August 14, 2023
Season 1, Episode 6
Richard Smith from Codification.io joins the podcast. He has helped build several enterprise cloud-native platforms, often modernising legacy applications with the Cloud. The key player involved in the DevOps team, and especially during a transformation, there are some everyday struggles. Richard shares the common DevOps struggles that he witnessed with us.
Guest Introduction: Richard Smith is a seasoned professional from Codification.io. With a rich history in IT, starting from 2001, Richard has been instrumental in building several enterprise cloud-native platforms. He has been pivotal in modernizing legacy applications with the Cloud. Over the years, Richard has witnessed the everyday struggles of the DevOps team, especially during transformation phases. His journey includes significant stints at the London Olympics, HSBC, and HMRC, where he played crucial roles in their cloud and DevOps transformations.
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
0:00locker and kubernetes it’s different underlying providers right so it’s never going to just run in multiple places and
0:07you know you’ve effectively got multiple data centers with complete different providers [Music]
0:23hello Welcome to Cloud unplugged this is season two episode eight I am here with
0:28Richard Smith from codification I’m John Shanks uh Richard do you want to
0:33introduce yourself yeah I’ll just give you a part of History John so uh I started working in
0:39it back in 2001 originally as a technician and then I moved into infrastructure engineering doing things
0:45like setting up data centers with VMware in 2012 that’s when I went became an
0:51independent contractor so I did site reliability engineering for the London Olympics
0:57um I helped startup go to Cloud basically from a from a
1:03software product to a SAS product uh helped a housing company get off their
1:09physical data center into private Cloud so third-party Cloud um and about eight years ago that’s when
1:16the assignments became kind of around Cloud native and devops so I was the devops lead for Department of work and
1:23pensions and I’d got their first microservice live known as the pension tracing service that same year 2016 was
1:32also when I joined HSBC as their Enterprise architect for platform so that was also about Cloud so AWS this
1:40time um the cloud native as well about exploring microservice style architectures and then the last project
1:46that I did before end of being full-time codification CEO was hmrc so that was
1:53their roll on roll off uh project which is better known as the brexit platform so I did the
1:59architecture for that um and all of those projects were supposed to be about
2:05agility and devops and I could see these kind of recurring patterns of failure uh
2:11patterns of success as well but but you know lots of the same mistakes being repeated so that’s when I decided to to
2:17set up codification so I had the first employee back in April 2019
2:23um and that’s what I do now so now we work with Enterprises with in-house engineering teams and help them on the
2:30same kind of Journey adopting cloud cloud native nice and and the challenges that you’re
2:35helping customers visit them at like current and at the moment and previously that’s all
2:42what is the type of constraints that you’re more focused on is it around modernization or
2:49applications itself or is it more like Cloud infrastructure automation or
2:54um or a bit of all of it is it all of those things I mean it it is a bit of all of it
2:59sometimes when when we’re kind of very heavily involved with the client but um mainly platform engineering is is the
3:06sweet spot so it is around like adopting kubernetes mainly but of course that
3:12then breaks into CI CD it breaks into the whole Devon Ops and and breaking
3:17down the organizational silos and then you know it bleeds into everything else where they don’t necessarily have the
3:24talent in-house so training the talents hiring the talent onboarding them
3:30um showing them the way all of those sorts as well right I see yeah that’s really good so the the stuff around the
3:37talent side is that is that the biggest construct obviously kubernetes I think
3:42is a big uh is a big topic obviously um and you know we’re in that space too so
3:50we kind of see how complicated well it is just complicated and that’s just factual what it’s trying to do is
3:56complicated um so is that what you’re seeing quite a lot of the challenges is people struggling for adoption and kubernetes
4:03it’s it’s harder than it they thought it was going to be or is it yeah definitely definitely so it’s the it’s the
4:10it’s the lack of Talent on the market for fast follow companies so of course
4:16the the very large tech companies don’t struggle with this as much of course but
4:22um outside of those I think there’s a shortage of talent for one thing and
4:27then it’s precisely what you said it’s very easy to underestimate you know in some ways it feels like Cloud native is
4:34is like an order of magnitude jump up from from virtualization and complexity so it’s very easy for for organizations
4:41who’ve done relatively okay with provisioning data sent as an in-house application development
4:47um to assume okay it’s it’s that kind of paradigm shift and I feel as though from
4:53what I’ve seen over the years that it’s that it’s much much harder than that and yeah it catches the complexity catches
4:59people off guard and what what do you what do you think is driving the adoption of kubernetes is more like the
5:05market or is it technologists coming in representing the
5:10need in the business and then trying to implement I guess what do you feel like it’s people that come in that then drive the
5:17adoption of kubernetes in like they make the decision for the business all the business is recognized the value of the technology for what
5:24they need or do you still think it’s a bit disconnected like the Technologies still focused on the technologists
5:29without the true value of understanding of why kubernetes
5:35um I think it’s a bit of like there’s a bit of that where the the Tails wagging the dog where the implementation technology
5:41kubernetes or or containerizations chosen first but I think
5:47you know early early days I guess like 2015 it was this agnosticism idea of
5:54like you know they’d experienced customers had experienced vendor lock-in
5:59before and were terrified of that um and so that was like can we build on an abstraction layer can we be you know
6:07multi um cloud or tool agnostic or whatever so
6:12I think that was driving it earlier now I actually think and I don’t know you know we’ll get into
6:17that probably but I don’t know that I fully agreed with those those sentiments but I think now it is
6:23it is kind of that there’s a pressure from the customer to be Omni Channel or to
6:30release software quickly and actually to achieve that you do need you know microservices and that then
6:36demands Cloud native approach which then kubernetes is the only game in town but actually when I started doing this sort
6:42of thing there was other valid options you know swarm was quite quite big at the time so we were doing a lot of swarm
6:47stuff so I don’t think it’s kubernetes as the driver uh but probably like microservices style architecture and
6:54needing to independently deploy services to do Omni channel is is probably what it is now yeah so like more
7:02containers it themselves unlike being more portable and slim which more developer friendly really than a VM yeah
7:09um I mean you know devs would never really be spinning VMS up or doing vagrant that was always like
7:16infrastructure exactly related wasn’t it and then suddenly you could then actually
7:21start packaging up your application into something like Docker and then that became easier
7:26for developers and then kubernetes came out later didn’t it really because it was just Docker at the beginning mostly then
7:34kubernetes came later mesos Marathon swarm then obviously kubernetes
7:39so but yeah that’s what we saw people successfully adopt containers
7:44um probably not true microservices but but you know containerizing to do lift and shift
7:52so finding that relatively doable in-house and then a lot of our
7:59customers came from those that then got stuck either trying to get it into production or they’re in production and
8:05trying to scale it or even more often the kind of in
8:10production at some kind of scale but can’t tell when it breaks why it’s broke so the observability stuff and stock
8:17tracing and that sort of thing so I think it all comes back to lack of platform engineering Talent so it’s good
8:24development Talent maybe good system engineering but that kind of crossover
8:29skill set is is missing and still now right yeah I agree with that I I find it like
8:36platform engineering is kind of like funny because the term could have been around probably for a
8:43while but it’s then either maybe like the central model versus disaggregated model so it’s like
8:49depending on how you perceiving it if you’re just having people in projects and you probably won’t call that
8:54platform engineering you just call up like devops you know devops embedded in a team delivering yeah with the
9:01application team but then the minute you kind of rolled into something Central so actually that model there’s not
9:07enough consistency or oversight or um you’re kind of relying on
9:12documentation and standards and all these other things to try and cohere some consistent approach or some
9:18security model and then kind of centralizing it was a better option because at least it’s in one place like
9:25you’re doing everything in one place yeah but then that’s risen to the concept of platform engineering is in
9:30like engineering everything in a central place but
9:36the term platform obviously is so used that it became you know like the cloud to platform isn’t it you know all those
9:43like platforms everywhere and platforms as a service all these different phrases yeah so it is quite confusing I’m not
9:51sure whether people do understand what platform engineering means it would devops suffers the same curse
9:57doesn’t it that’s true yeah devops seemed to me to sort that start out like
10:04agile so it was a culture culture of organic automation or a culture of
10:10breaking down organizational silos um that sort of thing and then you you
10:16know I was called a devops engineer at one point so that’s kind of an Andy pattern in in my mind so you know that
10:23that term also gets gets misused like devops engineer is that should that be a
10:28should that be a thing or you know but then as like you said there’s this kind of Federated view of where you have
10:34someone embedded in the team and sometimes they call that site reliability engineer but but
10:39whatever yeah I think I think for us we use it because that’s what we’re doing so it fits with our organization because
10:47you know we are specifically building a platform whether it’s whether it’s a cloud Landing zone or it’s a kubernetes
10:53platform or whatever so that’s what we’re doing I think it’s a separate question of should non-technology companies that are
11:00you know Finance Company becoming fintech or health becoming like should you should you have that you know either
11:06of those Concepts should you have a devops team or a platform engineering team I think probably no
11:13um but but we need one because that’s what we’re what we’re typically being contracted to do yeah yeah that makes a
11:19lot of sense it is very confused isn’t it because there’s loads of um people brand themselves of different
11:25things I don’t I don’t think a lot of people can keep up with the terminologies and the trends because obviously it’s rapid isn’t it and
11:32there’s so many different paradigms on top of it like you’re saying Sr read obviously born out
11:37of Google is mostly about scale you know they were high scale so you can kind of expect the need to be reliably
11:45engineering um for the scale that you’ve got to so it doesn’t go down and you can perform well
11:51and then devops obviously it’s more of a methodology than a role got coined mostly by
11:57um recruiters maybe in truth and then people were like branded as an engineer
12:03um and then they’d be kind of there was a period where there were two like people kept using them interchangeably like oh I’m an SRE but they were kind of
12:09just doing Engineering in a project maybe um so it still really devops because it
12:15wasn’t that they were doing reliability engineering yeah and platform engineering happened at the same time
12:20and then platform engineering with devops and sres because some companies have like all of them on the go yeah
12:27they like different shapes and sizes and it is quite complicated to know who’s doing what what are the roles and
12:33responsibilities of everybody absolutely um do you want to give your opinion on what like the the value of
12:40platform engineering is versus devops versus SRE that you see in customers just to
12:46clarify things a bit I mean I think if if our adopt this
12:52dance of the customer so if the customers let’s say they’re a software Enterprise or or a bank trying to
12:59release digital banking products or whatever right they should want to concentrate on building the product and
13:06so I think what you want is that to just be abstracted away and be self-service
13:11now there’s a there’s a road map there’s a transformation to get to that point
13:16but I think as a as a Target State that’s where you want to end up and
13:22therefore should the organization care about all of these things so this is what I see on
13:28a on a typical transformation program they’re trying to get the grips even still with agile you know a lot of people are still waterfall mindset
13:34prints too service management so you’ve got that that they’re trying to understand they’re trying to understand
13:40the transformation strategy the vision from the organization itself they’re trying to understand now this devops
13:46thing and like you see maybe maybe even two or three other competing things and I think that’s a bad idea I think
13:53probably under the under the cover as you want you know you want automation so you want
13:59you want that part of devops you probably want you know the E you know
14:05eat your own dog food concept of if you build it you look after it in production so I think you want devops culture part
