him so I apologize that this is totally out there but I think we can talk about it

I'd like to get an idea I think it's a rough order of magnitude from from abstraction

to orchestration to True automation so I think if we could just put an example of the value of abstraction first and then we'll go to Automation orchestration and then through automation from a business perspective.

Tim Fiola: [41:12] Yeah so that I yeah I love that let's let's talk about that abstraction is where we start.

Typically from the Grassroots up and if you stop at just abstraction for whatever reason if your organization just can't go further that's not to say you're not.

Adding it tremendous value and your organization is not going to be successful because what abstraction does is it.

It takes the task you do a thousand times and instead of doing a thousand times you make a script that can do it a thousand times.

[41:47] And in doing so you increase you can have a group of say four Engineers start to do the work of.

Six or seven perhaps,

if you can automate your OS upgrades if you can automate add in a VLAN or reconfiguring a segment of the network or you know we talked about,

couple months ago the an RSVP mesh putting a router into the RSVP mesh if you can automate these things the amount of throughput in your that the engineers themselves can do goes way up.

And so just with that the network engineer is taking the train saying I am now doing the work of,

you know two or three Engineers or increasing my productivity by ninety a hundred hundred fifty percent and most likely it'll be pretty easy to make that argument hey instead of doing this thing a thousand times at an hour each I wrote,

script in like four hours and now I just press a button tremendous payoff at the abstraction layer tremendous.

Zig Zsiga: [42:48] Yeah it's just a heat it's huge return on investment return on your time right and then now you can focus on other things while it's going on so that's cool so that's abstraction about orchestration.

Tim Fiola: [42:59] So orchestration cooperation between groups is always a great thing orchestration can start to.

And Shrine that cooperation in code so that it's harder to stop doing once,

you know say director level management perhaps if there's you know two or three groups coordinating with each other to get orchestration was the director levels realize what's going on.

And by the way the director level is not going to not know it's happening typically the director levels get involved too you know perhaps like the platform kind of maybe Corral the the

the group Center cooperate in doing the education piece like we talked about.

Now you're starting to to get efficiency with not just tasks and simple workflows your workflows get a bit more elaborate that you can that you can increase.

That you can how do I set your workflow start to get more.

[44:03] Duct of in that you are increasing your throughput your decrease in human error you're starting to impact,

the company's cash flow because you're getting stuff done quicker you're reducing quote to cash interval which is a cash flow bump so.

Orchestration you're starting to.

Get closer to corporate strategic objectives and when I say strategic objectives it's it's a high-level objective defined by the sea level,

folks that say this is where we're going to invest time because this is how we can differentiate yourself from the car competitors this gives us an advantage not easy for our competitors to replicate or,

everyone's moving here in order to stay competitive we have to have this capability automation is is one expression of that orchestration is where that starts to become noticeable.

Operation between groups more elaborate workflows.

Zig Zsiga: [45:04] So for orchestration I mean it's I know you've said it a couple of times right you have to collaborate which we need groups so here's a question I think that I resonate with and I'm wondering if other people that are listening to resonate with this too

so in all of my environments for the most part I'm going to some I'm going to generalize but I would say 95% of the

companies I worked with there was siloed groups and none of the groups work together none of the group's talk to each other it was always a pain in the butt,

to do anything something like I'm on the network team right

and maybe I have a good relationship with a security team because I don't know they came from the networking team so we mesh right but then the VMware team and the server team and the storage team and the application team the development team all those other teams they hate me

because I'm always the issue right I'm always the problem and they never helped me in so I'm just trying to think how do we.

Maybe you know this answer I don't know like maybe in your experience in that might be helpful here like how do we go from,

being that siloed traditional networking team where no one wants to help us and no one wants to partner with us how do we copy come that partner and say hey we're going to get some Automation and orchestration.

We want to work together and make your life easier how do we do that.

Tim Fiola: [46:26] The answer to that is that's why the c-level folks make the big bucks.

Zig Zsiga: [46:32] Okay yeah.

Tim Fiola: [46:33] That is absolutely their job to figure that out and in doing so they they,

they're advancing their strategic objectives which make the company more competitive or give it a differential that its competitors find hard to match.

Zig Zsiga: [46:48] That's great that's a great answer because I was over here thinking that we had to own that and try to find figure out ourselves right but but you put it to a leadership issue it's totally a leadership issue that,

leader has to own it and say okay hey these groups have to work together we have to we're going to do orchestration and now it's not one

individual group like networking group server group so on and so forth It's one Enterprise architecture group with automation every layer

all working together and truly putting in orchestrated solution together.

