You were told to install kubernetes, now what? A talk with Robert Kloosterhuis of ITQ

Episode 254 · Tue, 14 Feb 2023 07:30:00 +0100 · 53:41 · SRT · archive.org · all Tanzu Talk

well uh kubecon is coming up is that next month in march i've lost track or is it may it's april april okay right in the middle there sandwiched between the two m months i think uh you know this this is a good opportunity for me to do my uh my my uh periodic query i think there's only four months that share the same letter right march may june and july well i'll have to go through the rest of them and see if i can come up with with other ones i'm probably skipping something because i'm not good at these memory tests yeah don't do this to me well here's here's before we get to the the the meaningless of it since kubecon is coming up now i know i know you don't live in the amsterdam area or you're not an expert on it so we don't need to go over like restaurants and things like this but here's something that having lived here for a while and given advice to people i realized you have to cover is like what is your advice for figuring out how how like the trains and the trams work and and just because it seemed to us well well to us it's simple right but i feel like especially if you're coming over from the states you might just be like whoa right you've got a lot of stuff going on yeah oh i have good news for everyone just use google maps uh yes yeah google have they ingest all of the essential infrastructure data that our country provides and you can use it perfectly for public transport and public transport is the way to get around i should add do not expect to be driving to kubecon don't even try and and it is uh there are multiple stops on all i think even the metro goes there right on all all forms of transit to where all forms of transit but of course if you're going dutch you should go by bike yeah now now there is so here's a caveat to use google maps our dutch rail service ns does have their own app obviously and if you install that app then as part of your public transport

experience you can rent a bike at just about any dutch train station using the app and it gives you the perfect way from the train station to get to kubecon by bike if it's too far you can always use the subway but i would definitely recommend trying the biking experience i don't know if you know the youtube channel not just bikes um it's another it's a canadian that lives in amsterdam and his it's a very popular channel and um he um he uh all his entire content his whole thing is using amsterdam and dutch infrastructure as the shining example to the rest of the world what urban planning could look like yeah yeah and what what actual useful public transport actually can look like yeah that's it's fun it's fun to watch that you know living here because it's uh yeah that guy has a lot of opinions which are pretty much accurate yeah yeah and and you and you can speak from experience having how long have you lived in amsterdam now like four and a half years basically yeah so yeah yeah you know so so to build on what you're saying yes i didn't even see this is why it's great to ask you because i didn't even consider that part is like yeah get google maps and uh you can anywhere you want to get you just switch it over to the transit thing and it's good it'll tell you like down to like the platform to go to at what time and it's it's uh it's quite impressive so like when you get to uh when you get to the airport you can just put it in instantly and it'll tell you how to get to your hotel and where to go and the other thing now i haven't tested this as much as i want to do to give this advice but i'm going to give it anyways uh recently they finally like uh got to the point where you can just use your your i don't know if this works for androids but you can use your phone to pay for uh yes fair which means if you set up the transit stuff on your phone you can just go up to the um i don't know what you call them the gates and just put your phone

on it which is it's always mysterious because there's this piece of glass and it looks like there's nothing underneath it but magically it works yeah and i i think i i don't think you get a discounted fare but you probably have to have a bank account live here anyways to get a discounted fare but it's great you just like when you approach the gates you uh you tap in and when you leave you tap out and there's only one confusing aspect and that is if you're switching from uh from a metro to something else you've got to like tap out and tap back in but you'll figure it out it'll it'll it'll be cool yeah and another another interesting point that people might not be used to we have a single system across the entire country for this yeah one single you can get these cards anywhere or you can do the app on your phone it doesn't i don't think the the phone thing works everywhere yet you get the cards the cards will work in the entire country of a national system yeah one system and you know what i would recommend depend if you're staying here like however many days you're staying here like right when you're at the airport if if you're not there at like 2 a.m although you can probably buy it at 2 a.m too you should just get like the tourist card that's like for three days or five days or whatever for for the trains and obviously don't lose it but you'll be set it'll it'll be uh it'll be super easy and then yeah that's the other thing is like i tell only one person out only one family out of i don't know five who visit us here over the years has ever taken me up on the vice but you should just rent some bikes day one and and you know you can also go to the station and just rent them hourly or whatever but like just just get some bikes it'll be fine i mean you're here and and have a good time and uh go go around that way because that's that's the uh the as our canadian friend would tell you that's that's what this whole place is optimized around so it'll be the the easiest to do and then you could be those people that are dinging at the uh the other tourists which might feel gratifying for you i i have to i do kind of imagine you know americans you know biking maybe maybe not for the first time

