August 15, 2023
Season 1, Episode 7
In this episode, Jon and Jay discuss best practices in authentication and authorization, emphasizing the importance of establishing Single Sign-On (SSO) from the outset. They delve into the challenges of authorizing and managing access for multiple teams, especially when dealing with the three major cloud vendors.
Guest Introduction:
In this episode, Jon and Jay are joined by Brian Ross, a technologist based in Edinburgh, Scotland. Brian has recently taken on a role at VMware Tanzu as an executive advisor, assisting businesses in deriving value from IT. With a rich history at Sky Media, Brian has extensive experience in building platforms, VMs, containers, and more. He brings to the table not just technical expertise but also insights from his journey in the IT world.
In This Episode, You Will Learn:
Themes Covered in the Podcast:
Quick Takeaways:
Follow for more:
Jon Shanks: LinkedIn
Jay Keshur: LinkedIn
Jon & Jay’s startup: Appvia
0:00[Music] start small and build up alongside your customers do not disappear for nine
0:07months to a year building some platform that’s going to change the world go change the world in a series of small
0:13steps [Music] hello Welcome to Cloud unplug this is
0:20season two episode 11. I am here with Brian Ross
0:25um Brian do you want to introduce yourself in a bit of your background yeah sure thank you so much for having
0:30me on first of all so hello everyone my name is Brian Ross I am a technologist
0:36working in the Edinburgh Scotland I’ve got a new role recently at VMware tanzu where I’m an exec advisor working with
0:43other businesses to help them get value from it previous to that I used to work
0:48at Sky media for a long long time building out platforms and VMS containers paths all that kind of good
0:56stuff so I’ve uh I’ve been out there I’ve got the scars and the gray hairs to to prove that I know some some of what
1:02I’m talking about so I’m looking forward to a conversation today yeah it’s great yeah it’s awesome to
1:07have you on as well and I think what what we’re going to talk about today is um a little bit of the industry the
1:13industry trends that are going on um Gartner obviously 2023 Trends has
1:18kind of been announced some of these things may or may not be new to some people
1:24um I think one of them was the platform engineering Trend um I guess it’d be good to kind of pick
1:30on some of these things and and give a bit of uh a bit of opinions on whether we think
1:36they’re good good things bad things um you know where they’re going to help businesses or not so I don’t know if you
1:41want to kind of kick off from your perspective on what you think about this platform engineering Trend that by 2026
1:4680 of businesses maybe forcibly have to go to platform engineering because technology is too
1:53complicated to do it any other way it’s how I’ve kind of slightly interpreted it and therefore you need to do something
1:58and that’s called platform engineering but I don’t know how you’re um thinking about it well you know I think um first of all it
2:06must be exhausting for people not an idea in fact even people in it the number of new terms we come up with to
2:12to try and explain how we’re definitely going to do it better this time we promise
2:17um so I’m somewhat biased in that I think I’ve been doing platform engineering for the
2:24best part of 10 years and that’s probably due to my background I’ve always been in kind of managed services
2:31and kind of customer facing rules so I think platform engineering for me is more a response to
2:38how large Enterprises have kind of failed to adopt agile and devops
2:44necessary like I think we’ve got the job titles but I’m not necessarily convinced that we managed to get all of the
2:50culture baked in there because I think it’s difficult at scale so I think that platform engineering and and link to
2:56that platform as a product are kind of topics that enterprises are talking about as a way to achieve the kind of
3:03values and goals we were doing in that that cloudy thing that we’re all tied up in a couple years ago
3:10and do you think then the reason for it is because the the
3:16traditional kind the traditional models which most organizations I’ve seen adopt
3:22which was like a bit more product project autonomy we’re going to give you a resource a devops resource they’re
3:30going to sit with your team you know we’re going to Sprint plan we’re gonna we’re gonna do all this and then but
3:35you’ve got that times however many teams all kind of independently going about things and then times that by say
3:43the level of different technology choices that you could make to achieve the same outcome
3:48um and then suddenly like oh why have I got 15 of the same types of products across my organization and like what’s
3:55actually going on and do you think that’s caused it where people like this isn’t what I thought devops was
4:01um or do you think it’s been something else um I I think you’re a break on both
4:07counts I’d also throw in public Cloud into that too in that we’ve gone through
4:12a period where all of a sudden we were no longer constrained by our operations team so we
4:19could have any infrastructure we wanted anywhere loads of services there’s this
4:24fuel all-you-can-eat buffet of I.T capability and as you say I think independent teams
4:31made different choices and then we reflect back we’re beginning to question why are it bills are so high
4:39why we’ve got maybe 10 20 100 different devops teams doing similar activities
4:45but seemingly completely different ways with different tooling so yeah I think for for me the big one is
4:52um devops is fantastic as a culture let’s bring people together so we can make better choices together
4:59um and that’s really successful particularly in smaller businesses I think when Enterprises tried to scale
5:05that out like it certainly worked you know if you localize that knowledge you can achieve lots of things and you shift
5:12left you do everything in Dev rather than waiting until it’s thrown over the fence and prod I think those are all fantastic but scale out to 150 teams and