Tim Fiola: [47:24] That is correct that that's right.

Zig Zsiga: [47:26] That's a whole different world man it's whole different world.

Tim Fiola: [47:29] Yeah yeah it's there's a lot to think about and so if as a network engineer you can get to abstraction but for reasons outside of your control you know

simple orchestration or is not possible because the groups won't cooperate.

You'll just have to rest easy at night knowing that you're doing the work of,

maybe two or three engineers and you're adding a tremendous amount of value and when you are looking at your next role you can start to ask the the people you're interviewing with about the culture.

And how automated are you and hopefully this this podcast is giving people some reference points to think back on saying yeah it sounds like they're you know.

They've they've made it to orchestration and you know maybe they're on the path to automation maybe I'll know that I'll go to this new role and I'll still just I'll still be in abstraction,

I'll still just leverage abstraction but again if you are getting paid more for adding that additional value you can't say that's a bad career move.

Zig Zsiga: [48:37] No no of course not a hundred percent you're adding value and if they're realizing adding value you're going to get paid more that's a quick you're making a difference and you have a differentiator there for your role get going.

Tim Fiola: [48:51] The last part of that is if you have a vision if you see if you can say to make a case to your prospective employer that you're interviewing with look I've done abstraction I've

I can show you data that shows how much I've improved.

The productivity of the group I'm in or my personal productivity and I would like to.

Take up an orchestration role where I would like to work in an organization I would like to lead the charge to orchestration that depending on the culture of the company are going to that might turn some heads in a big way.

And get you a you know more high-level job than you thought you were interviewing for but if you have the vision and you have the capability and you show that.

Good things can happen without you planning on that and that's always nice.

Zig Zsiga: [49:44] Yeah oh yeah hundred percent I mean someone's going to buy into what you're saying if you can actually,

bleed that orchestration drive I mean that's a huge drive that you're leaving that's a huge initiative I mean that's not full on automation right but that's that's I would say more than halfway there and I think.

Your benchmarks right there were talking about our great

great Milestones or benchmarks is still good word to but milestones for people to think about hey this company is doing abstraction today it maybe they're not maybe not even doing anything

right let's there nowhere and then you okay well we're going to start doing abstraction let's abstract all these different tasks let's start small right doesn't have to be complicated,

hey let's figure out a way to abstract adding vlans to you know switches and trunk interfaces and you know we had to create a new VLAN well where does it need to propagate to okay these five switches okay how do we automate that

abstract that and create the script just to do that one single task

but now hey we had a thousand vlans well now we can do that right we have that one task that one script that can add a thousand vlans now we just got to call it a thousand times.

Um

But then if you have a workflow I mean that's the orchestration element right that's the way that's the orchestration outline right there and I had the I'm envisioning for workflows orchestration specifically like

you have this you have the network part of it the servers the VMware the like the.

[51:09] Um not the source of truth I notice what we talked about earlier but kind of like the inventory system and the ticketing system like I think itsm like it Asset Management Service management,

but then you're also going to have like like help desk part of that I mean it's a whole it's really truly a whole.

I don't want to say architecture but it's a whole workflow from beginning to end and it touches everything within your environment everything least that's how I Envision it right now I don't know if you think that's the same roof.

Tim Fiola: [51:41] No that's once you start looking at orchestration you're starting to deal with workflows.

And cross cross cross Silo workflows and then automation will start to go with you know you'll start to see more and and Company process.

Automation automation of entire end and workflows orchestration you might you know you might be automating a segment of the workflow.

But not the entire workflow.

Zig Zsiga: [52:13] Look at it I look at it like abstraction specifically probably to your field of view or your your area of responsibility like abstraction for Network Engineers going to be in the network space something that you can abstract and automate for yourself but it's not going to be

at the orchestration level in the most cases but that's the same

I think definition for like a server admin like there are scripts that you can use as a server admin to do certain things and then as a VMware admin same thing right there storage I've been everyone has their own abstraction tools,

to help them do their job a little more effectively efficiently

but then when we start talking about orchestration we're just pulling all that together right we're just pulling everything together now and it's not that VMware Eben doing things on his own it's hey here's the network engineer the you network happened the VMware admin the server admin everyone get together

yeah we're no longer on different teams we're working together and we're going to make it orchestration workflow so maybe we have to do

build out a new application write a new application that requires vm's and storage and networking and load balancers and I'm just talking at the top

my mind here right and then also oh hey there's an application development team that's building the application so that they need a QA environment they need a Dev environment they need a production environment of staging environment now you can see how this all

ties together and I think that right there if you have an organization to get actually embrace parts of that or all of that that's huge.