14:12uh I think you need platform engineering to get there right you know in a in a typical Enterprise it’s going to be one
14:19to five years depending on how much budget you put up front in how serious you are about and how many
14:24mistakes you make along the way so but that should be transitional in my view and then you should end up with
14:31something that’s self-service for the developer what what do you think
14:37um yeah I totally agree with that I think I’m not sure I think what’s difficult
14:43probably for customers is knowing when you need something like what are the markers that tell you when you need to
14:50change because I guess every company is so different yet the terminology is
14:55uh kind of singular they don’t they’re not conditionalized you know as it’s just like platform Engineers platform
15:01engineering devops is devops but there’s no condition of like which one do I choose if I’m a small company am I going
15:06devops straight away or like do I hire do I hire a lead in a small team who
15:12understands cloud or has maybe got some devops principle knowledge you know I think that’s quite hard because there’s
15:18no there’s no real good guidance probably on like what it means and then large
15:24Enterprises who are probably straddling really complicated
15:30um complicated environments and that’s nothing to do with the strategy that’s just history like you’ve been around
15:35long enough yeah to have to have been on a data center because you predate Cloud right so that’s just a historical thing
15:43and so then you’re into that like a totally different type of business and so what they’re trying to do might be
15:48completely different their needs will be completely different to say a really small exact coming that’s all in the
15:54cloud so I do think it is hard for people to know what does it all mean to me you know
16:00like my company when do I need these resources and when’s a good time
16:06to invest um and in the in that as a principle
16:12because most of the time I think people are getting people in when
16:17they’ve they’ve got the problem rather than you know then they’re in a bit of a mess so then they’re seeking
16:23help because they’ve got them to land themselves in a bit of a mess rather than understanding how they’re going to evolve in
16:28yeah it’s a delivery but yeah and you see this like cargo cult thing of you know this organization
16:36wants to do it because it’s competitors are all doing it without really understanding why I think that’s yeah
16:41that’s the tail wagging the dog I mean for me it should come from like a clearly articulated business strategy
16:48and then whoever’s doing a strategy in it should should articulate you know
16:53this is the vision this is the strategy and then it it can follow from that if that’s clear right if what are we trying
17:00to do are we are we trying to reduce you know is our main priority to reduce time to Market is it cost saving you know
17:07what like what is it because that that might dictate a different path are we
17:12wanting to bring all of the incumbent staff suppliers whatever on the journey
17:18that that would obviously tell you a lot about what you should do um and I think where it goes wrong is if
17:24you just kind of go to the market start hiring for buzzwords then you’ve been in
17:30quite a mess because you don’t understand why you’re doing it right yeah exactly and it’s the same with all the
17:36um like you’re talking even about multicult we’ll come on to in a minute but even those things strategically aren’t
17:42necessarily tied to business outcomes you know like what what’s the driver of
17:48multi-cloud exactly or you know people saying hybrid and it’s like right our strategy is hybridating
17:53like how is how is that an actual strategy because it’s not business driven I mean it’s like what is the
17:59outcome you’re even what do you get at the end of hybrid um and like what was the cost did you
18:06save money did you not save money like what was the objective actually so there’s all these things that are term strategies but they’re not really
18:12strategies because I agree with you I don’t I I don’t see how that can be a you know it gets termed as a strategy
18:17but I’ve never seen I’ve never heard a CIO or a strategist an organization articulate why that’s a strategy
18:24ultimately when they articulated it’s a tactic right it’s not a strategy we’re
18:29on Prem and would like to be fully in xcloud and so we’re hybrid well yeah that’s just the state of being rather
18:36than a than a strategy right so yeah but I I you know think maybe the same thing
18:42could be true of of having multiple public clouds as well I don’t know what you think about that but but is that a
18:49strategy is that a meaningful strategy that you’ve heard articulated uh from
18:54finances from financial institutions I mean over here in UK obviously have the pra which is like the regulatory body
19:01I’m not sure they say you need to you has to you have to be multi-cloud I
19:06think it’s then a bit like anything Chinese Whispers I think they say something around the resiliency there
19:13um and obviously like data recovery perspectives and uptime Etc like or if
19:19the cloud was to go bankrupt that you know for whatever reason um that you’ve Diversified enough to
19:26like reduce the risk I don’t think they ever say that then means you have to multi-cloud
19:32um it just means you need to get out a jail card right basically I think is really what they’re saying but
19:39I think just what you were saying before on the effort it can take some companies to
19:44even do one and the lack of understanding the hiring the talent and everything else I mean my view I’m I’ll
19:52ask yours as well is like it seems almost impossible to do two clouds after
19:57you’ve done one I don’t know you’re agree with that and that maybe puts like value
20:02and reward isn’t it it’s like is there value in it what’s the reward of doing both
20:08um and is it worth the effort in the end so I don’t know I’ve never seen anyone pull it off as in like they’re actually
20:16achieve the Strategic aim of being either on the one hand getting the best
20:21out of both clouds so sometimes that’s the use case right of like we want to use Google for developer experience and
20:28AWS for maturity or AWS maturity and as you are to be close to the data lake or
20:34so you sometimes hear that as a as a reason and I’ve never seen that realized
20:40and then the other reason is you know resiliently and availability over over monthly for that reason and
20:46I’ve I’ve never seen that either so I think you know these surveys come out and they say how you know there’s more people
20:52doing multi-cloud than single but I don’t I kind of imagine that there’s a
20:58very very small subset of those that are actually achieving that I mean yeah
21:04yeah I totally agree with that I have seen the using the right card for the right job but I think you have to be quite mature
21:12to know how to do that as in to have achieved
21:17I guess if you’ve if you can think about the infrastructure as separate from application in in some ways that if you
21:23can think that the application has dependencies outside of like how I run as a workload
21:29um and in the app dependencies could be like a database or an object storage or something that a
21:35much more simplistic than say infrastructure and networking is then if you can make the infrastructure in the
21:41networking very service driven as in like a service then it can work but that takes a lot of
21:48design and upfront thinking it’s the trade-off though isn’t it like it can work but
21:54I I think unless like you said Regulatory Compliance maybe could be one reason if but uh
22:02but you know I didn’t I didn’t see that as a as a dictat when I was at HSBC when they were doing it so
22:08they wanted agnosticism because they’ve maybe feared competition or other things
22:14like that right but so so maybe if you’re so large maybe if you’re in the top 500 largest organizations in the
22:20world it can make sense maybe if there’s true Regulatory Compliance so Middle Eastern banks for example so they they
22:27maybe have a real case where they can’t be on one American public Cloud um so that that probably makes sense but
22:34short of those the challenge is going to be what we were talking about before right it’s going to be the sprawl and
22:41the lack of talent to be able to do one so who which mid size or even large but
22:47not super large organization has the luxury of you know the trade-off right to try and do too is gonna essentially
22:54mean that you’re gonna take three times as long to accomplish the same goals as just focusing on one I
23:00would say yeah just a minute I’m gonna have to start terms and of
23:06um can you talk about like compliance um and the security aspects of things so
23:13when you’re promoting platform engineering are you seeing that as a better
23:18methodology on achieving the regulatory needs
23:24um for cloud and things like that even if it was like multi-cloud or single Cloud approaches and you’re getting that there
23:31and that’s all and then if you are doing that so it’s a
23:37bit of too many questions for you but if you are doing that then does that does platform engineering then make multi-cloud easier
23:45um in theory well I think multi-cloud always makes it harder from an engineering point of view obviously
23:52because you’ve got two different provide us to deal with so you know
23:59even if you kind of use consistent tooling as like terraform and
24:05um Docker and kubernetes it’s different underlying providers right so it’s never going to just run in multiple places and
24:12you know you’ve effectively got multiple data centers with completely different providers in I think that the time that
24:18it the the hybrid strategy makes sense is maybe this example of
24:23you have uh did a sovereignty concerns legitimate data sovereignty concerns and therefore
24:30you’re on Prem for core systems so-called banking systems or or
24:35government in the case of foreign governments right um and you want to make use of cloud for
24:42point of presence or things like um you know Edge you know so you want to do
24:48iot stuff so that that would be a legitimate strategy where you would have multi-cloud in the sense of something
24:55private VMware or red hat or Oracle or something like that some private cloud
25:02running kubernetes and then some um some Azure or something that runs
25:08stateless applications so we see that sort of thing so could you do that
25:14though with just because the data services the data not the workload
25:20well yeah so you have some you have some workloads that are Maybe that I made maybe stateful and you know
25:28even maybe you consider you know even user information
25:34Sovereign data which which would be the case in like a Middle Eastern Bank let’s say
25:39um so that would need to be private and then there’s like a very narrow circumstance in which you can use public
25:45Cloud where it’s like completely anonymized so the workloads are are purely non-customer you know like
25:51they’re interacting with an edge device or something like that so you see that sort of thing so that would be legitimately pretty educational right
25:58yeah yeah very very Niche but I I guess I
26:03struggle a little bit because the architectural the architecture design that then drives
26:10because there’s so much to now you know from obviously got kubernetes but you have
26:16public Cloud kubernetes you have anthos you have eks anywhere you’ve got you
26:21know Outpost you’ve got Azure stack on-premes there’s you know there’s so
26:27many other ways to think about the public Cloud even when it’s on Prem
26:34um and if the thing if there are I guess that’s where it becomes harder
26:39because now public cloud has evolved in such a way where it actually I mean
26:46what is public Cloud when they go on-prem are they classified now as hybrid I don’t even know what the market
26:52research says is now a public cloud provider now a hybrid cloud provider because
26:57they’ve got Amazon Outpost and other things and are they you know because yeah I would never I think I would never
27:02classify them as if somebody asked me what is the public Cloud you wouldn’t ever think on-prem but yet they’ve now
27:09got services that do support on-prem so there is a bit of a remarketing repositioning probably even for the cloud vendors where they might have to
27:16change how they’ve been seen um and then they’re entering the hybrid market then at that point which is what
27:23they’re doing so yeah I’m not sure if customers even know that like I’m not sure if people out there the different
27:28companies are even aware that it doesn’t mean you can’t now use public Cloud just because you are
27:34on-prem like things have changed um so yeah that’s a good that’s a good
27:40thought actually so you get as well we’ve seen where
27:46you’re wanting to be in let’s say as you are in your example you’re wanting to be in Azure for to be at the edge near the
27:54near the customer that you’re delivering the service to you’re wanting to make use of all the advanced Tooling in the
27:59cloud which if you wind back a couple of years I don’t know what is your stack is exactly like now but it was much more
28:05limited a couple of years back so but customer has a data center and they’ve
28:12put as you are stuck on it and then therefore I mean it’s a little bit of some cost fallacy I would say but but
28:20maybe uh at least in the interim period while that’s committed you know contractually
28:26or or strategic partnership or whatever um or they’re just wedded to it in some
28:32way then it’s cheaper right so it can often be cheaper to run than development and test workloads on-prem and then
28:39early production workloads in the cloud so you see that set up as well so that might be something where you see like
28:45you mentioned a hybrid on-prem and public cloud with the same provider but
28:51we see it more commonly weather with like I mentioned like pivotal Cloud Foundry or red hat or VMware or Oracle
28:58or something on-prem and then they’re using Amazon Google Microsoft in a limited way
29:05in the public and it’s normally a risk diversity do you think then because straddling like you were saying before
29:12you know multi-cloud the different apis different Services
29:18different ways you’ve got to do it terraform obviously is a standard abstraction to a point