5:20just the cost alone is going to be high so I think if we think about platform
5:25engineering it’s maybe worthwhile defining some of that and I’ll be honest that’s I’ve spent the last couple of
5:31weeks chatting with some friends in VMware trying to work out exactly what platform engineering really is and I
5:39don’t think we’re quite there yet the movement’s still going to taken off but it’s all about if you have a
5:45a hosting platform of some description that you’ve built and it’s all designed
5:51around consumption and empowering developer experience so that you can get
5:57developers getting to production faster getting rid of some of the operational
6:02treacle that we all seem to build up um and generally being able to be
6:07effective and be um impactful faster because I think this also ties back into some of the other
6:13problems we’ve had in the industry recently around um skills Gap and trying to retain good
6:18people is that we’ve we’ve almost kind of come the full circle we’ve we started
6:24off with developers not being able to get anything into production because there’s too many processes and we took a
6:29whole bunch of those processes out and let them run loose with a credit card on on Amazon or Google and now we’re trying
6:36to tell that back and say well how do we make you impactful but in a way that still kind of matches
6:42the corporate standards and we still have some control over expense and we still have some means of carrying on if
6:49you leave the company and that sort of thing so there’s quite a lot in that in that one topic but
6:54um yeah I certainly think it’s been a response to just trying to practice agile devops SRE
7:02in a more generalistic bigger company type way yeah and do you think because I have an
7:09opinion of it as well a little bit on I think you’re right the agile and devops movement I think was great for like Innovative
7:17thinking um because kind of like the boundaries were broken down at the same time I
7:22think it also allowed for a lot of Madness to probably ensue which is like you know
7:29some of that might have turned into really good high values outcomes but at the same time
7:35if like in two years or a year later you’re like okay well I need to go and get a another product because I have no
7:41idea about the risk of what’s just happened like I know we’re live but like I don’t really know what that means
7:47exactly to you know operational risk of what’s just being done and like you’re saying Talent moving around and all
7:54these other things my only concern a little bit with the trend like with anything is if you’re doing platform engineering
8:01like you’re saying defining it our company is going to have a clear enough vision for themselves internally about
8:08what that actually truly means to their business and how they equipped internally to
8:14you know represent what outcome they’re striving for because it could end up a bit like anything
8:20where it becomes just another technology play um and not really driving business
8:25outcomes and you kind of just going around in circles but just in different ways around the mess
8:32um don’t know what your views are but yeah so I I think that’s a really important
8:38thing so as you guessed with uh working VMware tanzu I spend a lot of time going out and talking to people about their
8:45container platforms and kubernetes and all that kind of good stuff it’s always really interesting though is that my
8:51opinion is if you’re using kubernetes that’s great why are you using it what are you looking to get out of it likewise if a
8:58customer says we’re not interested in kubernetes that’s also great but but why why are you not using it because there
9:04might be other industries that there are so I think bringing it back to business values always the most important thing
9:09otherwise it is just a technology play um it’s interesting what you’re saying about you know agile adoption and
9:16Enterprise I think that we give a really good shot but in many ways the
9:22fundamentals of business haven’t changed in a lot of the big organizations that I’m working with so when we talk about
9:29how things are funded they are still predominantly a project mindset we’re
9:35going to achieve this business goal we’re going to put this amount of money behind it we expect that project to
9:40finish by this date everybody needs to work on that critical path I think what we did then is we then iterated in
9:47little Loops to deliver that product and in the way that was the best way but the financing model was still the
9:54same it means at the end of that project somebody is given the keys to then look after that platform that product the
10:01tools the infrastructure I think with um platform engineering what we’re
10:06trying to say is we need to address the underlying business model and saying rather investing in those large kind of
10:14Monolithic projects we kind of want to invest earlier in the cycle
10:19for a platform that will deliver capabilities that we’ll use later
10:25and that’s I think that’s a really difficult sale from any big organizations because it is a mind shift in how they plan their annual finances
10:33right like all of a sudden you’re talking about building something that you’re not planning to use straight away
10:39and in itself does not deliver Direct business value but when you come to deliver 10 projects
10:46that all need something very similar they can whiz right through Straight to production and you’ve got the time to
10:52Market benefits that you’re just not going to see in a kind of Serial project delivery type way does that add up with
10:59what you’ve seen as well or is it it has different yeah I think
11:05you’re pretty much spot on I do think there is legitimate rationale probably
11:11behind why that happens and that is like probably very
11:17gets quite into like the capitalist mindset of like share prices and you know all these other things that it’s