but not being used to the how how many bikes there are in amstam and how dense that that that experience can be it presents different challenges to cars yeah it can be overwhelming if you're not used to it yeah yeah yeah you know there's there's a little bit of just like uh gird your loins and take your life in your hands but just do it for like 30 minutes and you'll be cool you just observe how people are going what what i've learned and then i think this is solid advice for kubcom people is like one of the things i remember i don't know what they teach around here but one of the things i remember from driver's ed is one of the more important things was look around at the other cars and do what they're doing right like just go with the flow of things like of course you want to follow the rules and things like that but it's also important to kind of read the road or read the room as it were and just do what they're doing and similarly like if you have that attitude just like spend 10-15 minutes as you're biking around looking at what all the other bikers are doing and just do that now you might want to actually stop when the bike lock lights tell you to stop unlike many people but that's fine if you you want to do that i i do i do notice the cars have really good brakes and radars these days they do well not all so so now now that we uh uh have established your your uh regional expertise here what what is what is your expertise relative to computers who who are you and uh what do you work on well my name is robert i uh i work for itq for the last three years now um and itq is um is a consultancy and uh vmware partner um for a long time um itq itself has been around about 20 20 years or so um um and myself well i've been in it for about 20 years as well um i'm one of those kids that that when you asked like what do you want to be when you grow up i had no answer but i did say something with computers there you go yeah so so um so basically uh once once we had a pc in the house

and olivetti 80 88 um i uh i owned it it was mine and i uh basically that was the start of uh of my interest um actually because i mean see i i think my dad took me to war games when it came out all right like it must have been like six or something which is not an appropriate age for that movie i believe but i remember being being so hooked on that movie i could literally um in the school yard like this little six-year-old was basically spouting all the lines from war games to his clueless and very confused six-year-old friends so that's probably where it started i see yeah yeah i can imagine oh it wasn't exactly like an elementary school favorite there uh like for for the uh the six-year-old set yeah i'm trying to remember what i guess just the general feeling of impending doom that that movie brings wouldn't be great for uh for for younger children like especially the questions about like could that really happen no one wants to discuss that yeah i think i was a little too young to understand the full context of that i remember when uh i remember it was the same year i remember chernobyl being on the news and not understanding when they were talking about a plant that they weren't talking about a leafy plant you know with green and in a pot somewhere right right that confused me yeah so so what uh i mean you had a practice there at itq right what what is what is this practice yeah so um so itq has been um um has been doing kind of general it consultancy for a long time we even did software development at some point in the past and i think the last uh last 10 years we focused on basically vmware we decided to essentially become a single vendor consultancy with a focus on on the vmware ecosystem um and um we're doing pretty well at that um and

we're kind of at the european level now so we have offices in various countries and about 150 odd people and i i think um about five years ago five or six years ago um some of some of us um decided to to look into this cloud native stuff that was starting to emerge and very early on we developed a partnership with pivotal um and um and we uh we would um help pivotal in engagements we would do the pivotal labs stuff we would uh we would be um we would help in the dojo engagements and for those those who don't know pivotal had a pretty interesting approach to customers as you'll be very familiar with um and we were a we're a big fan of the way they did things and the way they approach customers and and how they how they how they engage with customers how they helped train up people by pairing with them by doing these really kind of intimate and and um pretty intensive engagements with developers or with um or building a platform team around some of their products and um and um so yeah we from early on we were involved in that um in getting pivotal kind of started up in in the netherlands and in europe um and um and that's how we started our cloud native division as it were um and um and it's only kind of been growing since there and of course a few years ago vmware reacquired pivotal came back to the family back to the mothership and um and uh the tanzu portfolio because it was born around the same time and that's around the time that i joined so i joined um yeah about three years ago tanzu was pretty new and i i came in i've got a long-standing infrastructure background so i did you know i've done i've done v-sphere storage um vsan uh touch more or less every vmware product but i came to itq and i said