29:25um the cloud vendors have even though they are quite different API in points their services are not that dissimilar
29:33um you know they tend to match like for like but just slightly nuanced each time obviously kubernetes is the agnostic
29:40piece from an application perspective the app doesn’t know any different and probably developers wouldn’t that really
29:45different when it comes to kubernetes it could be anywhere at that stage um that how is it working then with your
29:51kind of straddling cloud and then something like VMware which is quite
29:57different you know it’s not the apis aren’t designed quite as well I would
30:03say arguably as say the public Cloud vendors who were born into a way where they were thinking
30:09about apis and abstractions right from the you know the get-go
30:14so does it is it more complicated or do people just get on with it it actually isn’t that
30:20complicating people have the skills I don’t I don’t know much more much more complicated I would say it’s uh
30:26it’s a situation which you want to be in tactically temporarily
30:31um you don’t want to be you don’t want to have two operating models which is ultimately what you end up with to
30:39separate sets of Ingress egress all those networking so you’ve still got a
30:44lot of you’ve still got basically all of the stuff teams licenses that you had
30:49before and now this you know yeah from your organization’s point of view
30:54Cutting Edge bit in separate teams and these people don’t want to work on that and these people aren’t able to
31:00understand that so I think that’s where they end up getting buried
31:06in technical debt isn’t it um and in Legacy you know running that’s what
31:12kills the transformation efforts because the budgets then go over and the
31:17stakeholders demand okay why are we not finished now and that’s how you end up where the they go go on for years and
31:23years you should from the beginning plan to go from one to the other as rapidly as you can but
31:29again it depends on the the constraints and you need a very well thought out
31:35strategy basically you do so you and then so you send people a bit more ring fenced from a more from a skills
31:42perspective so you’re gonna have um even in the devops space then as well or the platform engineering space yeah
31:48definitely right so you’ve got you’ve got obviously the infrastructure side of VMware Networking like you were saying
31:55load balances you’re gonna have that team right because it’s it’s not software defined as it kind of is
32:02after it’s there but somebody’s got to make it there and all yes you know we’re gonna pop the networks in place and
32:08actually configure it all beforehand which obviously don’t do with public Cloud because they’ve done that
32:14so you’ve got that tape and then you’ve then got then devops or platform engineering
32:20then ring fence to knowing the kind of VMware world and then you have then another team that kind of understands
32:26the more public Cloud world like Amazon or Google that’s interesting and do they like are you
32:34trying to get the skills you know when you go into an org where you’re trying up skill
32:40you know both or do do you think it’s better to kind of ring fence the teams rather than cross-pollinate and
32:47upscale well if I’m being very honest I think
32:54you need you need engineers and you need to retrain them in
33:00security principles and and things like this right you need people because the
33:05news the new platforms are made of software you need software engineers and
33:11can you reskill non-engineers you know it’s a lot a very
33:17long time to to learn to become an engineer and learn the software development life cycle and you know even
33:22even platforms now right the the software defined and we’re checking them
33:27in and out and we’re version controlling them and so I don’t think you would want to
33:34to retrain people in that I think you would want to hire software engineers
33:39and make them appreciate for the next you know four years or whatever we’re
33:45going to be in this hybrid State and and this is where our core banking function
33:50or Core Business function is and so you know we need to treat that with the same
33:55importance as this shiny new stuff so that’s probably the line if it was my organization I would go down is is
34:03training the software Engineers to have sympathy for the Legacy estate and and
34:08try and migrate you know have sun setting of of the Legacy as the priority and not often that’s not what you see
34:15happen you often see that it’s good for everyone’s CVS including the the managers just to push this bit and uh of
34:23course the debt is the is the Legacy I mean it’s Legacy for a reason right you want to wrap a veneer around it and then
34:29eventually Sunset it so I think more priority could go on on sun
34:34setting the the old stuff and understand that right at the beginning that that’s what you’re doing that there’s some end
34:40in mind yeah and do you think then because
34:45um do you think the reason it’s taking longer is because let me rephrase the question it what do
34:53you think slows down the migration process to public cloud
34:58um do you think the migration the modernization and migration is more from the application side
35:05um or do you think it’s more from the infrastructure side that kind of slows it down
35:13I think I think it’s the the infrastructure side
35:18but it’s again it’s like it’s software Engineers who know platforms isn’t it so it depends on it
35:26depends on your priori but I would say like a good devops engineer as a developer who’s keen on doing Ops not
35:32the other way not the other way around so there’s obviously abundance of people with with my past history you know system
35:40Engineers but unless they’re a bit like you’ve ever been like hardcore Linux
35:45first they’re not that useful right if it’s if it’s people who know you know
35:50primarily Cisco switches and operating systems and uh you know this
35:57patching in the data center that’s not really a transferable skill right um yeah some of it is some of it is but not
36:04not much of it so that’s the shortage you get right the shortages software engineers and then within the software
36:09engineer shortage you know people who are willing to give up craftsmanship of software and do
36:15building Utility Services which a lot of you know proper developers find boring they don’t want to go into that so
36:22that’s where you find the profound shortage so it’s it’s that bit that we see the most shortage of but then you
36:28know building building really good microservices you know people who can who can do that as well are also lacking
36:35I would say yeah that did I came across a company that could that like profiled a bit like um tracing
36:43it like Trace internally in the application um for like monolithic apps and then it
36:50could work out how to um
36:55kind of decouple the application into microservices so it could work out like what was isolated in the code as in like
37:01what groups how you could group the code into my Microsoft architecture and then
37:06kind of split the code out to be like this should be one microservice this should be another microservices so I think there’s a I think there’s things
37:12going on there in that space to help on the app side and that’s probably just going to keep happening you know
37:17Innovation just always is happening um I’m sure that’s useful but to me to
37:23me it’s it’s it’s kind of the wrong way around so so to me what the bit that often gets
37:28missed is like doing proper domain driven design like actually understanding okay what is your current
37:35application estate I mean actually it’s quite a few organizations that really understand what they’ve already got I’m
37:42sure that you’ve you’ve felt you’ve seen this as well right so so you need to understand that first and then to me you
37:49need to to do domain driven design right you need to understand the domains of the business you need to use the
37:54language of the business decide what it is that you’re building and then carve that out and do it as microservices so
38:01the danger of the danger of a tool like that example although I think if you’re on that path that I’m describing it
38:07would be great because it would then provide a utility to say okay like he has the low hanging fruit to do that but
38:13I think that it’d be dangerous for that to drive it because that’s like an anti-pattern of of domain-driven design
38:19right to have the the way that things were before driving how they should be in the future uh or
38:26the language of the developer you know the ubiquitous language should be the language of the the developer should
38:31learn the language of the business not the other way not the other way around yeah and I think I think yeah it’s spot on really
38:37and then the goals are even what we’re just saying before those aren’t strategies and goals and I think
38:44you’ve pulled on it a few times where you know what is the actual objective
38:49surely the objective has to be the business logic being deep in such a way
38:55that you know that’s the bit that’s making the money for the company yeah everything else really is all about is
39:01just friction isn’t it it’s friction to that like the infrastructure’s friction you know just to get the application
39:08hosted the the clouds can be friction because you’ve got to understand them all before you can actually deliver on
39:13the business logic you know it’s all these things all about like the strategy should be like friction free really that
39:19should be the strategy like just remove all the friction as much as possible so that we’re just down to the business logic and that should be kind of the
39:26strategy and everything else exactly Falls past that doesn’t it yeah and it’s strange because we knew this right it
39:31originally was like under finance and was a service service to the rest of the business and
39:38and and now you know it’s right that that software developers are really
39:44cherished because they’re building the services now because the services are digital but yeah I don’t
39:50I don’t quite understand how it became that the devops called the shots and site reliability
39:57called the shots and obviously those things should be a mean to an end so I think I think that’s maybe what was
40:02behind your comment about I mean like to go back to that you comment about um Recruitment and you know there’s a
40:09whole there’s a whole thing where the industry is driving the behavior in the Enterprise and surely it should be look
40:15this is our strategy this is what our business needs to go this is how we’re competitive and then we have a hiring
40:20strategy that meets that need so yeah what did you mean by by that comment about
40:26you know recruitment driving things yeah recruitment driving things I think
40:32you only know what you need yeah I I it’s a to be fair I think maybe because
40:39we’ve been in the industry a while and probably seen a lot of things you take for granted your experience you know
40:46because you’ve probably seen what works and what doesn’t work over time and you’re watching where the industry is going and you’ve seen good patterns and
40:52bad patterns and I suppose that comes from dealing with multiple customers you know
40:58and being privileged to see and compare and contrast whereas some people might be in the same company for quite a while
41:04and not necessarily have that compare and contrast elements so that’s when the market starts to drive
41:10you know them in some ways because they’re hearing these things and it’s like right okay well it’s it’s devops
41:16and it’s this you know yeah that’s what I need right that’s yeah and people only talk about the successes right people
41:22aren’t saying hey yeah you know we do this thing but we actually failed eight million times before we even got home
41:27and it cost us like loads of money and no one’s really willingly gonna put themselves forward and say that like was
41:33that even cost effective that Journey that we just went on um was that business value so
41:40I think there isn’t enough in the industry to talk about what doesn’t work and what isn’t working well rather than
41:46just you know the the illusion that Innovations always positive in the roles are always
41:53positive but I I think no there probably needs to be something
41:59in the industry to bring or my advice for like ctOS and others would be to join groups Community groups of yourself
42:06so they actually can share ideas I don’t think there’s enough of that maybe the cloud vendor should be bringing the
42:11customers together better um to talk through things and share knowledge because that’s the only right
42:17way and then the roles would sit behind that because then you’d be like right okay now I get what I’m really trying to
42:24do and these are the types of hires I need and now I can go to market and get them
42:29um whereas without that happening I kind of feel sorry for like if you don’t know you just don’t know and if you’re
42:36learning second hand from the market without the experience then obviously you’re not going to know either you’re just going to do whatever it’s telling
42:42you and imagine the cloud vendors probably equally is frustrated because the more
42:50you know the more a customer struggles the less money they’ll make right so in truth so they’re going to be motivated
42:56to have the customer doing well and also the whole cost thing they don’t really want customers I don’t
43:03think spending that amount of money um in the cloud because it risks it’s not getting not getting a result that
43:10they can showcase yeah exactly it’s a risk to them you know that if that noise starts to happen where
43:16it’s not worth the money it’s more expensive and all this other stuff it’s like well
43:21you can do lots of things badly not just the cloud lots of things can be done badly and cost a lot of money it doesn’t
43:28mean you know that you go backwards you need to kind of work out the right thing
43:34so I don’t think the car vendors are sent are incentivized to either to want it so I think the talent
43:42lacking and the groups and the roles it probably needs to be more community centricity
43:47but I think the roles like you’re saying with you don’t if you can’t clarify the
43:52roles then you’re not gonna know why you need them other than the fact