11:22very hard to like control guts systemically you’re in a system at that point it’s like how much can you influence because you know how much are
11:28we going to invest what’s the bottom line how do we give us how do we give assurances back to shareholders in the
11:34end you know it all rolls up into something everything is rolling up into something somewhere so I think sometimes that can be quite
11:40hard for Enterprises where that is a reality versus a smaller company where it probably isn’t and you’re all about the go to market how I’m gonna so I can
11:47kind of see why that institution the system that you’re kind of in ripples down and then
11:53actually how you’re supposed to deliver them becomes a bit more constrained but
12:00I think there is still maneuver in doing it well I I it’s interesting that you said about
12:06you know not necessarily because of what you’ve got to do because you’re basically designing for some element of
12:11scale so therefore the results might not be yielded what would you suggest for de-risking
12:18though because I guess what you wouldn’t want is then 10 projects to start and then none of them can go live
12:25is either extreme end result of what you’ve built right and you’re like oh yes sorry about that um yeah so I guess
12:33it’s the what would you be advising there just so you didn’t necessarily hit that problem
12:40later so so I think the the sad truth is I’ve I’ve worked with a number of large
12:45organizations I’ve had exactly that problem in that the infrastructure team has gone away and built a platform for
12:53developers and it’s going to be magnificent and they’re a little disappointed that once they’ve built it developers refuse
13:00to use it and of course my first question is always well who did you involve in the design
13:07process like surely the developers you were building it for them so you asked them what they wanted right
13:15and then there’s usually some blank faces you know no no we know what they need
13:20you know that’s that’s the first mistake so as always I’m I’m strongly of the opinion that most of our I.T problems
13:26were solved years ago or the new ones are relatively easy to solve the bits
13:31where I think we need to really work on agile and devops and devsecops and now
13:37platform engineering’s still the same it’s the people it’s communication it’s trying to work together in getting an
13:43organization pooling in the same direction so I would say if you’re thinking about building anything for
13:49someone else the first thing to do is involve them in that process I think as well you need to partner with them so in
13:57my previous role it was really important whenever we’re trying to build some new functionality we would try and find an
14:04internal customer that was looking for that and then that meant that we delivered together so it kept us honest
14:10which I think is important in it as well to make sure we actually get it done you know we’re not building this Rolls-Royce
14:15thing what we kind of need is a you know a go-kart that we can then add bits onto right
14:21um yeah let me just as well at the end of
14:26that at the end of it it means as well that you don’t have to go market yourself the customer that you’ve just done a great job for is going to help
14:33you in that process another thing I think I’ve found as well is that measurement of your success is really
14:38important so reporting back to your Finance teams and business saying here’s
14:43why we’ve been successful in what we’re doing here’s the time we’ve saved this project in rollout and it kind of about
14:50matches up to the to the investment we’ve made um and the more you can get that
14:56customer to do that on your behalf the better right and it’s really interesting once once one developer team talks about
15:03how quickly they were able to do something because the developers didn’t have to worry about infrastructure
15:08all of a sudden other developer teams go wait a minute why are we why are we
15:13building out infrastructure on Amazon which we shouldn’t have to do that and they’re absolutely right they shouldn’t have to
15:20do that so um yeah good news travels fast you just need to make sure you’re successful because bad news travels
15:26faster so if you’re building it you need to involve people without doubt yeah true and and what about the uh like
15:34operational end State because you know you can always be this is how I’ve seen as well is you’ll always be a
15:42hero at the beginning right when you’ve got new projects starting and everyone
15:47wants to get going and you’re enabling you’re giving the environments you know that can get access to things they’re
15:53deploying things and the constraints obviously are less because it’s about
15:59speed you’re at this you’re at that Tipping Point of agility so obviously
16:04at that time agility is seen as like the hero and then any other kind of more boring non-functional requirements that
16:11don’t really seem to benefit anything obviously or tangibly until they’re
16:16really needed like monitoring a login becomes very important when you need it
16:22um and then observability in general obviously and thinking about all those other things and security which also can
16:27be a bit of a for most people quite contentious um and a bit of a red feels like red
16:34tape but then where you end up as you’re iterating can always sometimes not end up in a great
16:41place either because you could have built a bit of a Frankenstein you didn’t know you were building Frankenstein right it didn’t you didn’t realize like
16:48an arm was attached to the ear and that the year was on the because you know you weren’t thinking about as a whole thing
16:53at the end um but as you zoomed out you’re like oh wow yeah I did put an arm on the head
16:59and you know the head’s kind of upside down you know a little bit like that but you don’t really think like that at the time so
17:05I have definitely seen that a lot but it’s a very difficult thing to also know
17:12that that’s what’s going to happen because you can’t see into the future um necessarily to know where you’re