this cloud native stuff is really exciting that's what i want to do um i i'm i know all these other products i don't want to do any of that anymore um uh just just put me on this this cloud native stuff and that's what they did and since then i've kind of um you know i like to be really broad so i i know a bit about just about everything in the tanzu portfolio a little bit you know i'm not i don't go very deep well some things i go deep but but um this broad kind of um view of the whole portfolio allows me to to to be the guy who can explain it to everyone else inside itq and then to our customers so i've ended up as one of the tanzu practice leads in itq we call them technologists um and it's basically me and my colleague sandra who's more on the dev developer side and so i bring the platform side he brings the dev side and together we kind of do our our um the tanzu practice or the cloud native practice inside itq um and it's headed up by uh by um our uh our valent uh uh vp of uh of cloud native martine um and yeah we're trying to build out this and we uh of course we see ourselves as the uh the kind of most cutting edge it's part of itq obviously um but it's it's a it's been a wild ride the last three years um we've really um you know a lot of customers a lot of products of course the tanzu portfolio is really evolving um so keeping up with that has been a real challenge but it's been it's been an amazing ride so far so so let's start with like uh you know uh for people aren't watching the live screen you have a v-san sticker on your laptop a testament to your exciting infrastructure background but so so let's let's start with the infrastructure background like when you're talking with people whether they're customers or not customers and they're kind of like infrastructure people um like how do you explain like this platform thing to them right like and and i ask

this because i talk with infrastructure people as well but i come from the application developer perspective right so i'm sort of like anytime i talk with infrastructure people i feel like i'm always learning more than i'm telling them about about how they think and kind of like map their notions up to developers but like what what do you what do you do to like or what do you talk about to get infrastructure people to be like oh developers that's fantastic right like like to kind of get them to be uh work with them and be you know have a good time with developers yeah well that no that goes straight to the heart of it doesn't it um so so it's funny the best just if i go back to my experience you know on in infra you know doing infra in all kinds of different ways i mean the thing that kind of stands out is that when you never really ever deal with developers directly um traditionally so um so that the the closest the closest you usually get um from a kind of that v sphere infra kind of uh area is um is you deliver a vm um and you might provision you know you might be the same team responsible for provisioning what goes into the vm and then you hand the vm over to someone else and then the vm becomes someone else's problem and um and oftentimes especially in larger organizations where you have thousands of vms running you don't want you don't know what people are doing on those vms you don't know what's running on them um and it's it's in that area where developers uh will start to to do their thing so traditionally there's this huge culture gap and when we go into customers the first thing the first thing i do the first questions i ask are about to see where that culture gap is is there a culture gap is there a gap how big is it um how close are they to the developers is it the same is it separate lines of business um you know is it uh is it uh do they fall under different directors you know what's the political situation because because knowing how far away um an it team is from

from um the uh you know what developers are usually doing building kind of you know value around the the main business process um understanding how far away from that they are really tells you a huge amount about how to approach that customer or how to how to do an engagement or even the kind of language to use um so so and i notice i find it's different with every customer right there's no i mean there's there's always a there's always a gap of some kind um but um but sometimes uh especially if the customer if they if the companies are smaller or newer that gap will be smaller usually they'll be closer together if it's a startup or a startup mentality uh it'll usually be closer but it really really depends and um and so it's it's it's hard when that when there's a gap when they they really have never kind of sat down down with their own developers to understand their world it can be really hard to to talk to them about shipping code or um yeah or on and it goes to the fundamentals of what containers and kubernetes and all that stuff is for um because you you have to maybe take quite a few steps back and start building up a picture of how modern software development goes to to actually educate them about how it probably happens inside their company um so so i end up with these engagements often um getting people together sometimes for the very first time and and and having them talk to each other and learning from each other how the other the other side's world actually is um and the first thing i do when i do a workshop around i know implementing tanzo kubernetes grid or something like that i say like you know what are you doing it for you're doing it for developers so we're gonna have a workshop and the developers are going to be there you know although the line of business is going to be there the project is going to be there it's not just going to be with infra i mean infra might might own the product and that even that is sometimes an open question um but i get them together and it's often for the for the first time um and that's a

those are interesting conversations you can just kind of you you set it up you sit back and then you see them start asking each other the right questions uh if you're lucky sometimes you have to help a bit uh to get that conversation going yeah that that's i i like i like the idea of you know it's it's making me realize i have to do this too and i'm not very good at it it's like calibrating pretty quickly where everyone's knowledge is of you know with with i mean it goes in both directions right like whether you're a developer or an infrastructure people like it's like i'm i'm quickly thinking through like the last you know conversations and meetings i've had for the past year and you've got to like go a couple of levels below how are you set up to figure out like how people are actually set up like i remember one conversation i had with someone in you know this this group of people had built out uh i don't know cloud stuff and kubernetes stuff and so they were like we're cool and then you know you you you as you're getting to you get closer to like yeah so uh how do your developers work and what do they do and then they're kind of like that's a great question i mean yeah and and but but and then and then as you're saying the the next good part after that and i think it would be interesting also to contemplate how again for developers you would have this same conversation in the other direction is like uh it's often handy to just go over how software development works like like what all the components of it because it can be a a bit of a odd mystery of of how these things go you know you you were making me think in in describing this again drawing from your uh your history like i'm i wasn't in this area like whenever virtualization became a big thing but you've worked in that area and i remember like i'm always curious like was it a similar structural issue because virtualization was and is kind of like a cross-cutting thing that doesn't really care about your organization structure now it's been