43:58that you probably need them because you are they because other people have them so you’re like well that’s just what
44:03people have so therefore I’m just gonna get the same um but
44:09again it’s not down to a strategy so I think with customers you are you helping
44:14do you help them with the strategy then do you get involved to try and coerce the right conversations
44:21so that you can at least help the customers properly you know rather than just do we
44:28do we try and give them our opinion especially if it’s early and and you know they’re not
44:34suffering from some cost fallacy and it can be difficult right if you’ve already committed if you’ve already if the CEO
44:41has already gone into a strategic relationship with someone or you know they’ve already you know spent a year
44:47working up this plan but yeah the sooner that you can get it back on track and that’s that’s what we try and do
44:54we we generally discourage them from trying to build platform engineering teams
45:00in-house um and that’s why we exist as an organization is because that you know like we talked about before that should
45:07be temporary and yeah we we try and build communities as well so especially when we do government stuff
45:14um you know they’re not competitors right so they often see each other as competitors but shouldn’t
45:20um so we try and get them together as you suggested so that they can follow the ones that have got the
45:25success patterns um I do think there’s like perverse incentives right to like you rightly put
45:33your finger on if um if you’re trying to attract Talent then the mindset is okay we have to project
45:38success and um you know there’s then this race condition of who’s the furthest ahead
45:45and like you like you suggested there’s a there’s a cherry picking of only the success cases which doesn’t really
45:51represent the reality so there’s those perverse incentives and then obviously recruitment agencies are typically set
45:59up that they make a percent of uh of a placed head Council so the higher that
46:04price then the better they do obviously the contractor or the individual does better if that price is inflated so I
46:11guess it follows the hype curve like almost everything else in that sense I’m sure I’m sure it’s not different to any
46:18other hype curve or any other rule but I don’t know you notice it and I noticed that it does seem quite profound in this
46:24particular case um but yeah you should be driven they should be driven by strategy not by
46:31recruitment do you think the industry is being honest then about the problems that
46:36exist as in like you know all the all the terms as well like all the
46:42different Paradise like get tops and talking about that but let’s be honest that’s like not really anything new just
46:48infrastructure is code rebranded um but then it doesn’t really tell you
46:54anything it’s just like a way of working kind of and then again I’m not sure even how
47:01how mature the strategies are there like you know how do you when how does that work at scale
47:07you know how what happens when you’ve got multiple environments and a multiple Cloud accounts multiple environments per
47:13cloud account how do you do all that with Git Ops are you branching are you like I mean is that all these different
47:19things that maybe haven’t been thought thought about but everyone’s talking about how amazing it is yeah and like
47:24you need to do it and you’re like a bit like devops was and all these other things and it tends to be a bit tends to be a bit like that in the
47:30industry it’s all quite hyped up as if everything’s got the answer in it and it’s like but so do you think there’s
47:37honesty you know as much transparency and honesty in the industry that people are
47:42talking enough about what isn’t working um
47:47well obviously it’s not like a singular uh conscience so I don’t think like it’s
47:54it’s necessarily honest or dishonest but it’s exactly as you said it’s it’s more attractive to talk about the
48:00latest shiny thing and the hype and um I don’t know like for me I often think
48:08can we not just concentrate on getting like what’s in the agile Manifesto like
48:13you know working software before documentation like we don’t need a new principle like we have that very old
48:19principle and how how many organizations are observing it or you know test driven development or I guess even devops is
48:27relatively old now um so yeah I would rather say
48:33you know things like XP and I would rather see some Mastery or at least
48:39at least a high level of attainment on some of these bold patterns that really
48:44matter um before before being driven by hype on
48:50particular implementations and um so there’s no doubt right when it be something like kubernetes I mean you
48:57know you see you might have seen that deal about cartoon uh to take off of it and you know so that can suffer from it
49:03as well but I think the difference is there that it’s that it’s already ubiquitous right it’s the only game in
49:09town so for that like you know if you are going to do microservices style
49:14architecture and you’re going to hire real humans to come and do it like it is going to be kubernetes right so so I
49:20think once they become ubiquitous that that type argument goes away right but um yeah I think for for these other new
49:28shiny things I mean they’re all great actually I don’t think there’s any there’s any Pretenders out there that
49:34have you know some are more useful than others and you can argue that all day but but something like get and then get
49:40Ops and infrastructure is clear I mean they’re all you know when you read through the best practices that they’re all
49:47Noble and you should be doing them but I would just ask yeah have you mastered you know is the organization agile first
49:54have you adopted kind of immutable infrastructure first are you following even git
50:01um and that’s often not the case right often you have these different site it’s so ironic right but
50:06you have these silos that the organizations are more siled than ever in some ways you’ve got all these different Clans are there in the
50:12organization you have centralized platform teams as you alluded to and they’re they’re adopting their own thing
50:18so one of the things we try and use is like um is like a technology roadmap and a
50:24technology radar and we try and we try and centralize the design
50:29Authority in the decision making and bring experts in from all of those different teams and you know it can be
50:35seen as putting putting the brakes on and dampening The Innovation but but really they’re off
50:41customers often trying to do too much and not getting anything done so we bring it back to okay like what is the
50:47principle that we’re trying to adhere to are we trying to test first or you know are we trying to improve agility and
50:53okay should we just maybe talk about these foundational blocks in our opinion
50:58that you need first so that’s that’s one thing that we try and do is just bring it bring it back down to Basics and
51:04often it’s hard right because they’re politically there are already three
51:09years into that journey and they don’t want to go back to that you know but they’re not really going back they’re really mastering something
51:16that they should have mastered right at the outset so that’s the challenge right if politically part of the organization thinks it’s very you know in your
51:22example ready for githubs we would have you actually got infrastructure as code
51:28everywhere yet are you doing policy you know are you actually doing get in development successfully yet
51:35yeah it’s your spot on there I think that’s good that you’re doing that because you’re right I suppose it is
51:43yeah if you’ve got so far to admit that you’re going too fast in too many
51:48directions and you need to reorder yourself it’s just basically like regrouping
51:54um but the education I think that you were saying if there’s a skills deficit um in the industry
52:00like anything you’re only as good as the talent that you managed to get in that’s the same with anything you could you
52:06could like to literally Builders any Tradesmen like literally anything right the job’s going to be is done well as
52:12the person and the people that you’re bringing in yeah but the education though the guidance and the for the
52:18customer on what it all means and the principles you were saying like what’s the value out of those ways of working
52:25and why are they important and to educate everyone on it first before you then Implement and do
52:31um so that everyone’s on the same page and there’s no illusions on like what it actually meant
52:37um because it’s funny because I think I’ve been in meetings where people have different perceptions
52:44of what the things mean you know and they’ve been doing them for a while and you could be in a meeting you can
52:50clearly tell that everyone’s got a slightly different view of roughly the same types of things so people would be
52:56saying words but what they’re meaning to someone else is slightly different to what they might be doing with someone else I mean you’re all you know really
53:02the rooms in this like state of confusion about what it all actually means and you can see that you think
53:08well if that’s in your company then yeah you’re probably not going to get very far because you’re already starting off on
53:14the wrong thought so um yeah yeah and and so what do you see next then for
53:19yourselves at codification here um
53:24progressing with the clouds doubling down on the platform engineering side helping customers more on the agile
53:31aspects or a bit of everything well the best the best relationship for us is when we’re offering all of that
53:37and we’re like a trusted partner and the customer accepts that it’s a
53:43complex challenge it’s a multi-year challenge so
53:49where we can have a relationship where we can honestly assess where they’re at so that’s one of the tools that we use
53:54is you know you can make those arguments but ultimately they can be political that can be another supplier there or
54:01even when there’s not very opinionated people in the organization like you rightly said you can have
54:07you know multiple people using a word but meaning different things by it so so one of the things we try and do is like
54:13the terms of reference and we try and use you know standard you know try not let the
54:19organizing come up with its own terminology but use standard tools and standard language to describe those
54:25tools and standard methodologies that are that are proven um and we try and do
54:31um assessment somewhat by stealth so so rather than going around saying okay
54:38we’re going to assess everyone we’ll often be engaged to do narrow form of assessment which we’ll
54:45then broaden out so we’ll be we’ll be engaged to do like a platform maturity
54:51assessment but then we’ll broaden it out into saying okay how is the testing being done what’s the self-service like
54:56for developers and then you’re actually into people in process
55:01um so where we can get that level of trust to be able to do those sorts of
55:07things we can get much better results and then especially if they engage us to kind of assess everything out in the
55:13open which is which is rare to be fair um because we’re a because we’re a small
55:18startup and not that well known so that can happen and you know maybe we’ve done some small one-off projects and then
55:25we’ve built up the trust and then you know we’ve made our way up to the kind of senior leadership and then they say
55:30okay you know tell us where we’re really at and then what we can do is compare the
55:35scores on multiple narrow things and say okay when you when you compare these in aggregate it kind of gives you a level
55:41of maturity let’s compare that with our other customers
55:46let’s compare that with people in your industry or different industry or where
55:51you’re trying to get to um so that that can work giving them this like objective comparison so
55:57obviously we don’t we don’t disclose who that other customer was or break any kind of confidentiality there but we just say
56:04okay another another client in this sector this is how they scored this is where they’re up to on their Journey
56:09this is their perception of themselves before the assessment this is the reality of what the assessment showed so
56:15we do that sort of thing and we do um like lab-based training of their
56:21people as well so we can actually get the capability assessment there so they send you know again we can ask them
56:27where do you think this team is at okay this you know after six weeks of instruction this is how well they
56:34followed the instruction this is how well they completed the assignments this is how your team compares to a different team in the organization that’s maybe in
56:40a completely different leadership structure different incentives so they can find that you know eye-opening as
56:47well and then again you can compare the capability to the outside world so
56:52that’s what we like and and we can compare new joiners that maybe have been through some kind of codification
56:59bootcamp before joining and then put on the project and then assessed versus someone that they’re trying to upskill
57:05or someone they haven’t given the proper time to achieve the upskilling so all of those things we can
57:11we can kind of help them with so it doesn’t matter to us if they’re really
57:16struggling or they’re really Advanced you know we’re our team is technically always in advance of their
57:22team because they’re not a technology company they’re they’re trying to become one so so we’ve got no
57:28no real strong opinion of what they ought to be and we just try and show them okay this is where this is where
57:34you thought you were and this is where you really are this is where you’re trying to get to and this is the gap and
57:39then we can help across the board right we can help by augmenting them with Engineers be it platform Engineers or
57:45Cloud Engineers or whatever we can train retrain the teams or we can hire
57:51hire new permanent staff for them nice and if somebody wanted to get in touch how do they get hold of
57:57yourselves if they’re like listened and like oh that is what we need like we’re in a bit of a tricky situation at the
58:03moment and maybe need to get yourselves in the website codification.io would be the