17:18going to end up and also how much it’s going to cost to operate um as well right so yeah so
17:26um I wish I had you before uh VMware Explorer in Europe because the
17:32conference talk that I did was really around the subject it’s kind of awkward turning up to a technical conference and
17:37then not talking about anything to do with technology but it was exactly this it’s about when you’re building a
17:43platform how do you do that and what are the stages that you’re going to go through as you build it so I talk a lot
17:51about platform as a product so trying to develop something on an ongoing basis so
17:56first get it in your head that you’re you’re never going to be finished there will be no end state so to speak uh it
18:02will be a constant Improvement of what you’ve done um I think that it’s really important to
18:11manage your product so although you’re building on behalf of
18:16developers or whatever customer you’ve got it’s your job to kind of curate those
18:21requirements and make sure you build a product which will solve a problem for
18:26The Wider Market if you listen to every requirement you get in and Implement
18:32that very thing you might be really popular and it might be really quick to do as well and really pleasing from an
18:38internal perspective because someone asked for something you built it and now they’re really happy but yeah
18:44fast forward a year or two years and you’ll end up with this Patchwork quilt of functionality that doesn’t quite mesh
18:52and we all recognize in our day-to-day life right if we see a a product that’s
18:57just been put together badly like the usability goes really way down
19:02and we don’t like it anymore and as you say the early days of your building this new thing and it’s super exciting you’re
19:08getting things done quickly you’re solving challenges it feels great and then all of a sudden you kind of come to
19:15a hump where the platform you built is now beginning to Creak a little bit and you might need to spend some time doing
19:22some back end improvements or it may be that the problems that you’re trying to solve are getting harder
19:28because you had all that low-hanging fruit to begin with but but now it’s getting harder to solve so I think it’s
19:34really important to manage the backlog of work that you’ve got coming in and
19:40think about how much time you’re gonna um spend on new features
19:45bug fixes and improvements to existing features and back-end features that nobody is going to appreciate
19:52until six months nine months time and ideally it’ll be invisible by that point right they wouldn’t have seen that
19:58Improvement um I have certainly had moments where I’ve had the engineering team going to
20:05come to me and say you know Brian we’re in trouble we’re falling behind we need to do X Y and Z and we’ve had a a week
20:12of what we call let the world burn and my job if you like as the product manager was to to hold back the flow
20:20from Project managers developers support all of those things just to give the
20:26team time to to work on that back end and make the improvements they need because I think it’s it’s really
20:32important again for them as well not to burn out that they can spend time doing things that they see is important if
20:38you’re just constantly adding more cherries on top of the cake it’s not necessarily a good thing right
20:43no it’s not it is right it reminds me of like a um when you just stayed holding back everybody like a Simpsons episode
20:49of people with the Torches you know I mean like all running towards the team like get them the book I was just
20:55running away and having to hide but um in terms of like the the when you’re doing this then if you were to try um
21:03this is a bit of a complex question it’s a bit off the cuff to be there but if you’re trying to
21:08budget for a platform how would you set a budget according to say projects like
21:16would you would you work out the overall budget as a percentile of the Project’s overall spend to say
21:24right okay well let’s like a tax or you know because I guess in some ways if
21:29that pocket if that pot of money isn’t constrained you know from the beginning then how
21:35you’re going to innovate is also unconstrained and that might leave you to like like endless resources or
21:42endless technology decisions you know and then list this and then this that because there’s no real defined
21:47you know there’s no constraint at all and something has to constrain people to then Define what you can and can’t do
21:53the parameters have to be set clear otherwise it’s just a bit of an open-ended book so I don’t know if
21:59you have advice on how you might work that out or not I don’t know if you’ve given any thought well the the spoiler
22:06alert I think is that there’s no there’s no one answer if it’s all um however I think you’ve hitting a
22:12really important part around the finance so the
22:18big thing is that if you are developing something of value then it should have a
22:24value and if you’re working with a a finance team to help fund that platform then you
22:32need to have a system there that they can see value coming back in so I think having a either a cost model or a
22:39recharge model or even just a showback model is really important um various ways you can do it I think
22:44you need to understand the mechanics of the finance in your company which is very boring but it’s very much worth
22:50investing in so if you’re a capital company where you have capex money but
22:56Opex is difficult to find then it’s probably best working with what you’ve got so you previously had these projects
23:03that got capex and then the spent capex on servers and
23:09engineer time to build something great work with those projects to get some of their money and use that to fund
23:16improvements to your platform the net result is still the same the project gets delivered but it’s the way that it
23:23got delivered is slightly different you’ve helped enhance the platform rather than just build a one-off bespoke