around so much that was just like we take it for granted but back then it was just like yeah what if we did everything and so it's i would imagine the conversations were kind of similar yeah well it's so i've i've i've i've experienced this this cyclical thing twice now in my career because i came in around 1999 2000 windows 2000 had come out just come out and this introduced a whole bunch of new stuff that windows nt4 and previous administrators were not used to things like active directory and um and a group policy and things like this and i remember and i was this is when i started my consultancy career and they they do what most consultants do this they send you to get the you know microsoft mcse you know so now you're a paper mcse you don't actually know anything but you have the thing right and and what you notice is that there was a generation of um of admins who had come up from the kind of the the kind of the old kind of mainframe data center days had grown through novel network and into into into um microsoft nt you know nt 311 nt4 um but 2000 introduced stuff and that was a real challenge for them um you know a lot of you really saw this drop off at that point where people struggled with that with new concepts with that kind of new those new things and then when when virtualization came around not that long after um you saw a very similar kind of effect where um where younger people had an easier time adapting to this concept of now we can do hardware as a kind of this virtual construct you could and i remember i remember the first time i struggled with this i was like i don't want to do this dual boot thing no it's not dual boot you just you just run this thing called vmware workstation um and it wasn't even called that it was the previous name and then you just and you can spin up a server inside on your laptop and it doesn't affect anything else it's like how can it not affect everything else that's not going to wipe my boot record come on this is

nonsense um a lot of people had trouble um adapting to that that that idea and even to this day i still do come across hardcore database guys who say listen i'm never going to virtualize this this this microsoft sql server that we've had for 15 years um because it'll never run as well in the inner vm and of course they're usually right because vmware you know virtualization as all as all virtual you know kind of constructs have an overhead and things are always you know things are always never clear cut but um but it was yeah i was in the middle of that revolution when it came up um where everyone just said like oh the real you could see the realization coming into people's heads we don't need hardware for everything you know we can set we you know we can separate that ads and then and then of course vmotion which always blows people's minds when they see it the first time yeah even today by the way um you know someone coming in fresh from college i've seen this experience recently they come in they've never seen the vmware stuff you know they've been trained on maybe amazon or something and you show them you know vmotion at the esx level and they're like wow it didn't even have to go down you know this this uncoupling of hardware and software they've never realized that that's the thing um so even so if the emotion that vmotion movement even happens today which is which is funny to me and you know yeah yeah no no that i mean that's that matches like my my theory as you were saying that there was uh uh i don't know every 10 or 15 years you just gotta like change yeah the the way you think about like your the the is responsibility the right word but like like what you're doing as an infrastructure person and and how you're doing it right and and if you you know if if you uh if you allow yourself to kind of float down that river of improvement like things are cool right but it's it's hard to uh not not to say it's hard to go against it but it's just like yeah it just stuff

changes like you know i'm i'm i'm remembering that like at uh at explore eu you know kelsey hightower was up there and it was it was like a short five minute talk but i think that was pretty much the point of his talk is just like yeah the stuff that you all do used to be crazy and now it's normal and just remember that at one point you thought it was crazy and and you know it was fun and interesting to get through it so like speaking of things becoming normal and not crazy like with with the so with the organizations you work with right like with the customers like what are like what i'm always curious about is like so they they want to put a platform in place so they're talking with you and like what are they running on it like what's motivating them to like put this platform in place rather than just do what they're already doing like what are the reasons that they're shifting over so that's that's an interesting question so um so most of our customers that we have most of our network that we have they're the same as vmware's network they're um you know they're people that we've known for years running vsphere running you know vra running you know nsx um so our contacts and our network are mostly infra right so they'll be it directors there'll be central it inside these companies they're not crucially that they're not necessarily lines of business they're not necessarily where the developers live so when we get the ask and this is i think similar for vmware when we get the ask it's usually coming from central it from infra and they don't know why they're being asked they're simply being asked to provide say kubernetes as a service to the rest of their organization or to a specific project that's running or you know when they're just starting up um or to uh you know whatever line of business represents a group of developers and when you ask them well what's this for you know what's the what's the actual use case behind this they don't they often won't know um because