Tim Fiola: [53:40] You're absolutely right it is and you started to touch on the infrastructure piece of it you know the source of Truth is one part of your infrastructure but your,

CI CD environment your pipeline your coding pipeline to check in code review code do unit tests of code before it goes into your production environment these are all parts of it and

they require resources which is why the director start getting involved with orchestration,

there's cooperation like we talked about and also resourcing which we didn't talk about but like you brought up it's a very important piece to the orchestration because you start dealing with platforms and digital resources so it that's,

that is part of the evolution is that director-level roughly by in.

Zig Zsiga: [54:31] Yes I was talking to another group of people on another show that we did a couple weeks ago or maybe a month ago at this point and one of the things we talked about is like a true

Enterprise architect not like not like an Enterprise architect in this terms of networking but an Enterprise architect that knows the business like at the higher level knows the business understands business speak language Etc but then also understands

every it I guess that's the common denominator here is it technology architecture,

so they understand development they understand automation Anderson application development

lifecycle management for applications but then they understand networking and then distant servers and load balancing and storage and VMware everything already mentioned right and more I'm sure they understand all this,

so that they can be that quarterback or that kind of leader when we're talking about automation because that's really what is needed I think I almost think it's a new role,

I don't know Enterprise Architects to right title but there's got to be this this kind of quarterback person

roll that's facilitating this change in my opinion,

in these organizations and then I think I think would be beneficial to hear some examples if you have any on how

organizations are are adopting this change right how are they they going from the ground up are they going from the top down what is the way to go right there's our lessons learned maybe don't go ground up or top down or whatever,

I think that be beneficial to here.

Tim Fiola: [55:59] Sure the abstraction part like we touched on before that's ground up that's an engineer just getting sick of doing something and saying I'm going to have a script or ansible Playbook do it straight up that is Grassroots.

Zig Zsiga: [56:14] That's me being lazy that that's me being lazy I got.

Tim Fiola: [56:16] Me being lazy call it what you will it's you being productive and lazy.

Zig Zsiga: [56:21] That's me telling my teammates I'm lazy and that's me tell my manager hey I'm productive I'm an efficient I'm saving you money I'm going to get a promotion all that stuff but like when I'm talking to my wife or you know I'm talking to my friends say I'm just lazy just lazy I just don't want to do work.

Tim Fiola: [56:35] As we started abstraction where encapsulating work in scripts and playbooks simple bits of work and scripts and playbooks,

abstraction we might stop there if we want to move further we need buy-in from the from the top like we talked about complex abstraction and automation,

tend to be implemented from the top down.

Because they affect multiple systems multiple processes across multiple organizations and management needs to coordinate that management also needs to fund the resources for the infrastructure.

Which which includes development of a CI CD pipeline

and in orchestration or an automation platform and some let me just throw out some examples of these platforms of these automation platforms I've seen yeah so the your automation platform.

Saltstack is one that I have that I see that's fairly new but it's open source and it's a really cool platform.

It's a true closely by automation capable platform you can also use it for abstraction orchestration if you like but it's capable of true closed-loop automation because it has a

it has an event bus and you have certain processes listening for certain events to kick off.

Simple bits of work.

Zig Zsiga: [58:04] See you tracking events and based on the event that it's receiving you're doing a work okay.

Tim Fiola: [58:08] Yep this interface went down or this bgp pair went down okay great do this do this troubleshooting thing look at the ticket the truck that ticket is you know the interface didn't come up okay then your face still isn't up kick off the next event.

Zig Zsiga: [58:21] What's the type of event is it like SNMP or is it some sort of other protocol do you happen to know.

Tim Fiola: [58:26] You know on saltstack it is a I believe it's a 0 mq message bus.

Zig Zsiga: [58:31] Okay all right cool got you.

Just an idea right like how it works and then and then for saltstack but I never heard it before I've heard ansible and some others were going to talk about but can you lab it like is it something that I could just put in like

my lab environment and run it on like a ton of virtual routers and switches.

Tim Fiola: [58:48] You you know in my opinion I've worked with saltstack a bit it is something you wouldn't want to if you don't have experience at least scripting with python I wouldn't I try and lab saltstack it.