23:29solution or as you say if you’ve got enough projects coming through perhaps you’ve been at this for a little while
23:35and now you’ve got the backing well that’s great maybe rather than just if you like billing on those big projects
23:41that are coming you could apply attacks across all projects and that might be 10
23:46it might be 20 whatever it is whatever you think the infrastructure cost should be or it may be that you can
23:54um have a more elaborate system whereby you have a pay per use but that would mostly
24:01be in an organization that is opex-sled um now you’re talking about billing back
24:06to a business unit and and they’ll just have that like their their Staffing cost or their
24:11public Cloud cost or their software subscription costs it’s just another part but it needs you need to work with
24:18the the constraints within your business but I would say I think my biggest warning to anyone building
24:24infrastructure out there is that unless you are part of the kind of value chain
24:31and it you know if if you’re just a cost center then the 90 is just a a line on a
24:39spreadsheet that needs to be reduced at some stage and that’s really no one sees a value in the in the service that you’re
24:45delivering but if you’re able to be part of the project on a direct line kind of matter you’re like a steel thread
24:51between the the revenue that was generated by the project and the cost of the I.T to implement it that’s a much
24:58better place to be because now you can talk about the value of what you did not in terms of gigabytes per pound or
25:05anything like that but instead no no this platform now supports three five ten million pounds about the revenue
25:11every year that’s why it’s really important that we have the right staff it’s why we have to make sure we’ve got
25:17new equipment it’s why we should continuously improve it you know I think that you need to be part of the value
25:22chain not part of the the cost if it is just cost that you are being measured on
25:28only going to stretch you in One Direction and that’s not a good place to be unfortunately that’s the reality for
25:35I would say 80 of the Enterprises that I talk to is that it is a central cost
25:41that is budgeted for and that’s not great it doesn’t drive the right behaviors internally because
25:47you’re always looking to cut Corners to save money it doesn’t drive the right behaviors with developers or project teams because they just want what they
25:54want faster but you don’t have any money to do it faster so and they’re not interested in saving
26:01money and in fact if an asset is harder to get hold of like a server a VM a
26:06container people are more likely to hoard them and then you end up with Shadow I.T
26:12black markets of VMS you know I’ve I’ve lived through all of those things of well no it takes us too long to get a VM
26:19so why don’t we just clean up these old ones and reuse them you know guaranteed
26:24how things happen I think that’s a really smart idea actually the the switching the context to like how much
26:31revenue you’ve supported for Generations a much better optic than just how much
26:37you’ve cost the thing I have seen is when people start especially with platform engineering if it’s a project
26:42some projects have zero context to costs right so like hosting costs or Cloud
26:48costs or any costs right in general other than like people costs so what they anchor is like tnm you know I know
26:55roughly like what I’ve resourced it’s very tangible whereas there’s no real expectation so
27:02what you do see is because I haven’t got a comparison as in they’ve never baselined anything before
27:07um then any cost of them is quite a shock so a little bit like sorry what you know when they start to get the
27:13recharge or the bill and all these other things and then some other people were like but surely and I can go on the internet and maybe there’s like a Heroku
27:19thing or maybe there’s like some other thing and that can drive Shadow it was just like this is so much cheaper like why are we just not using this and it’s
27:26like not no context for you know how much it might cost in prod or you know
27:32the security implications or any of those things because obviously that’s not their domain all they’re looking at
27:37is a like for like you know you host they host this price that price right so
27:43they just kind of look at it in real simple terms so it can be hard to to get by in
27:48you know what if you don’t set expectations first on the value
27:54and then you charge later um and there’s no context to it that can also backfire I’ve kind of I’ve
28:00witnessed that a fair chunk you know behaviorally internally I I think it’s um like I was lucky I I
28:07was involved in helping build a small company up from the ground and I learned
28:12some really important lessons about money during that type and no matter how complicated the
28:20finance of anything is there is still a really simple equation that you should always apply to what you’re doing and that’s money in and
28:27money out right yeah and I think if um if you have a lot like a project that
28:35you’re going to deliver how can that not have an I.T cost associated with it
28:40because you’re seeing one of two things either you’re just gonna have to make do
28:46with what you’ve got which is terrible right because that means that you’re going to restrict the revenue your
28:52business can generate or you’re seeing that there is no there is no limit to how much I.T like I I
29:00heard a quote a while ago from a large customer that it shouldn’t constrain the business
29:06and on face value you’re like well absolutely apart from the business is already
29:13constrained in many other ways like the number of people that they have on the project the software that they buy to
29:20facilitate the project the other services that they’re purchasing in order to make that success it might be
29:26the marketing spend Etc why would I T spend not be part of that
29:33project’s accountability like when they’re putting their business case together why would it spend not be