there's this gap um and and i love customers like that because you know because it allows me to go on a journey with them which excites me um because that's the journey i've been on so it's like oh you're in for a ride um and um you know and of course you know and some people are really all in on that ride and some of less so but that's that's part of you know what makes this uh an interesting job um it's very sometimes we'll we'll we'll come in and it'll be developers it'll be a line of business or a specific project or something saying look we need we uh we need say you know kubernetes on premises um you guys specialize in this vmware stuff apparently vmware also does something with kubernetes why don't you go help our central it set this up and we're like great we can do that that's what we do um so you always end up at the place where where um the vmware products are owned which is almost always still central it um now we could we can have a discussion whether that should remain the normal uh because i don't think it should i think that um i think that that maybe when you especially when you talk in terms of platform teams and the stack of technology you you need the full stack of technology you need to um to create say you know an internal developer platform or some kind of developer experience something that's going to be consumed mostly by developers maybe it shouldn't be owned by central it maybe it should be distributed or maybe you you need to create a separate division for this so this is the kind of these are the kind of conversations i have too so it's not just about product and about how to implement you know tkg or avi or that stuff but it's also like how do you position this inside your company how are you going to sell this to the business or how are you going to how are you going to make sure that that it doesn't immediately fall into the same list of building blocks as say virtualization it may not fit best there maybe this is a separate category i think it's a separate category because these technologies

like let's say kubernetes and up um is way closer to software development you know it's it's um the way i explain kubernetes to to people is is you know see it as see it as a framework and infrastructure to build a scaffolding to build modern application architectures so that means you know this and it goes so that goes way beyond just vms vms are like individual building blocks you can use you know to to do runtime kubernetes does a lot more you can describe the whole architecture in it so it is naturally going to be closer to the people building that architecture and who builds software architecture it's going to be the developers it's going to be some project or it's going to be something something that's probably outside of central it you're going to have to be pretty close to that wherever that's coming from in order to make the adoption of these technologies be successful to actually fit the use case you have to bridge the gap so so that that creates all kinds of interesting discussions um within within within companies within organizations that we talk to because it will be the first time that they've been confronted with the fact that they're building a platform specifically to enable software development in a modern way um and but to make it successful they they they it has to be you know they have they have to bridge that gap that sometimes for the first time they're really really forced to bridge the gap um and and then we start talking talking about well let's talk about self-service instead of doing everything through a ticketing system for example um and let's talk about automating dns and certificates and uh and data services instead of having to put in a ticket um and and this is where things get crazy because the scope of what what it takes to make um a kind of a modern app platform actually successful then suddenly they realize it goes beyond just um you know building some technology and raising a stack inside central it but it requires a

bit of a cultural shift it requires an organizational shift and that's that's really at the heart of it um all the technology is really cool and it's going to keep changing and it's cyclical but what the harder thing to change is the people in the process right it's the mindset and the organizational stuff you need to put in place to to actually make the adoption of this stuff actually successful that's the real challenge you know you may maybe you can help me figure out how to word this thought that you're uh you're you're uh making me have which is like so you know the the the easy main way to talk about like you know whenever using cloud native stuff like building stuff on kubernetes or using like a paths thing like you know like a cloud foundry thing or whatever whatever it may be building these platforms developers is often around like automation and speed and kind of as you were alluding to not um not have like avoiding having to file tickets for getting stuff right just self-service and and all the good ways but there's another angle that like it's always well i don't know it feels like it's weird to talk about if only because we don't talk about it this explicitly but it's also like whenever you're putting a platform in place there's both like the opportunity and the need to do a lot of customization of that platform for how it works in in your organization how your developers use it and you know build a platform out on top of it and then as you were kind of getting to that's where if you are building kind of a a customized platform then of course you should talk with the developers to see what works for them and you know it doesn't work right and in contrast right like i mean as we were talking virtualization has been around for a while and and sort of like i mean i work at vmware so i know there's a lot of interesting innovation and stuff that goes on there but it's like it's we've got it figured out right like as you were saying there's like the building blocks of virtualization and like what you do with them whereas like when you're building like cloud

native platforms like i don't know if it's all figured out let alone like how inside your organizations and with various groups of developers how they want to use it and so it seems like that's why you have this pull if you're just more of a a traditional infrastructure person to think a lot more about just not so much how you provide and ensure that like the infrastructure you're giving developers like runs and that it has all of the illities that you need but also that it just has like the features and the functionality that you the platform people are kind of like own like you're in charge of that and and building it out which is a different i think that's a different function than infrastructure people have had for a long time like customizing things out yeah this is um this is also where the um this is also a bit of a trap that people can fall into really easily um so so one of the things pivotal was like to talk about was the concept of platform as a product and the idea behind that is is think of think of an app you know think of um think of a version 0.1 like an alpha version of an app or maybe a a game on steam right you decided to fund it on the patreon you get access to the uh the early alphas and what you what you get is a you know a minimal viable product you get you get a beta version of the game or a beta version of an app you know that it's not going to do everything that you there's a vision there's somewhere you want to get but the initial version won't have that but you go on a journey with um with with uh you know whatever team is behind the game or the app and they have a road map and they have an ambition and they have a vision and and it gets better and better over time um so so the what what a lot of especially in in corporate it um they're used to just buying buying products off the shelf