58:09one to go cool um all right well um it’s been great chatting to you
58:15Richard um and thanks for spending the time obviously to come on the podcast and talk about these things with everybody
58:21hopefully somebody got something out of it um who’s listening and obviously as you
58:26heard you can reach out to codification.io if there’s things there and you think that can help them please do reach out so awesome thank you John
58:33thanks cheers thanks bye [Music]
0:00Docker and kubernetes it’s different underlying providers right so it’s never going to just run in multiple places and
0:07you know you’ve effectively got multiple data centers with complete different providers [Music]
0:23hello Welcome to Cloud unplugged this is season two episode eight I am here with
0:28Richard Smith from codification I’m John Shanks uh Richard do you want to
0:33introduce yourself yeah I’ll just give you a part of History John so uh I started working in
0:39it back in 2001 originally as a technician and then I moved into infrastructure engineering doing things
0:45like setting up data centers with VMware in 2012 that’s when I went became an
0:51independent contractor so I did site reliability engineering for the London Olympics
0:57um I helped startup go to Cloud basically from a from a
1:03software product to a SAS product uh helped a housing company get off their
1:09physical data center into private Cloud so third-party Cloud um and about eight years ago that’s when
1:16the assignments became kind of around Cloud native and devops so I was the devops lead for Department of work and
1:23pensions and I’d got their first microservice live known as the pension tracing service that same year 2016 was
1:32also when I joined HSBC as their Enterprise architect for platform so that was also about Cloud so AWS this
1:40time um the cloud native as well about exploring microservice style architectures and then the last project
1:46that I did before end of being full-time codification CEO was hmrc so that was
1:53their roll on roll off uh project which is better known as the brexit platform so I did the
1:59architecture for that um and all of those projects were supposed to be about
2:05agility and devops and I could see these kind of recurring patterns of failure uh
2:11patterns of success as well but but you know lots of the same mistakes being repeated so that’s when I decided to to
2:17set up codification so I had the first employee back in April 2019
2:23um and that’s what I do now so now we work with Enterprises with in-house engineering teams and help them on the
2:30same kind of Journey adopting cloud cloud native nice and and the challenges that you’re
2:35helping customers visit them at like current and at the moment and previously that’s all
2:42what is the type of constraints that you’re more focused on is it around modernization or
2:49applications itself or is it more like Cloud infrastructure automation or
2:54um or a bit of all of it is it all of those things I mean it it is a bit of all of it
2:59sometimes when when we’re kind of very heavily involved with the client but um mainly platform engineering is is the
3:06sweet spot so it is around like adopting kubernetes mainly but of course that
3:12then breaks into CI CD it breaks into the whole Devon Ops and and breaking
3:17down the organizational silos and then you know it bleeds into everything else where they don’t necessarily have the
3:24talent in-house so training the talents hiring the talent onboarding them
3:30um showing them the way all of those sorts as well right I see yeah that’s really good so the the stuff around the
3:37talent side is that is that the biggest construct obviously kubernetes I think
3:42is a big uh is a big topic obviously um and you know we’re in that space too so
3:50we kind of see how complicated well it is just complicated and that’s just factual what it’s trying to do is
3:56complicated um so is that what you’re seeing quite a lot of the challenges is people struggling for adoption and kubernetes
4:03it’s it’s harder than it they thought it was going to be or is it yeah definitely definitely so it’s the it’s the
4:10it’s the lack of Talent on the market for fast follow companies so of course
4:16the the very large tech companies don’t struggle with this as much of course but
4:22um outside of those I think there’s a shortage of talent for one thing and
4:27then it’s precisely what you said it’s very easy to underestimate you know in some ways it feels like Cloud native is
4:34is like an order of magnitude jump up from from virtualization and complexity so it’s very easy for for organizations
4:41who’ve done relatively okay with provisioning data sent as an in-house application development
4:47um to assume okay it’s it’s that kind of paradigm shift and I feel as though from
4:53what I’ve seen over the years that it’s that it’s much much harder than that and yeah it catches the complexity catches
4:59people off guard and what what do you what do you think is driving the adoption of kubernetes is more like the
5:05market or is it technologists coming in representing the
5:10need in the business and then trying to implement I guess what do you feel like it’s people that come in that then drive the
5:17adoption of kubernetes in like they make the decision for the business all the business is recognized the value of the technology for what
5:24they need or do you still think it’s a bit disconnected like the Technologies still focused on the technologists
5:29without the true value of understanding of why kubernetes
5:35um I think it’s a bit of like there’s a bit of that where the the Tails wagging the dog where the implementation technology
5:41kubernetes or or containerizations chosen first but I think
5:47you know early early days I guess like 2015 it was this agnosticism idea of
5:54like you know they’d experienced customers had experienced vendor lock-in
5:59before and were terrified of that um and so that was like can we build on an abstraction layer can we be you know
6:07multi um cloud or tool agnostic or whatever so
6:12I think that was driving it earlier now I actually think and I don’t know you know we’ll get into
6:17that probably but I don’t know that I fully agreed with those those sentiments but I think now it is
6:23it is kind of that there’s a pressure from the customer to be Omni Channel or to
6:30release software quickly and actually to achieve that you do need you know microservices and that then
6:36demands Cloud native approach which then kubernetes is the only game in town but actually when I started doing this sort
6:42of thing there was other valid options you know swarm was quite quite big at the time so we were doing a lot of swarm
6:47stuff so I don’t think it’s kubernetes as the driver uh but probably like microservices style architecture and
6:54needing to independently deploy services to do Omni channel is is probably what it is now yeah so like more
7:02containers it themselves unlike being more portable and slim which more developer friendly really than a VM yeah
7:09um I mean you know devs would never really be spinning VMS up or doing vagrant that was always like
7:16infrastructure exactly related wasn’t it and then suddenly you could then actually
7:21start packaging up your application into something like Docker and then that became easier
7:26for developers and then kubernetes came out later didn’t it really because it was just Docker at the beginning mostly then
7:34kubernetes came later mesos Marathon swarm then obviously kubernetes
7:39so but yeah that’s what we saw people successfully adopt containers
7:44um probably not true microservices but but you know containerizing to do lift and shift
7:52so finding that relatively doable in-house and then a lot of our
7:59customers came from those that then got stuck either trying to get it into production or they’re in production and
8:05trying to scale it or even more often the kind of in
8:10production at some kind of scale but can’t tell when it breaks why it’s broke so the observability stuff and stock
8:17tracing and that sort of thing so I think it all comes back to lack of platform engineering Talent so it’s good
8:24development Talent maybe good system engineering but that kind of crossover
8:29skill set is is missing and still now right yeah I agree with that I I find it like
8:36platform engineering is kind of like funny because the term could have been around probably for a
8:43while but it’s then either maybe like the central model versus disaggregated model so it’s like
8:49depending on how you perceiving it if you’re just having people in projects and you probably won’t call that
8:54platform engineering you just call up like devops you know devops embedded in a team delivering yeah with the
9:01application team but then the minute you kind of rolled into something Central so actually that model there’s not
9:07enough consistency or oversight or um you’re kind of relying on
9:12documentation and standards and all these other things to try and cohere some consistent approach or some
9:18security model and then kind of centralizing it was a better option because at least it’s in one place like
9:25you’re doing everything in one place yeah but then that’s risen to the concept of platform engineering is in
9:30like engineering everything in a central place but
9:36the term platform obviously is so used that it became you know like the cloud to platform isn’t it you know all those
9:43like platforms everywhere and platforms as a service all these different phrases yeah so it is quite confusing I’m not
9:51sure whether people do understand what platform engineering means it would devops suffers the same curse
9:57doesn’t it that’s true yeah devops seemed to me to sort that start out like
10:04agile so it was a culture culture of organic automation or a culture of
10:10breaking down organizational silos um that sort of thing and then you you
10:16know I was called a devops engineer at one point so that’s kind of an Andy pattern in in my mind so you know that
10:23that term also gets gets misused like devops engineer is that should that be a
10:28should that be a thing or you know but then as like you said there’s this kind of Federated view of where you have
10:34someone embedded in the team and sometimes they call that site reliability engineer but but
10:39whatever yeah I think I think for us we use it because that’s what we’re doing so it fits with our organization because
10:47you know we are specifically building a platform whether it’s whether it’s a cloud Landing zone or it’s a kubernetes
10:53platform or whatever so that’s what we’re doing I think it’s a separate question of should non-technology companies that are
11:00you know Finance Company becoming fintech or health becoming like should you should you have that you know either
11:06of those Concepts should you have a devops team or a platform engineering team I think probably no
11:13um but but we need one because that’s what we’re what we’re typically being contracted to do yeah yeah that makes a
11:19lot of sense it is very confused isn’t it because there’s loads of um people brand themselves of different
11:25things I don’t I don’t think a lot of people can keep up with the terminologies and the trends because obviously it’s rapid isn’t it and
11:32there’s so many different paradigms on top of it like you’re saying Sr read obviously born out
11:37of Google is mostly about scale you know they were high scale so you can kind of expect the need to be reliably
11:45engineering um for the scale that you’ve got to so it doesn’t go down and you can perform well
11:51and then devops obviously it’s more of a methodology than a role got coined mostly by
11:57um recruiters maybe in truth and then people were like branded as an engineer
12:03um and then they’d be kind of there was a period where there were two like people kept using them interchangeably like oh I’m an SRE but they were kind of
12:09just doing Engineering in a project maybe um so it still really devops because it
12:15wasn’t that they were doing reliability engineering yeah and platform engineering happened at the same time
12:20and then platform engineering with devops and sres because some companies have like all of them on the go yeah
12:27they like different shapes and sizes and it is quite complicated to know who’s doing what what are the roles and
12:33responsibilities of everybody absolutely um do you want to give your opinion on what like the the value of
12:40platform engineering is versus devops versus SRE that you see in customers just to
12:46clarify things a bit I mean I think if if our adopt this
12:52dance of the customer so if the customers let’s say they’re a software Enterprise or or a bank trying to
12:59release digital banking products or whatever right they should want to concentrate on building the product and
13:06so I think what you want is that to just be abstracted away and be self-service
13:11now there’s a there’s a road map there’s a transformation to get to that point
13:16but I think as a as a Target State that’s where you want to end up and
13:22therefore should the organization care about all of these things so this is what I see on
13:28a on a typical transformation program they’re trying to get the grips even still with agile you know a lot of people are still waterfall mindset
13:34prints too service management so you’ve got that that they’re trying to understand they’re trying to understand
13:40the transformation strategy the vision from the organization itself they’re trying to understand now this devops
13:46thing and like you see maybe maybe even two or three other competing things and I think that’s a bad idea I think
13:53probably under the under the cover as you want you know you want automation so you want
13:59you want that part of devops you probably want you know the E you know
14:05eat your own dog food concept of if you build it you look after it in production so I think you want devops culture part
14:12uh I think you need platform engineering to get there right you know in a in a typical Enterprise it’s going to be one
14:19to five years depending on how much budget you put up front in how serious you are about and how many
14:24mistakes you make along the way so but that should be transitional in my view and then you should end up with
14:31something that’s self-service for the developer what what do you think
14:37um yeah I totally agree with that I think I’m not sure I think what’s difficult