29:39included there that seems really weird for me because again it’s the back
29:44this equation money and money out show me all the costs and I think historically it’s because it weren’t
29:50really able to tell you how much it cost but I think in this day and age of public cloud of platform engineering we
29:57should know how much it costs deliver what we do and we should be able to articulate why it costs that much or
30:03that value bearing in mind I am still willing to retire at any stage when a couple of things happen one of those
30:10things is I meet a customer that spends less money than they planned to in public cloud
30:15and I think that you know whenever we estimate the cost of doing something we tend to make it too low I think that we
30:22sell internal I.T Too Short sometimes like I know sometimes you’ll be slow and cumbersome but if we could improve those
30:29processes I think we can make it a really really valuable part of the business because whether we like it or
30:34not I think most businesses now are digital first um because our client base our digital
30:40natives they don’t really know any other way to operate yeah that’s true I mean it’s funny
30:46because I had I was at a round table and and we were talking about it was a
30:51some people were saying that inside their customers because the customers have gone into
30:57cloud and then there’s like been a recognition of the cost implications that now they’re like oh my God like we
31:04should have stayed on Prem which I kind of understand but at the same time
31:09like the things you were just saying even when you buy a car you don’t really know the fuel costs like it’s not like
31:15you’re like right I reckon it’s gonna cost me like some things are just so hard to equate but at the same time
31:20you’re not going to say right horse and car this is ridiculous I’m going back to horse and car I think we all should you
31:26know I mean it’s like you know it’s kind of like an obvious thing it’s like you don’t behave like that but it’s kind of
31:31strange when it gets to technology people for some reason can like revert back to Old behaviors
31:38and be like better the devil you know but it’s like yeah but you’ve also got to innovate and you’re gonna like if
31:44you’re doing that somebody or the majority definitely isn’t and the probability of them succeeding is probably higher and it might just be
31:51that you just need to find the right people to manage that problem um maybe or look at alternative ways of
31:57managing the problem at least rather than like maybe reverting backwards and kind of running back but yeah
32:03100 I’ve never felt so old in I.T than when public Cloud was coming about first
32:09of all public cloud is an amazing feat and also as well we should all be thankful for it because it made IIT
32:15interesting and sexy again no one really paid attention to it before public Cloud but now we’re back in the front seats I
32:21think we’ve got a lot to thank there um I would say though that I I feel like an old man shouting into the clouds for
32:28quite some time because it’s been businesses moving everything to public cloud and claiming that they’re going to
32:33save money and again I’m it it doesn’t always make sense to me like the money in Money out part I’m like but we’re a
32:40capex company how is that going to work that’s going to affect our bottom line um
32:45likewise yes I’ve I’ve met other companies as well which I know well it was too high so we’re moving everything
32:51back no no there was some good things there right I think the challenge I’ve
32:57seen a lot of time is that people adopted public Cloud as a and expect it to be a silver bullet to their
33:03operational organizational and cultural problems they thought if they did public
33:08Cloud then they would be devops and groovy and it just didn’t happen at the Box in many times what I’ve seen is say
33:15there was a process to get from Dev to prod public cloud empowered people to just go ahead and
33:23build things themselves and what they ended up doing was first of all removing some people out of the
33:28process so we don’t need to speak the network team or the or the storage team anymore we’ve just got instant access to
33:33those Primitives that’s phenomenal but it also meant that some things like you mentioned like security
33:39software compliance audit those steps got missed as well so some of that
33:45velocity we got from public Cloud was not as improving it was as avoiding and
33:50I think that’s where we began to go wrong now we look back we go oh my God we’ve we’ve we don’t have any
33:55architecture we’ve got 15 different tools for logging we’re not even sure which SAS companies were using anymore
34:01but we’re pretty sure our data’s somewhere um we’ve had an S3 bucket been exposed
34:07and now our customer data is out there that should never have happened like those are all problems that can
34:14happen when we skip steps I think now with um with platform engineering saying look public Cloud’s not evil private
34:21cloud is not evil process Improvement is where we should all sit so we want to do all the checks
34:26and balances we want to see things like compliance and security be part of our feature set not a gate to be crossed and
34:34we still want that to happen quickly like the the example I always talk about with the things like that is um mobile
34:40banking like two-factor authentication can sometimes be a pain the certain apps
34:45that will make you blink in the carrying move your phone further away or closer or text messages but would you use a
34:51mobile banking app that didn’t have two-factor Authentication right so you need to be able to trust things that you’ve got and I think in
34:57this day and age now trust in applications is becoming more and more important for any industry because the
35:04risk of Cyber attack identity theft data loss is just too big so I think we need to bake all those
35:10things in not throw the baby out with the bath water um yeah yeah and use the mixture of