and saying here's a product here's a feature set this product's been around for for ages it's supported by some vendor it does you know x y and z and that's what you get and um and it's a kind of a closed loop you know the if you if you wanted to do more than that well we can maybe put in a feature request with the vendor but that's not necessarily going to be you know it's it's you know it'll be in a road map somewhere but but when you when you look at kind of you know indie games or apps it's a lot more iterative right it's a lot it's a lot more cloud native shall we say um and the way i advise people is think of platform teams and the product that they're building in the same way start small you know don't don't don't try to reach for the moon don't don't don't you know go go into your organization find requirements write them all down but create a road map that's very restricted in scope you don't need to solve every organizational problem every developer need in day on day one um you might want to start with the most important ones but sometimes the most important ones are also the most complex to solve so i'll give you i'll give you an example to make it a little bit more tangible um data services so okay you've got kubernetes you've got a platform team that runs kubernetes for you developers want self-service data services and they want to be able to just consume a you know postgres database um now you have to provide them a postgres database so that's already a challenge for some platform teams because they might not know anything about databases certainly not dbas might know a bit about kubernetes um so you can you know use use an operator spin up postgres deliver postgres as a service and then if it's um if it's a windows uh kind of based organization they're used to developing in windows they might say hold on this postgres thing can this uh can this can this can this integrate with microsoft kerberos and and then you go like uh that sounds like fun no and you go go to the operator it's like uh no i don't think so and this is all we we definitely

need we definitely need kerberos authentication because that's how we're used to working with with databases you know we need uh we need we need we need the uh the microsoft service accounts that do the service identity go straight into the database that's how we do things but this is this is where things can kind of get off the rails because the the their requirements is is also a bit of a a mental trick people do they're used to something they're used to a way of working and when you ask them what they need they will say they will give you a version of reality that matches what they used to it doesn't necessarily map to what they actually need so part of having a successful platform team and building a successful road map is being able to pass the difference between those two things um this is this is what they do this is what they say they need this is what they actually need but to understand what they actually need you have to know a little bit about their process you have to understand what it is they're building and able to and then you can say look how about this we're going to give you on-demand databases so you'll never have to do a ticket ever again how about that does that sound good and they say yes okay but it won't ever have kerberos authentication because you're going to have to learn how to do you know a different way of authentication but you get this amazing self-service experience back in return what do you think of that oh that sounds kind of cool they might say well they might not say that um so so building a successful platform and the road map for it is always a negotiation um is basically my point um yeah you're never going to be able to give them everything they need but you can give them most of what they need yeah and and to your your i mean you know it's a it's an excellent uh change agent tactic to give people things they want to have them stop doing things the old way or or to put it in a more positive way to have them start doing things in a new way right as you know give them advantages and and tasty treats and things like that and you know you're you're also like in in what you're going over there right like there's

maybe the the notion that that i was looking for some good wording on is like when you're putting a platform in place you have the option to like to make the platform better for your needs as you were saying right instead of like waiting for another road map or trying to influence features being added that just kind of come out of the box you can actually like change things not everything but you can customize what things are like add in different services like you're saying uh like how integrations work all sorts of stuff like that but what that means is that you have to be able to basically like product manage it like you have to go through that process of figuring out what to do but you know having having the uh having your infrastructure finally behave in a way that your organization wanted in the first place should be pretty attractive like and and it's also like i keep saying this word customize and i think it's like a uh uh commentary on the the fraught nature of this idea is like at some point i learned that customizing was a bad idea but but in fact in this case you know it's it's like kind of like you were saying the whole not the whole but a major component of doing like platforms in this area is that it is a system that has been developed so that customizing is okay is not a bad idea to kind of be lost on your own custom island but customizing has a cost associated with it and i think that's the most important thing to always you know be aware of because um because say that that it becomes an actual requirement to to make sure that that postgres can be accessed through some kind of kerberos integration i mean there are solutions for that there are technical solutions for that but they they take time and expertise to implement and they add an overhead to maintain and that and that can be the role of that platform team right there's nothing wrong with that but um but what you can this is the trap you can end up in everything being um super custom and impossible to maintain it's like spaghetti code but at the