14:43probably for customers is knowing when you need something like what are the markers that tell you when you need to
14:50change because I guess every company is so different yet the terminology is
14:55uh kind of singular they don’t they’re not conditionalized you know as it’s just like platform Engineers platform
15:01engineering devops is devops but there’s no condition of like which one do I choose if I’m a small company am I going
15:06devops straight away or like do I hire do I hire a lead in a small team who
15:12understands cloud or has maybe got some devops principle knowledge you know I think that’s quite hard because there’s
15:18no there’s no real good guidance probably on like what it means and then large
15:24Enterprises who are probably straddling really complicated
15:30um complicated environments and that’s nothing to do with the strategy that’s just history like you’ve been around
15:35long enough yeah to have to have been on a data center because you predate Cloud right so that’s just a historical thing
15:43and so then you’re into that like a totally different type of business and so what they’re trying to do might be
15:48completely different their needs will be completely different to say a really small exact coming that’s all in the
15:54cloud so I do think it is hard for people to know what does it all mean to me you know
16:00like my company when do I need these resources and when’s a good time
16:06to invest um and in the in that as a principle
16:12because most of the time I think people are getting people in when
16:17they’ve they’ve got the problem rather than you know then they’re in a bit of a mess so then they’re seeking
16:23help because they’ve got them to land themselves in a bit of a mess rather than understanding how they’re going to evolve in
16:28yeah it’s a delivery but yeah and you see this like cargo cult thing of you know this organization
16:36wants to do it because it’s competitors are all doing it without really understanding why I think that’s yeah
16:41that’s the tail wagging the dog I mean for me it should come from like a clearly articulated business strategy
16:48and then whoever’s doing a strategy in it should should articulate you know
16:53this is the vision this is the strategy and then it it can follow from that if that’s clear right if what are we trying
17:00to do are we are we trying to reduce you know is our main priority to reduce time to Market is it cost saving you know
17:07what like what is it because that that might dictate a different path are we
17:12wanting to bring all of the incumbent staff suppliers whatever on the journey
17:18that that would obviously tell you a lot about what you should do um and I think where it goes wrong is if
17:24you just kind of go to the market start hiring for buzzwords then you’ve been in
17:30quite a mess because you don’t understand why you’re doing it right yeah exactly and it’s the same with all the
17:36um like you’re talking even about multicult we’ll come on to in a minute but even those things strategically aren’t
17:42necessarily tied to business outcomes you know like what what’s the driver of
17:48multi-cloud exactly or you know people saying hybrid and it’s like right our strategy is hybridating
17:53like how is how is that an actual strategy because it’s not business driven I mean it’s like what is the
17:59outcome you’re even what do you get at the end of hybrid um and like what was the cost did you
18:06save money did you not save money like what was the objective actually so there’s all these things that are term strategies but they’re not really
18:12strategies because I agree with you I don’t I I don’t see how that can be a you know it gets termed as a strategy
18:17but I’ve never seen I’ve never heard a CIO or a strategist an organization articulate why that’s a strategy
18:24ultimately when they articulated it’s a tactic right it’s not a strategy we’re
18:29on Prem and would like to be fully in xcloud and so we’re hybrid well yeah that’s just the state of being rather
18:36than a than a strategy right so yeah but I I you know think maybe the same thing
18:42could be true of of having multiple public clouds as well I don’t know what you think about that but but is that a
18:49strategy is that a meaningful strategy that you’ve heard articulated uh from
18:54finances from financial institutions I mean over here in UK obviously have the pra which is like the regulatory body
19:01I’m not sure they say you need to you has to you have to be multi-cloud I
19:06think it’s then a bit like anything Chinese Whispers I think they say something around the resiliency there
19:13um and obviously like data recovery perspectives and uptime Etc like or if
19:19the cloud was to go bankrupt that you know for whatever reason um that you’ve Diversified enough to
19:26like reduce the risk I don’t think they ever say that then means you have to multi-cloud
19:32um it just means you need to get out a jail card right basically I think is really what they’re saying but
19:39I think just what you were saying before on the effort it can take some companies to
19:44even do one and the lack of understanding the hiring the talent and everything else I mean my view I’m I’ll
19:52ask yours as well is like it seems almost impossible to do two clouds after
19:57you’ve done one I don’t know you’re agree with that and that maybe puts like value
20:02and reward isn’t it it’s like is there value in it what’s the reward of doing both
20:08um and is it worth the effort in the end so I don’t know I’ve never seen anyone pull it off as in like they’re actually
20:16achieve the Strategic aim of being either on the one hand getting the best
20:21out of both clouds so sometimes that’s the use case right of like we want to use Google for developer experience and
20:28AWS for maturity or AWS maturity and as you are to be close to the data lake or
20:34so you sometimes hear that as a as a reason and I’ve never seen that realized
20:40and then the other reason is you know resiliently and availability over over monthly for that reason and
20:46I’ve I’ve never seen that either so I think you know these surveys come out and they say how you know there’s more people
20:52doing multi-cloud than single but I don’t I kind of imagine that there’s a
20:58very very small subset of those that are actually achieving that I mean yeah
21:04yeah I totally agree with that I have seen the using the right card for the right job but I think you have to be quite mature
21:12to know how to do that as in to have achieved
21:17I guess if you’ve if you can think about the infrastructure as separate from application in in some ways that if you
21:23can think that the application has dependencies outside of like how I run as a workload
21:29um and in the app dependencies could be like a database or an object storage or something that a
21:35much more simplistic than say infrastructure and networking is then if you can make the infrastructure in the
21:41networking very service driven as in like a service then it can work but that takes a lot of
21:48design and upfront thinking it’s the trade-off though isn’t it like it can work but
21:54I I think unless like you said Regulatory Compliance maybe could be one reason if but uh
22:02but you know I didn’t I didn’t see that as a as a dictat when I was at HSBC when they were doing it so
22:08they wanted agnosticism because they’ve maybe feared competition or other things
22:14like that right but so so maybe if you’re so large maybe if you’re in the top 500 largest organizations in the
22:20world it can make sense maybe if there’s true Regulatory Compliance so Middle Eastern banks for example so they they
22:27maybe have a real case where they can’t be on one American public Cloud um so that that probably makes sense but
22:34short of those the challenge is going to be what we were talking about before right it’s going to be the sprawl and
22:41the lack of talent to be able to do one so who which mid size or even large but
22:47not super large organization has the luxury of you know the trade-off right to try and do too is gonna essentially
22:54mean that you’re gonna take three times as long to accomplish the same goals as just focusing on one I
23:00would say yeah just a minute I’m gonna have to start terms and of
23:06um can you talk about like compliance um and the security aspects of things so
23:13when you’re promoting platform engineering are you seeing that as a better
23:18methodology on achieving the regulatory needs
23:24um for cloud and things like that even if it was like multi-cloud or single Cloud approaches and you’re getting that there
23:31and that’s all and then if you are doing that so it’s a
23:37bit of too many questions for you but if you are doing that then does that does platform engineering then make multi-cloud easier
23:45um in theory well I think multi-cloud always makes it harder from an engineering point of view obviously
23:52because you’ve got two different provide us to deal with so you know
23:59even if you kind of use consistent tooling as like terraform and
24:05um Docker and kubernetes it’s different underlying providers right so it’s never going to just run in multiple places and
24:12you know you’ve effectively got multiple data centers with completely different providers in I think that the time that
24:18it the the hybrid strategy makes sense is maybe this example of
24:23you have uh did a sovereignty concerns legitimate data sovereignty concerns and therefore
24:30you’re on Prem for core systems so-called banking systems or or
24:35government in the case of foreign governments right um and you want to make use of cloud for
24:42point of presence or things like um you know Edge you know so you want to do
24:48iot stuff so that that would be a legitimate strategy where you would have multi-cloud in the sense of something
24:55private VMware or red hat or Oracle or something like that some private cloud
25:02running kubernetes and then some um some Azure or something that runs
25:08stateless applications so we see that sort of thing so could you do that
25:14though with just because the data services the data not the workload
25:20well yeah so you have some you have some workloads that are Maybe that I made maybe stateful and you know
25:28even maybe you consider you know even user information
25:34Sovereign data which which would be the case in like a Middle Eastern Bank let’s say
25:39um so that would need to be private and then there’s like a very narrow circumstance in which you can use public
25:45Cloud where it’s like completely anonymized so the workloads are are purely non-customer you know like
25:51they’re interacting with an edge device or something like that so you see that sort of thing so that would be legitimately pretty educational right
25:58yeah yeah very very Niche but I I guess I
26:03struggle a little bit because the architectural the architecture design that then drives
26:10because there’s so much to now you know from obviously got kubernetes but you have
26:16public Cloud kubernetes you have anthos you have eks anywhere you’ve got you
26:21know Outpost you’ve got Azure stack on-premes there’s you know there’s so
26:27many other ways to think about the public Cloud even when it’s on Prem
26:34um and if the thing if there are I guess that’s where it becomes harder
26:39because now public cloud has evolved in such a way where it actually I mean
26:46what is public Cloud when they go on-prem are they classified now as hybrid I don’t even know what the market
26:52research says is now a public cloud provider now a hybrid cloud provider because
26:57they’ve got Amazon Outpost and other things and are they you know because yeah I would never I think I would never
27:02classify them as if somebody asked me what is the public Cloud you wouldn’t ever think on-prem but yet they’ve now
27:09got services that do support on-prem so there is a bit of a remarketing repositioning probably even for the cloud vendors where they might have to
27:16change how they’ve been seen um and then they’re entering the hybrid market then at that point which is what
27:23they’re doing so yeah I’m not sure if customers even know that like I’m not sure if people out there the different
27:28companies are even aware that it doesn’t mean you can’t now use public Cloud just because you are
27:34on-prem like things have changed um so yeah that’s a good that’s a good
27:40thought actually so you get as well we’ve seen where
27:46you’re wanting to be in let’s say as you are in your example you’re wanting to be in Azure for to be at the edge near the
27:54near the customer that you’re delivering the service to you’re wanting to make use of all the advanced Tooling in the
27:59cloud which if you wind back a couple of years I don’t know what is your stack is exactly like now but it was much more
28:05limited a couple of years back so but customer has a data center and they’ve
28:12put as you are stuck on it and then therefore I mean it’s a little bit of some cost fallacy I would say but but
28:20maybe uh at least in the interim period while that’s committed you know contractually
28:26or or strategic partnership or whatever um or they’re just wedded to it in some
28:32way then it’s cheaper right so it can often be cheaper to run than development and test workloads on-prem and then
28:39early production workloads in the cloud so you see that set up as well so that might be something where you see like
28:45you mentioned a hybrid on-prem and public cloud with the same provider but
28:51we see it more commonly weather with like I mentioned like pivotal Cloud Foundry or red hat or VMware or Oracle
28:58or something on-prem and then they’re using Amazon Google Microsoft in a limited way
29:05in the public and it’s normally a risk diversity do you think then because straddling like you were saying before
29:12you know multi-cloud the different apis different Services
29:18different ways you’ve got to do it terraform obviously is a standard abstraction to a point
29:25um the cloud vendors have even though they are quite different API in points their services are not that dissimilar
29:33um you know they tend to match like for like but just slightly nuanced each time obviously kubernetes is the agnostic
29:40piece from an application perspective the app doesn’t know any different and probably developers wouldn’t that really
29:45different when it comes to kubernetes it could be anywhere at that stage um that how is it working then with your