35:15public Cloud private Cloud SAS all other things where it makes business sense
35:21fundamentally bring it back into that business case let’s talk about the best way of delivering it not just the way we
35:28know or what didn’t work last time so if I’m sorry apologies if you’re then
35:34thinking about the platform engineering team so you’ve got a platform engineering teams
35:39let me let me create a bit of a scenario I suppose so if you’re an Enterprise
35:44the thing that I find very interesting and also just quite
35:50nice in the industry at the moment I’m sure there’ll be Enterprises that will hate me saying this but
35:56cloud has created a Level Playing Field right so I could have X number of
36:01credits and I not even have very much money like generally I might only have like 100k seed investment or something
36:08I could probably deliver on some service pretty radically fast and work with the end users pretty
36:16quickly and disrupt a market um with not huge cost implications and
36:24and also I don’t have the baggage of an Enterprise either right so it’s like I don’t have to deal with bureaucracy I
36:30don’t have I’m not institutionalized I don’t have to worry about you know your processes and all the other things
36:35so you kind of have that track that’s going on in in the markets and the markets are getting disrupted because of
36:41that and the clouds had a lot to do with that right because the and I think that’s what makes it quite incredible I
36:47think that gets overlooked by accident people just think markets are just disrupting and it’s like well they’re
36:53disrupting more because of the accessibility of being able to deliver quickly on something
36:58um but then you’re an Enterprise and you’re not that so meanwhile you’re kind of
37:04like plodding along yeah I mean and I I kind of Imagine I see and this you can challenge me if
37:10you don’t agree but the platform engineering teams should really not to
37:15sacrifice in any of the security or anything else but you’re in an Enterprise so all those things matter and the matter to any company anyways
37:21it’s not that you’re going to discredit them just because you’re a starter but they’re all important because it’s
37:27customer data and customer support you know so you could still be fined Etc but um
37:33but the job of the platform engineering team is to make that accessibility as as disruptive
37:39for that business as much as any other um and you’re there you know then you’ve
37:46kind of got the ambition of the goal but you’re all there to do it in a in a way that is equally as cheap in some ways because
37:55you don’t want it to cost more with more people people take time to hire anyway so that isn’t fast you know you’ve got
38:00to go and hire 15 people to that’s not a fast execution plan right so there is
38:06elements of like a process of elimination for your strategy that you’ve probably got to work through where the market won’t care necessarily
38:15about your business right it just the consumers in control
38:20in the end and so you’ve kind of got to recognize that and if you’re not meeting
38:26that customer in the end then you’re kind of going to get left behind like it or not it’s kind of the reality
38:32but that’s how I look at platform Engineering in the end is if that’s the circumstances you’re in I mean there’s loads of other benefits but I kind of
38:39just think most my instinct is most markets are going to get disrupted I don’t think there’s really any now given
38:46the breadth of cloud services that aren’t up for grabs and it’s more a matter of time than anything else so I
38:53kind of feel it probably becomes one of the most important aspects as an Enterprise is to respond
38:59that’s my personal view obviously people could disagree with it but that’s what I kind of witnessed
39:06I feel like I’d be disagreeing with you more just just for uh just for clicks and Views yeah I think the the interestingly hit
39:15on there is that the only unique thing about a small business that you
39:20mentioned was that they don’t have the same baggage as a large business
39:26isn’t it interesting that’s not an I.T problem yeah it’s definitely not an I.T program yeah so I would say that you’re
39:32absolutely right I think the benefit of um like what we’re trying to achieve with devops and all these other things
39:37and agile was to make big businesses faster and when I say faster I mean the
39:44velocity time from I’ve got an idea about what I want to change in production and I want to get there like
39:50the one line change is the metric that I use in a lot of businesses or the other one is you’ve just got a brand new
39:56developer through the front door how long until they provide their first pull request to your production
40:04um git Repository and those numbers can be frightening really frightening whereas the teams I
40:10ran we took great pride in congratulations it’s your first day like your outcomes are really to know where
40:17to get food where to go to the toilet oh by the way you’re gonna need to push the prod like it’ll be a small change it
40:22will hold your hand but you will be pushing to prod today just to get that out of the way um yeah and it just sets you off on the
40:28right on the right path um again if we think about and again I feel
40:34really old but if we think about platform engineering being a broker for I.T services
40:40and the reason I feel embarrassed is because I was selling the same message a decade ago saying well
40:46virtualization was kind of like a broker for compute and storage not so long ago yeah yeah so it’s kind of like the same
40:52yes people selling but now I think again platform engineering is not just on-prem platform engineering is how do I make a
40:59developer’s job easier irrespective of where they’re running and it is where technology now has
41:05helped us so if you want some really complex high-end service like you want
41:13to you want to do some um voice recognition or a huge machine learning