platform level you end up with with um you know 20 point solutions built into your platform to address 20 needs how big is your platform exactly supposed to be how many people in your platform team do you want to have how many different skills are they supposed to have is it still viable now at larger and larger scale of course you can get efficiencies around this um but if this is every every company every organization is it has um you know a different dividing line where where they where they where they draw that where they where where the the effort becomes too much um i think i i mean we touched on the role of the product owner or the role of the project leader or whatever who is managing those expectations and managing that roadmap that role is super super important um um and it's certainly going to become more important um to uh to to create you know to create platforms and services like this um and that and of course the and i touched on the skill thing um and maybe we could just expand on that one a bit because that's a huge problem right the the amount of people currently in the kind of workforce that know kubernetes is still relatively small uh the people understand cloud native architectures and all the stuff that goes into building a cloud native architecture is relatively rare still amongst developers amongst uh you know infrastructure people ops people um and i don't see that problem going away anytime soon in fact i kind of i see signals that get it's getting worse and it doesn't doesn't help like as you pointed out that this this ecosystem is still like really in flux still and uh very much uh kind of we're in the early days of it um but um but this you know i i really see that the problem around around skills being a real impediment to progress it's that's again it and when it goes back to the platform team this is what i say like keep your scope really constrained because you're not going to find you're not going to be able to add a dba to this team um you know anytime

soon you're not going to be able you try to focus on one or two services first become really good at delivering those before you go wide um because you're not going to have you know the skill sets you know you it's a distributed mongodb um you know you have to find a mongodb guy right um yeah yeah yeah i mean to to oversimplify it almost to the uh the uh i don't know the level of cuteness and triteness right like it's it's i mean it's it's always good to know how your tools and technology work like i always think that like yeah culture is great but if you don't have tools it's just a bunch of talking uh so you you always need tools but also like kind of the it seems like in in this fluxiness period that we have like the most durable thing you can do is like understand what your needs are right like understand like what the developers need what your organization needs and understand also that kind of like product owner product management process of continually refreshing and understanding it and that way whenever some new technology or you know different versions of the same technology come around you can always start with with what the needs are like what what people are doing instead of having to um rediscover those needs based on the tool you end up with right like you know i'm a big fan of changing the way you work according to the tools that you have but not like all the way it's also good to be realized like here here's like what we need and so we're gonna get tools that work this way or force our tools to work that way right like a good a good middle path uh oh yeah that's um that's a question that customers often don't expect um you know when i when i say this this use case you think you have does it actually need kubernetes right it's actually needs containers does this need does this need to be a microservices architecture um and they're like well you know it shocks them um but but i mean there's there's there's there's a place for all of this stuff there's a place for monoliths you know

there's a there's a place for spaghetti code somewhere there's a there's a it it it doesn't you know that nothing things often don't need to be perfect to be successful um you know we we see that um you know we see that all the time um i mean it's twitter just take twitter for example right i mean if you if you if you believe the twitter files and if you if you're following all the stuff coming out of twitter now you think oh it was a mess inside but it was good enough it was good enough to create a service that millions of people were using um successfully happily um you know even even even what we think of as the most cutting-edge services that are actually really messy on the inside and we often don't know because they work well enough um and and but you know engineers like doing things properly they like to you know they often feel a little bit thinking in black and white they they like to to make have the best technical solution they want things you know implemented in the best technical way um and that that's of course and this is not related to cloud natives just just generally to get some way it doesn't always have to need doesn't always need to be perfect it needs to be good enough to address the need and the same is true for for platform engineering or what you what you want to what you want to deliver even the quality and what you want to deliver it i mean of course there are best practices but you can you can get in in the end you want to you want to be able to achieve something that's a value to your business or to your use case um and so you you can compromise um but you need to compromise in the right ways um and that can be that would be a real trick to understand like you know what's important what is not important yeah i mean it definitely like all of this uh shifting to a platform stuff or when when when when you get the request to build up a kubernetes thing like it's it's a good it's a good moment to just do your standard five wise thing not to be an annoying contrarian you know the the the uh the consultative person who's always like let's back up and start with why but really just to