It's it gets kind of involved I'll say that.

Zig Zsiga: [59:07] Not something you want to start out with if you don't know what you're doing.

Tim Fiola: [59:09] No for sure no so the thing was saltstack is it's a platform capable of closed-loop automation which is awesome.

Because it can do so many cool things it's I'll say on the spectrum of complexity it's it can be a bit complex there's a lot of components you need to be aware of and understand and what not,

your automation platform can just be a pure python infrastructure,

if you want to just roll your own and build it from the ground up with your within your organization another one I've heard about but never worked with is combining ansible,

automation platform behind you combining ansible with the rabbitmq message bus

so the rabbit the message bus starts to you know do the communication between elements and then ansible interacts with the network elements directly I've heard of that I've never worked with that but I've.

I've heard of that a few times.

Zig Zsiga: [1:00:05] So I've heard I've heard a have experienced ansible not with rabbit rapid mq.

Tim Fiola: [1:00:12] Rabbitmq.

Zig Zsiga: [1:00:12] Yeah not with that specifically at least I don't believe so

I could just be be naive and not know enough are know what I don't know but I've definitely had experienced the ansible and I do have a lot of actually customers using ansible today

so it's definitely something that people are leveraging and I would say it's a little easier to roll out ansible than it is to roll out like something that you have to know python for.

Tim Fiola: [1:00:34] So and here's why that rabbitmq ansible model,

like say I've never used it it's interesting to me because you can start with abstraction just doing ansible playbooks simple playbooks

then the rabbitmq architecture Ali comes in and gets an acts as the platform so it can start to trigger your ansible playbooks.

To do and that's how you start stringing together an entire workflow one event happens something happens in response that.

Same complete says I'm done now something else oh this thing is done that means I'm supposed to go and that is how you string together,

on saltstack a non-revenue ansible you just string together a series of events.

Of of encapsulated work each one listing for a certain event and that's when I trigger and then I tell the system I'm done,

that becomes that completed effort becomes an event for something else to take to take to take as their trigger,

along the workflow so rabbitmq ansible like say it's cool it's interesting to me just because ansible is a great platform to start ups tracting with and then you can keep your ansible playbooks which you've invested in along the way,

with your orchestrated and perhaps automated infrastructure.

Zig Zsiga: [1:02:01] Okay all right there any commercial software out there proprietary commercial.

Tim Fiola: [1:02:08] You know I've heard you know I've you know like I'm I,

I have no real interest in any of these companies you know commercially obviously I would I you know there's I tential which I've heard as you know which I've no of and then there's a newer one I think called pliant which,

I know a little bit about but they start to let you encapsulate your scripts you've already done.

Amongst other things this is just one of the things I do and then serve as a platform to

use those to automate your full workflow so Architects really those are all very interesting to me because they let you start with abstraction and keep that investment As you move

to the more sophisticated side of the spectrum.

Zig Zsiga: [1:02:54] Okay all right cool so what about like automating things company-wide workflows and what do you see what do you recommend.

Tim Fiola: [1:03:06] I'm sorry about that platform or just what do I see as far as impact.

Zig Zsiga: [1:03:11] Yeah so automating what company away workflow is what are the the the tens the there's a Trends kind of yeah.

Tim Fiola: [1:03:17] Okay I see ya so.

[1:03:22] These will start to affect what we talked about before a strategic objectives.

Improving quality reducing error improving velocity or throughput reducing your quote to cash interval these all tie into or are explicit,

high-level strategic objectives for the company and as a network engineer running a network,

efficiently just as an engineer that is important but it's a lot it's a few rungs lower down on the you know it's a little bit further away from these explicit objectives but,

when I say hey.

I'm starting to improve quality my errors dropping were increasing our cash flow and I'm having a more direct impact on those I as a person and contributing more value.

It would pay when you're started on this journey to understand your organization's or maybe your you know your directors objectives and strategic objectives so you can start talking and communicating in those terms when you're communicating

what you're trying to accomplish with abstraction with orchestration and then with automation they'll probably say.

[1:04:50] We're buying it we're going to invest in a platform we want you to be on this team.