41:19lab that needs to do some massive learning piece as an I.T company we would have to buy a whole bunch of
41:24Hardware not so long what we’re trying to achieve that whereas now like well no no we’ll just we’ll just get that
41:30from a public cloud provider but we don’t want developers going with a credit card we’d like to make sure that we’ve got that as part of our
41:37service catalog so they can get it easy also we don’t want to be in listing 10
41:42companies to do the same work we want to have one contract and that way we can make sure the t’s and C’s are all
41:48correct Etc because it might be customer data that we’re shipping into that service in order to do translation for
41:53example um so the other platform engineering is an opportunity for us to be Brokers so
41:59it’s not about being anti-public cloud or Pro public Cloud it’s just about saying look how do we how do we get rid
42:06of the of that trickle that you’re having to Wade through in your day-to-day job get rid of the toil
42:12um and I think as I.T Engineers we’ve got an opportunity to kind of prove our
42:18value again and it’s not it’s not in the depths of kubernetes it’s about bringing
42:24our heads up and talking about selling and marketing our services and being really clear about what we do and what
42:30we don’t do it’s about changing the way we do customer Journeys and customer support so you know I think it tickets
42:38will be here for quite some time for things like production downtime where you have to have an audit of what
42:43happened when but if somebody just wants advice on how to do something better
42:49then you know slack teams or some kind of chat channels much it’s gonna be much better that so yeah I think um customers
42:56can no matter who your customer is I think journey is more important than technology and everything we do has to
43:03be about the people involved in the business not the not the the stuff we spent years in universities and other
43:10companies training in right yeah totally agree so I guess
43:15if it were to wrap up a little bit then I guess two two questions for you
43:21um one is if you’re thinking about creating a platform team what are like maybe the
43:30top I don’t know how many you want to list but I guess not too exhaustive like
43:36maybe three to five things that you’d recommend um somebody to
43:43I guess anchor themselves within just to make sure that they’re actually gonna align themselves to the
43:49business properly like whatever they’re gonna do making sure it’s always anchored in some value somewhere else and it isn’t just a
43:56bit of a an engineering like you’re saying a bit of an educational thing or an engineering exercise for engineers or
44:02you know all the markets saying this so we should just do it and it’s like you know let’s get a team together and just
44:07do what the Market’s saying so just kind of avoid that and then second it’ll be how do people find out so if someone’s
44:14hit listen to you how do they find you you know they’re like actually I really want to go and speak to Brian he knows
44:20what he’s talking about is probably the final for the final one for that okay well that last one soon sounds like
44:26a a scenario is not going to happen but you know we’ll we’ll put it out there anyway um so uh number one we’ve talked about
44:33already you have to engage with your customers be customer Centric
44:39um and that that involves the finance part as well but you need to be part of the
44:44value chain not just an I.T service so work work all all your efforts over there
44:52number two I would say start small and build up alongside your customers do not
44:58disappear for nine months to a year building some platform that’s going to change the world
45:04go change the world in a series of small steps and none of this should be new right we’ve been talking about this for
45:09years through agile and devops and everything else um and I think the third is
45:17is don’t get lost in technology I think as an I.T organizations we need to spend
45:23more time marketing what we do thinking about what we provide not as a capability like we
45:31can do anything you want with Linux instead think about how you’re going to package that up into being some kind of
45:36consumable product like we will give you a Linux box with a configuration that
45:42meets all corporate policies that you can have route access to and will be here 24 7 to support you in whatever
45:48you’re doing that’s a very different proposition make it really easy consume so those are the three things
45:53work with people start small build up
45:58and um think about the product that you’re selling and that means by the way you’ll
46:04need to go and research marketing and some boring subjects like that but when you get into it I think it’ll be really important
46:10as for contacting me um I’d say obviously I have the sales pitch obviously is a good times website
46:15there’s lots of stuff on there about all these types of challenges but if you want to reach me personally you can get
46:20me in on Twitter or LinkedIn you’ll find me on most social platforms either as Brian Ross or Brian Ross UK
46:28um but yeah by all means reach out I’m more than happy to talk about the stuff I’ve been doing it for a good number of
46:33years and I think it’s we’re a really interesting time we’ve got a an inflection point to change the way we do
46:40things so now it’s time yeah that’s great it’s been awesome chatting to you as well and actually
46:45just speaking about these complex probably it’s good to kind of make sure that we uh keep the trends honest
46:51without kind of getting too uh getting too carried away so it’s good to kind of bring them back down to earth a little
46:57bit I’d offer advice but it’s great so I think you know obviously reach out to Brian as well if you know we want the
47:03experience of others to kind of help you and guide you so it’s fantastic but thanks for joining as well and chatting about this it’s been really really good
47:10thanks for having me on really appreciate it
47:23thank you
Part of the Appvia family.