29:51kind of straddling cloud and then something like VMware which is quite
29:57different you know it’s not the apis aren’t designed quite as well I would
30:03say arguably as say the public Cloud vendors who were born into a way where they were thinking
30:09about apis and abstractions right from the you know the get-go
30:14so does it is it more complicated or do people just get on with it it actually isn’t that
30:20complicating people have the skills I don’t I don’t know much more much more complicated I would say it’s uh
30:26it’s a situation which you want to be in tactically temporarily
30:31um you don’t want to be you don’t want to have two operating models which is ultimately what you end up with to
30:39separate sets of Ingress egress all those networking so you’ve still got a
30:44lot of you’ve still got basically all of the stuff teams licenses that you had
30:49before and now this you know yeah from your organization’s point of view
30:54Cutting Edge bit in separate teams and these people don’t want to work on that and these people aren’t able to
31:00understand that so I think that’s where they end up getting buried
31:06in technical debt isn’t it um and in Legacy you know running that’s what
31:12kills the transformation efforts because the budgets then go over and the
31:17stakeholders demand okay why are we not finished now and that’s how you end up where the they go go on for years and
31:23years you should from the beginning plan to go from one to the other as rapidly as you can but
31:29again it depends on the the constraints and you need a very well thought out
31:35strategy basically you do so you and then so you send people a bit more ring fenced from a more from a skills
31:42perspective so you’re gonna have um even in the devops space then as well or the platform engineering space yeah
31:48definitely right so you’ve got you’ve got obviously the infrastructure side of VMware Networking like you were saying
31:55load balances you’re gonna have that team right because it’s it’s not software defined as it kind of is
32:02after it’s there but somebody’s got to make it there and all yes you know we’re gonna pop the networks in place and
32:08actually configure it all beforehand which obviously don’t do with public Cloud because they’ve done that
32:14so you’ve got that tape and then you’ve then got then devops or platform engineering
32:20then ring fence to knowing the kind of VMware world and then you have then another team that kind of understands
32:26the more public Cloud world like Amazon or Google that’s interesting and do they like are you
32:34trying to get the skills you know when you go into an org where you’re trying up skill
32:40you know both or do do you think it’s better to kind of ring fence the teams rather than cross-pollinate and
32:47upscale well if I’m being very honest I think
32:54you need you need engineers and you need to retrain them in
33:00security principles and and things like this right you need people because the
33:05news the new platforms are made of software you need software engineers and
33:11can you reskill non-engineers you know it’s a lot a very
33:17long time to to learn to become an engineer and learn the software development life cycle and you know even
33:22even platforms now right the the software defined and we’re checking them
33:27in and out and we’re version controlling them and so I don’t think you would want to
33:34to retrain people in that I think you would want to hire software engineers
33:39and make them appreciate for the next you know four years or whatever we’re
33:45going to be in this hybrid State and and this is where our core banking function
33:50or Core Business function is and so you know we need to treat that with the same
33:55importance as this shiny new stuff so that’s probably the line if it was my organization I would go down is is
34:03training the software Engineers to have sympathy for the Legacy estate and and
34:08try and migrate you know have sun setting of of the Legacy as the priority and not often that’s not what you see
34:15happen you often see that it’s good for everyone’s CVS including the the managers just to push this bit and uh of
34:23course the debt is the is the Legacy I mean it’s Legacy for a reason right you want to wrap a veneer around it and then
34:29eventually Sunset it so I think more priority could go on on sun
34:34setting the the old stuff and understand that right at the beginning that that’s what you’re doing that there’s some end
34:40in mind yeah and do you think then because
34:45um do you think the reason it’s taking longer is because let me rephrase the question it what do
34:53you think slows down the migration process to public cloud
34:58um do you think the migration the modernization and migration is more from the application side
35:05um or do you think it’s more from the infrastructure side that kind of slows it down
35:13I think I think it’s the the infrastructure side
35:18but it’s again it’s like it’s software Engineers who know platforms isn’t it so it depends on it
35:26depends on your priori but I would say like a good devops engineer as a developer who’s keen on doing Ops not
35:32the other way not the other way around so there’s obviously abundance of people with with my past history you know system
35:40Engineers but unless they’re a bit like you’ve ever been like hardcore Linux
35:45first they’re not that useful right if it’s if it’s people who know you know
35:50primarily Cisco switches and operating systems and uh you know this
35:57patching in the data center that’s not really a transferable skill right um yeah some of it is some of it is but not
36:04not much of it so that’s the shortage you get right the shortages software engineers and then within the software
36:09engineer shortage you know people who are willing to give up craftsmanship of software and do
36:15building Utility Services which a lot of you know proper developers find boring they don’t want to go into that so
36:22that’s where you find the profound shortage so it’s it’s that bit that we see the most shortage of but then you
36:28know building building really good microservices you know people who can who can do that as well are also lacking
36:35I would say yeah that did I came across a company that could that like profiled a bit like um tracing
36:43it like Trace internally in the application um for like monolithic apps and then it
36:50could work out how to um
36:55kind of decouple the application into microservices so it could work out like what was isolated in the code as in like
37:01what groups how you could group the code into my Microsoft architecture and then
37:06kind of split the code out to be like this should be one microservice this should be another microservices so I think there’s a I think there’s things
37:12going on there in that space to help on the app side and that’s probably just going to keep happening you know
37:17Innovation just always is happening um I’m sure that’s useful but to me to
37:23me it’s it’s it’s kind of the wrong way around so so to me what the bit that often gets
37:28missed is like doing proper domain driven design like actually understanding okay what is your current
37:35application estate I mean actually it’s quite a few organizations that really understand what they’ve already got I’m
37:42sure that you’ve you’ve felt you’ve seen this as well right so so you need to understand that first and then to me you
37:49need to to do domain driven design right you need to understand the domains of the business you need to use the
37:54language of the business decide what it is that you’re building and then carve that out and do it as microservices so
38:01the danger of the danger of a tool like that example although I think if you’re on that path that I’m describing it
38:07would be great because it would then provide a utility to say okay like he has the low hanging fruit to do that but
38:13I think that it’d be dangerous for that to drive it because that’s like an anti-pattern of of domain-driven design
38:19right to have the the way that things were before driving how they should be in the future uh or
38:26the language of the developer you know the ubiquitous language should be the language of the the developer should
38:31learn the language of the business not the other way not the other way around yeah and I think I think yeah it’s spot on really
38:37and then the goals are even what we’re just saying before those aren’t strategies and goals and I think
38:44you’ve pulled on it a few times where you know what is the actual objective
38:49surely the objective has to be the business logic being deep in such a way
38:55that you know that’s the bit that’s making the money for the company yeah everything else really is all about is
39:01just friction isn’t it it’s friction to that like the infrastructure’s friction you know just to get the application
39:08hosted the the clouds can be friction because you’ve got to understand them all before you can actually deliver on
39:13the business logic you know it’s all these things all about like the strategy should be like friction free really that
39:19should be the strategy like just remove all the friction as much as possible so that we’re just down to the business logic and that should be kind of the
39:26strategy and everything else exactly Falls past that doesn’t it yeah and it’s strange because we knew this right it
39:31originally was like under finance and was a service service to the rest of the business and
39:38and and now you know it’s right that that software developers are really
39:44cherished because they’re building the services now because the services are digital but yeah I don’t
39:50I don’t quite understand how it became that the devops called the shots and site reliability
39:57called the shots and obviously those things should be a mean to an end so I think I think that’s maybe what was
40:02behind your comment about I mean like to go back to that you comment about um Recruitment and you know there’s a
40:09whole there’s a whole thing where the industry is driving the behavior in the Enterprise and surely it should be look
40:15this is our strategy this is what our business needs to go this is how we’re competitive and then we have a hiring
40:20strategy that meets that need so yeah what did you mean by by that comment about
40:26you know recruitment driving things yeah recruitment driving things I think
40:32you only know what you need yeah I I it’s a to be fair I think maybe because
40:39we’ve been in the industry a while and probably seen a lot of things you take for granted your experience you know
40:46because you’ve probably seen what works and what doesn’t work over time and you’re watching where the industry is going and you’ve seen good patterns and
40:52bad patterns and I suppose that comes from dealing with multiple customers you know
40:58and being privileged to see and compare and contrast whereas some people might be in the same company for quite a while
41:04and not necessarily have that compare and contrast elements so that’s when the market starts to drive
41:10you know them in some ways because they’re hearing these things and it’s like right okay well it’s it’s devops
41:16and it’s this you know yeah that’s what I need right that’s yeah and people only talk about the successes right people
41:22aren’t saying hey yeah you know we do this thing but we actually failed eight million times before we even got home
41:27and it cost us like loads of money and no one’s really willingly gonna put themselves forward and say that like was
41:33that even cost effective that Journey that we just went on um was that business value so
41:40I think there isn’t enough in the industry to talk about what doesn’t work and what isn’t working well rather than
41:46just you know the the illusion that Innovations always positive in the roles are always
41:53positive but I I think no there probably needs to be something
41:59in the industry to bring or my advice for like ctOS and others would be to join groups Community groups of yourself
42:06so they actually can share ideas I don’t think there’s enough of that maybe the cloud vendor should be bringing the
42:11customers together better um to talk through things and share knowledge because that’s the only right
42:17way and then the roles would sit behind that because then you’d be like right okay now I get what I’m really trying to
42:24do and these are the types of hires I need and now I can go to market and get them
42:29um whereas without that happening I kind of feel sorry for like if you don’t know you just don’t know and if you’re
42:36learning second hand from the market without the experience then obviously you’re not going to know either you’re just going to do whatever it’s telling
42:42you and imagine the cloud vendors probably equally is frustrated because the more
42:50you know the more a customer struggles the less money they’ll make right so in truth so they’re going to be motivated
42:56to have the customer doing well and also the whole cost thing they don’t really want customers I don’t
43:03think spending that amount of money um in the cloud because it risks it’s not getting not getting a result that
43:10they can showcase yeah exactly it’s a risk to them you know that if that noise starts to happen where
43:16it’s not worth the money it’s more expensive and all this other stuff it’s like well
43:21you can do lots of things badly not just the cloud lots of things can be done badly and cost a lot of money it doesn’t
43:28mean you know that you go backwards you need to kind of work out the right thing
43:34so I don’t think the car vendors are sent are incentivized to either to want it so I think the talent
43:42lacking and the groups and the roles it probably needs to be more community centricity
43:47but I think the roles like you’re saying with you don’t if you can’t clarify the
43:52roles then you’re not gonna know why you need them other than the fact
43:58that you probably need them because you are they because other people have them so you’re like well that’s just what
44:03people have so therefore I’m just gonna get the same um but
44:09again it’s not down to a strategy so I think with customers you are you helping
44:14do you help them with the strategy then do you get involved to try and coerce the right conversations
44:21so that you can at least help the customers properly you know rather than just do we