like you know every now and then it's good to like go down a five or so wise just to find out like let's just make sure we're doing the best here that like we're uh we're meeting the needs uh that people have well speaking of meeting the needs that people have if someone had some needs that they needed addressed how would they uh how would they get in contact with y'all what should what should they look up to see if uh you're someone who could help them out well um so um well i mean i'm on twitter and um but i'm not as active as i used to be i've also been playing around with mastodon a bit um as we probably all have been recently um but um but i do occasionally put some blurbs out there um i used to blog more i find it hard to keep up with blogging um when it comes to my company anyways that's that's the problem with well very unsatisfying well i'll tell you though i mean the the the reason i've if i block something the reason i block something is because i wanted to show up in a google result in six years time right right that's my motivation is that one guy in the middle of nowhere who's run into a problem and it's like my blog is the answer and it doesn't matter that it was like you know a seven-year-old blog post i get comments like that it's like oh my god you saved me i'm like why are you still working with exchange version you know 2010 or clear or clearly how to get microsoft kerberos to work with postgres i think oh yes yeah exactly yeah um but uh when it comes to the company i work for itq um so um so so we yeah we specialize around the whole vmware stack so we're the people to go to for that but um and this is kind of um to go back to the skills issue um we think we think a lot of people are going to struggle with this cloud native stuff we can see i mean we see them struggling with this cloud native stuff the skills aren't there uh we saw an opportunity like a like a you know a gap in the market to address so we've uh we teamed up with vmware to create to create a managed service around tanzu uh so this is kind of a new concept

uh new for vmware too uh to actually do to do managed services around around some of the cloud native stuff that vmware's building um so we've just launched that we launched that at explore in barcelona um and and that's what i'm currently working on the most that's my whole focus right now is building out that whole service um so so that's i just needed to pitch that in there you know have my commercial hat on um because because if people want to contact us they should definitely contact us about that yeah so i wrap that up yeah you know i i i you know related to that right like uh i'm always like tossing around in my head like the notion of like doesn't everyone just want public cloud and and what i mean by you know again to apply the needs analysis is like you just want to like not really there's a certain level of infrastructure management you just don't want to do anymore right and whether whether that's like someone else is managing it for you or like you get it from like public cloud over here or whatever like it seems like wherever that line may be there's a lot more uh interest in just like having someone else do that but without the um without the like we have to file a ticket and be compliant with our uh five-year contract that you hear about from traditional outsourcing right you want something a little more fungible uh which which i think it's it's exciting to see how that's uh that's panning out with people well well thanks for being on it was it was fun uh i i had a few things to uh noodle around on here to use some other mind idiom uh after talking with you and you know i i'm maybe we'll see each other at kubcon we always uh we always sort of accidentally run into each other yeah uh when we're out and about yeah okay so yeah well we should remind people and this is some more relevant for the dutch markets there's a vmug coming up in march so the dutch vmug which is the biggest vmug uh vmware user group meeting in the world um it's that's coming

up soon i'll be i have a session there around um uh some some best practices and kind of lessons learned around uh tansy kubinators grid and the avi and basically you know some of the stuff we talked about how to do a platform team and then yeah we're going to kubecon um not speaking there but we're taking 10 of our consultants to kubecon because we just think it's an important thing for them to to be exposed to that ecosystem it's right around the corner it's amsterdam which is great and um and i'm planning to organize a vbeers around kubecon so that's an interesting combination because vbe is very v community thing very much a vmware thing and then that's at kubecon so we co-located it you know somewhere in the amsterdam city center if you have yet to pick a place for it um and of course i want to see all our vmware friends show up there including those who uh who are at the who are doing booth duty uh at the uh at kubecon uh all the developer advocates and uh all those all those friends but uh but anyone's welcome to to show up um i haven't published anything um just keep an eye on my twitter for that um and uh it'll just be a cool way for the um the more v communities kind of people that are walking around kubecon to get together maybe um but i would definitely um drag anyone who's dead deliberately not from the v community so people outside of our ecosystem drag them along um because meshing you know the the cloud native the cncf community together with the v community sounds like a a pretty good uh pretty good plan for all of us yeah yeah no that should be good because i i think i think people around here on on uh what whatever whether whatever if if you're a v or a k so to speak uh everyone around here is always very curious and open about what's going on and has uh is always very forthcoming with their opinions so there's always a uh there's never a boring conversation up here about computers and stuff well great well i'll uh i'll try to list those things in the uh the show notes which which you

can get dear listener if you go to tanzu talk.com and uh it's easy to find this episode but you can uh you can see the the v mug that's going on i i spoke at that last year and it is it's massive it's a it's a very large v mug uh which was nice and then also uh links to the other relevant stuff well thanks for being on yeah you're really uh you're welcome and that was a lot of fun hope we can do this again at some point

Tanzu Talk · 🍕 Get more pizza.