Zig Zsiga: [1:04:54] You know I think that's exactly

how to do it so I correlate those strategic objectives to business success specifically the business goals are trying to achieve which those goals related to logging that the priorities of the company of the organization so like that's how I look at it those strategic objectives are part of those goals and so

everything you mentioned improving quality reducing are improving velocity which I've never heard that before but that's awesome I reducing the quote to cash interval those are all things that are going to make,

business successful so I really we've talked about today literally comes back to making that business successful I think that's extremely imperative going to realize that you know automation,

abstraction orchestration whatever buzzword you want to leverage right these are all tools all things that you can help make a business successful.

Tim Fiola: [1:05:48] That is the name of the game that is the name of the game.

Zig Zsiga: [1:05:51] Awesome awesome um so do you have anything else you want to talk about before we kind of wrap up.

Tim Fiola: [1:05:57] No I you know we've covered everything I you know wanted to make sure we communicated I hope the engineers the audience this was helpful to them and understanding just if you're going to start the journey if you're already on the journey.

Hopefully this will frame where you're at now and let you know where you want to go and give you some some additional knowledge to get there.

Zig Zsiga: [1:06:21] That's awesome Tim that's awesome hey where can our listeners if they want to keep the conversation going which I'm sure most of them will where can they find you on the interwebs today.

Tim Fiola: [1:06:31] Sure I got I have a personal website lock,

s-foils.com that's lock - s - Voyles.com the Star Wars reference.

Zig Zsiga: [1:06:41] Star Wars.

Tim Fiola: [1:06:42] On that on that website I have you no more technical blogs and then I also have a worldly observation,

section where I comment on,

you know my thoughts on tequila and why it gets a bad rap tortilla chip deception

why I think taking pictures of your food is is a ridiculous thing to do

all these things I discuss on on on the worldly observations part so feel free to check that out there's a

few links there where you can contact the webmaster and that'll get back to me via email I'm on linkedin.com Timothy Fiola all one word,

if you search for Timothy Viola you should see me and I'm on github.com.

My user id there is Tim - viola.

Zig Zsiga: [1:07:36] Awesome awesome and I will have all those links for Tim in the show notes you can easily click those links and get a hold of him and ask him a whole bunch of questions about tequila and Legos and of course abstraction you know in orchestration automation as well.

I mean you know we care about tequila and Legos first and foremost of course.

Tim Fiola: [1:07:56] Talk your ear off on tequila.

Zig Zsiga: [1:07:57] Right hey Tim I appreciate it buddy thanks again for joining me I know we got a couple other shows that we're going to do in the future so I'm looking forward to those thanks again I hope you have a great day this is a great show.

Tim Fiola: [1:08:10] Zig it's been a pleasure thank you.

Zig Zsiga: [1:08:11] Hey friends nerds geeks and ziglets that's going to close out today's episode of the zigbits network design podcast where we had a discussion with my good friend Tim Fiola on how to how abstraction orchestration,

Automation and we coined it true automation can make your career successful.

[1:08:30] Make sure to follow Tim he's great follow on his social accounts you can follow him on his website you can check his get did get home hub repositories

he's actually in the Discord server to so if you want to reach out to Tim he's in there lurking around and again if you want to find all this information all our show notes today is zigbits

text / 90 I have opened up the doors to my network design course called designing Network architectures and ensuring

business success if you're interested to hear more about my course or are ready to enroll right now,

visit zigbits dot tax / DNA if you want to have a live Network design conversations right now with myself and other skilled design experts join the zigbits Discord Community you are truly missing out,

now you can said there's a ton of Highly skilled experts not just design but they're skilled in all these other ass areas

of expertise and they're just really ready to help you there their kind of chilling having conversations with other people and they're there to help you so if you want to join have your design questions asked have real conversations.

Technically highly skilled people go to zigbits top Tech / Discord it's 100% free community there is no requirement on price,

payment or donations or anything like that obviously donations are welcomed but they're not required but it's just an in community for all of us just to be.

[1:09:59] Us and to talk and to help each other that's all it is,

if you like today's episode let us know you can find more zigbits network engineering Network design and the network architecture content including the technical podcast like this one you're listening to right now our monthly webinars,

our YouTube videos and a dedicated community on Discord all of this content I just mentioned is free content,

fun all this and much more at zigbits. Tech you can follow us on Twitter at zigbits and you can find us on LinkedIn.

Hey sign up for our free Weekly Newsletter the network design digest filled with the best network design content in network engineering today and that's just at zigbits dot text last newsletter,

as always I appreciate you and I thank you for listening don't forget to attack your goals attack the day attack your life,

and make progress my friends until next time bye for now.