44:28do we try and give them our opinion especially if it’s early and and you know they’re not
44:34suffering from some cost fallacy and it can be difficult right if you’ve already committed if you’ve already if the CEO
44:41has already gone into a strategic relationship with someone or you know they’ve already you know spent a year
44:47working up this plan but yeah the sooner that you can get it back on track and that’s that’s what we try and do
44:54we we generally discourage them from trying to build platform engineering teams
45:00in-house um and that’s why we exist as an organization is because that you know like we talked about before that should
45:07be temporary and yeah we we try and build communities as well so especially when we do government stuff
45:14um you know they’re not competitors right so they often see each other as competitors but shouldn’t
45:20um so we try and get them together as you suggested so that they can follow the ones that have got the
45:25success patterns um I do think there’s like perverse incentives right to like you rightly put
45:33your finger on if um if you’re trying to attract Talent then the mindset is okay we have to project
45:38success and um you know there’s then this race condition of who’s the furthest ahead
45:45and like you like you suggested there’s a there’s a cherry picking of only the success cases which doesn’t really
45:51represent the reality so there’s those perverse incentives and then obviously recruitment agencies are typically set
45:59up that they make a percent of uh of a placed head Council so the higher that
46:04price then the better they do obviously the contractor or the individual does better if that price is inflated so I
46:11guess it follows the hype curve like almost everything else in that sense I’m sure I’m sure it’s not different to any
46:18other hype curve or any other rule but I don’t know you notice it and I noticed that it does seem quite profound in this
46:24particular case um but yeah you should be driven they should be driven by strategy not by
46:31recruitment do you think the industry is being honest then about the problems that
46:36exist as in like you know all the all the terms as well like all the
46:42different Paradise like get tops and talking about that but let’s be honest that’s like not really anything new just
46:48infrastructure is code rebranded um but then it doesn’t really tell you
46:54anything it’s just like a way of working kind of and then again I’m not sure even how
47:01how mature the strategies are there like you know how do you when how does that work at scale
47:07you know how what happens when you’ve got multiple environments and a multiple Cloud accounts multiple environments per
47:13cloud account how do you do all that with Git Ops are you branching are you like I mean is that all these different
47:19things that maybe haven’t been thought thought about but everyone’s talking about how amazing it is yeah and like
47:24you need to do it and you’re like a bit like devops was and all these other things and it tends to be a bit tends to be a bit like that in the
47:30industry it’s all quite hyped up as if everything’s got the answer in it and it’s like but so do you think there’s
47:37honesty you know as much transparency and honesty in the industry that people are
47:42talking enough about what isn’t working um
47:47well obviously it’s not like a singular uh conscience so I don’t think like it’s
47:54it’s necessarily honest or dishonest but it’s exactly as you said it’s it’s more attractive to talk about the
48:00latest shiny thing and the hype and um I don’t know like for me I often think
48:08can we not just concentrate on getting like what’s in the agile Manifesto like
48:13you know working software before documentation like we don’t need a new principle like we have that very old
48:19principle and how how many organizations are observing it or you know test driven development or I guess even devops is
48:27relatively old now um so yeah I would rather say
48:33you know things like XP and I would rather see some Mastery or at least
48:39at least a high level of attainment on some of these bold patterns that really
48:44matter um before before being driven by hype on
48:50particular implementations and um so there’s no doubt right when it be something like kubernetes I mean you
48:57know you see you might have seen that deal about cartoon uh to take off of it and you know so that can suffer from it
49:03as well but I think the difference is there that it’s that it’s already ubiquitous right it’s the only game in
49:09town so for that like you know if you are going to do microservices style
49:14architecture and you’re going to hire real humans to come and do it like it is going to be kubernetes right so so I
49:20think once they become ubiquitous that that type argument goes away right but um yeah I think for for these other new
49:28shiny things I mean they’re all great actually I don’t think there’s any there’s any Pretenders out there that
49:34have you know some are more useful than others and you can argue that all day but but something like get and then get
49:40Ops and infrastructure is clear I mean they’re all you know when you read through the best practices that they’re all
49:47Noble and you should be doing them but I would just ask yeah have you mastered you know is the organization agile first
49:54have you adopted kind of immutable infrastructure first are you following even git
50:01um and that’s often not the case right often you have these different site it’s so ironic right but
50:06you have these silos that the organizations are more siled than ever in some ways you’ve got all these different Clans are there in the
50:12organization you have centralized platform teams as you alluded to and they’re they’re adopting their own thing
50:18so one of the things we try and use is like um is like a technology roadmap and a
50:24technology radar and we try and we try and centralize the design
50:29Authority in the decision making and bring experts in from all of those different teams and you know it can be
50:35seen as putting putting the brakes on and dampening The Innovation but but really they’re off
50:41customers often trying to do too much and not getting anything done so we bring it back to okay like what is the
50:47principle that we’re trying to adhere to are we trying to test first or you know are we trying to improve agility and
50:53okay should we just maybe talk about these foundational blocks in our opinion
50:58that you need first so that’s that’s one thing that we try and do is just bring it bring it back down to Basics and
51:04often it’s hard right because they’re politically there are already three
51:09years into that journey and they don’t want to go back to that you know but they’re not really going back they’re really mastering something
51:16that they should have mastered right at the outset so that’s the challenge right if politically part of the organization thinks it’s very you know in your
51:22example ready for githubs we would have you actually got infrastructure as code
51:28everywhere yet are you doing policy you know are you actually doing get in development successfully yet
51:35yeah it’s your spot on there I think that’s good that you’re doing that because you’re right I suppose it is
51:43yeah if you’ve got so far to admit that you’re going too fast in too many
51:48directions and you need to reorder yourself it’s just basically like regrouping
51:54um but the education I think that you were saying if there’s a skills deficit um in the industry
52:00like anything you’re only as good as the talent that you managed to get in that’s the same with anything you could you
52:06could like to literally Builders any Tradesmen like literally anything right the job’s going to be is done well as
52:12the person and the people that you’re bringing in yeah but the education though the guidance and the for the
52:18customer on what it all means and the principles you were saying like what’s the value out of those ways of working
52:25and why are they important and to educate everyone on it first before you then Implement and do
52:31um so that everyone’s on the same page and there’s no illusions on like what it actually meant
52:37um because it’s funny because I think I’ve been in meetings where people have different perceptions
52:44of what the things mean you know and they’ve been doing them for a while and you could be in a meeting you can
52:50clearly tell that everyone’s got a slightly different view of roughly the same types of things so people would be
52:56saying words but what they’re meaning to someone else is slightly different to what they might be doing with someone else I mean you’re all you know really
53:02the rooms in this like state of confusion about what it all actually means and you can see that you think
53:08well if that’s in your company then yeah you’re probably not going to get very far because you’re already starting off on
53:14the wrong thought so um yeah yeah and and so what do you see next then for
53:19yourselves at codification here um
53:24progressing with the clouds doubling down on the platform engineering side helping customers more on the agile
53:31aspects or a bit of everything well the best the best relationship for us is when we’re offering all of that
53:37and we’re like a trusted partner and the customer accepts that it’s a
53:43complex challenge it’s a multi-year challenge so
53:49where we can have a relationship where we can honestly assess where they’re at so that’s one of the tools that we use
53:54is you know you can make those arguments but ultimately they can be political that can be another supplier there or
54:01even when there’s not very opinionated people in the organization like you rightly said you can have
54:07you know multiple people using a word but meaning different things by it so so one of the things we try and do is like
54:13the terms of reference and we try and use you know standard you know try not let the
54:19organizing come up with its own terminology but use standard tools and standard language to describe those
54:25tools and standard methodologies that are that are proven um and we try and do
54:31um assessment somewhat by stealth so so rather than going around saying okay
54:38we’re going to assess everyone we’ll often be engaged to do narrow form of assessment which we’ll
54:45then broaden out so we’ll be we’ll be engaged to do like a platform maturity
54:51assessment but then we’ll broaden it out into saying okay how is the testing being done what’s the self-service like
54:56for developers and then you’re actually into people in process
55:01um so where we can get that level of trust to be able to do those sorts of
55:07things we can get much better results and then especially if they engage us to kind of assess everything out in the
55:13open which is which is rare to be fair um because we’re a because we’re a small
55:18startup and not that well known so that can happen and you know maybe we’ve done some small one-off projects and then
55:25we’ve built up the trust and then you know we’ve made our way up to the kind of senior leadership and then they say
55:30okay you know tell us where we’re really at and then what we can do is compare the
55:35scores on multiple narrow things and say okay when you when you compare these in aggregate it kind of gives you a level
55:41of maturity let’s compare that with our other customers
55:46let’s compare that with people in your industry or different industry or where
55:51you’re trying to get to um so that that can work giving them this like objective comparison so
55:57obviously we don’t we don’t disclose who that other customer was or break any kind of confidentiality there but we just say
56:04okay another another client in this sector this is how they scored this is where they’re up to on their Journey
56:09this is their perception of themselves before the assessment this is the reality of what the assessment showed so
56:15we do that sort of thing and we do um like lab-based training of their
56:21people as well so we can actually get the capability assessment there so they send you know again we can ask them
56:27where do you think this team is at okay this you know after six weeks of instruction this is how well they
56:34followed the instruction this is how well they completed the assignments this is how your team compares to a different team in the organization that’s maybe in
56:40a completely different leadership structure different incentives so they can find that you know eye-opening as
56:47well and then again you can compare the capability to the outside world so
56:52that’s what we like and and we can compare new joiners that maybe have been through some kind of codification
56:59bootcamp before joining and then put on the project and then assessed versus someone that they’re trying to upskill
57:05or someone they haven’t given the proper time to achieve the upskilling so all of those things we can
57:11we can kind of help them with so it doesn’t matter to us if they’re really
57:16struggling or they’re really Advanced you know we’re our team is technically always in advance of their
57:22team because they’re not a technology company they’re they’re trying to become one so so we’ve got no
57:28no real strong opinion of what they ought to be and we just try and show them okay this is where this is where
57:34you thought you were and this is where you really are this is where you’re trying to get to and this is the gap and
57:39then we can help across the board right we can help by augmenting them with Engineers be it platform Engineers or
57:45Cloud Engineers or whatever we can train retrain the teams or we can hire
57:51hire new permanent staff for them nice and if somebody wanted to get in touch how do they get hold of
57:57yourselves if they’re like listened and like oh that is what we need like we’re in a bit of a tricky situation at the
58:03moment and maybe need to get yourselves in the website codification.io would be the
58:09one to go cool um all right well um it’s been great chatting to you
58:15Richard um and thanks for spending the time obviously to come on the podcast and talk about these things with everybody
58:21hopefully somebody got something out of it um who’s listening and obviously as you
58:26heard you can reach out to codification.io if there’s things there and you think that can help them please do reach out so awesome thank you John
58:33thanks cheers, thanks bye
Part of the Appvia family.