Arun Narayanaswamy, Amadeus

Thank you for tuning in to today’s episode of the Breaking Changes podcast. I’m your host and chief evangelist for Postman, Kin Lane. With Breaking Changes we explore topics from the world of APIs, but through the lens of business and engineering leadership. Joining me today we have Arun Narayanaswamy, director of engineering at Amadeus Labs. Listening to Arun talk about the relationship between APIs and the cloud took me back to the early days of my API work in 2010, while also helping me understand the symbiotic relationship between being API-first and being cloud-native. Let’s start with the basics: who are you and what do you do?

Hi everyone, hi Kin. I’m Arun, I’m more of a cloud and a DevOps evangelist, that’s what I call myself. My responsibilities have been in transforming companies into DevOps ways or cloud ways of working. I’ve led large transformations in big enterprises, and obviously started off my career as an entrepreneur, so I sort of have that zeal to keep making changes in a setup, bring in new things in everybody’s life, and focus on developer efficiency, which I believe — if we deliver great to our developers, our developers can develop great to our customers. So that’s typically who I am, focusing on open source cloud, multi-cloud, hybrid cloud kind of experience that I bring with.

I want to continue with the basics, beyond just you, for my listeners, because this show tends to be for leadership and folks who are trying to make sense of this digital transformation that we’re all going through. Let’s lay some fundamentals. What is DevOps?

Well, okay, on a very light note, it’s the most abused term, right? The way I would put it, it’s very clear, it’s a very large bucket of things which helps the organization transform. I would classify them as largely a cultural change that the organization has to bring in top-down. Everybody’s assumption, when we talk about it from a leadership point of view, is they believe DevOps is like a ground-up thing, saying, oh, somebody is doing automation and it’s going to work. But yeah, it’s the convergence of a top-down initiative to say, as an organization I want to make a change, help our developers and customers move fast, and then obviously a grassroots initiative where people are thinking automation-first, people want to codify everything, from infrastructure to configuration to the code obviously, and the convergence of these two is DevOps. It’s a very different take to it. When people say it’s a convergence of dev and ops, I would say it’s a convergence of culture, from a top-down initiative and a bottom-up initiative to help move things faster.

I really like that definition, because that’s exactly where I’m going with this show and trying to reach leadership, because I think it has that perception of being just purely bottom-up, but I think you’ve probably seen it, without that top-down it’s not going to be what it needs to be. So I think that’s really an important definition. The other kind of fundamental is, what’s the reality of the cloud transformation and going cloud-native? There’s a bunch of terms folks are using when they describe this. What are you seeing when it comes to cloud adoption at enterprise scale?

Well, okay, I mean, again, like DevOps — thankfully cloud is not an abused term, right, it really drives value instantly. But again, at an enterprise level, unfortunately it’s sold to an enterprise with a mindset of saying this is going to save cost. And again, maybe the big cloud providers don’t sell it that way, but it’s bought into an enterprise with that intent. Having said that, thankfully most of the enterprises realize that that’s not the only thing, and the biggest transformation which cloud brings in is, I would say, twofold. One, very importantly, is you being close to the customer. You could set up — instead of having a data center in one location and serving the globe, if you are an enterprise serving customers in all different locations, you could set up your cloud data center or a cloud setup closer to the customer, so that sort of enriches the customer’s experience, helps in data residency, which is also quite important in the current scenario and what sort of panned out in the recent times. The second bit, which is again when I focus or have the developer mindset, or the developer efficiency mindset, cloud gives you the flexibility to work on fancy new stuff without large investment, right? So a developer could quickly try out the latest technology, quickly try out, let’s say, serverless, the latest of hardware, ARM processors, whatever they might be, without seeking for a large investment, pilot it and showcase what they want to do, which is the real or the true feeling of agile, which comes with this advent of cloud. So a slightly different take of why cloud transformation is happening, but it’s twofold, as I summarize: one, to entice the developer, and two, to entice the customer by being closer to them.

That makes a lot of sense. I do hear though some people have misconceptions that this means that the data center is going to probably go away, that everything’s going to move to the cloud. Do you feel that’s a reality, or is the data center fine, and the cloud’s going to be for bursting, for these data residency, these different purposes and use cases, and they kind of coexist?

Oh, I would say, Kin, this is quite a tough question to see from an enterprise viewpoint. Companies that are focusing very heavily to say being close to the customer is the key, they go very heavily towards a cloud-only approach, because your data center probably will get old very soon. You’ve bought your capex, or you’ve put your investments three years, five years ago, and the kind of demands that you have from the customers these days — if you’re bursting or you’re setting up your Kubernetes instance, or you’re setting up API-first kind of development, all of those are easier done in the cloud. When you start focusing, and when your infrastructure team starts focusing on cloud, the investment starts going down on your data center, which eventually means it becomes old and not maintainable. So in those scenarios I would say enterprises are heavily moving towards cloud-only.

But again, when I used to manage the cloud center of excellence, one of the bits that I realized is sort of an intermediate approach where companies are taking to keep data, or the database, on-prem, your logging, your monitoring etc. happening on your on-prem instances, whereas your, let’s say, the compute is happening on cloud. So this is the second set where both cloud and the data center is going to coexist. The third, which is also a very known model being adopted, is bursting, bursting on both fronts. Bursting in terms of, let’s say, an e-commerce website trying to manage peaks, they are obviously using cloud to scale up during some seasons. But then there is the other segment who bursts during the development cycle, so they say my production is stable because they’re potentially a B2B ecosystem or even a B2C ecosystem where they have a predictable load, they set up production instances on-prem, but they have the test instances, they have the dev ecosystem etc. sitting on cloud, so they could kill those instances as soon as they are done. So it’s a mixed bag in the enterprise segment, where all combinations of this exist. Having said that, one of the key things which we sell heavily to our employees is, irrespective of whether it’s the data center or the cloud, the skills sort of remain the same. The people, the skill is needed, maybe where it’s hosted, the underlying infrastructure could be changing —

Sorry, Kin, I got here, sorry, there was a weed eater, a lawnmower outside and I went mute, so let’s try that again.

That’s why we’re not live streaming, so we can do this.

Sure. So that makes a lot of sense. I would say it really depends on the use case and the type of resources you have on hand, the staff, how they’re going to be able to do things, what your customers are needing, and I would say the industry you’re in really influences it. One of the things I hear a lot, and I’d love to get your opinion on, is what’s the reality of multi-cloud? Can you be multi-cloud, and what are the reasons why enterprise organizations are exploring being multi-cloud?

Okay, I mean, this is a debate that starts on day one. The reason being, most of the organizations, large enterprises for sure, are transitioning from their data center to cloud, which means they have a hundred percent control on their data, on their infrastructure, on their security, on their skills, everything combined, they have the control on their own data center. When they move to the cloud, the first conversation is, oh, if I choose X cloud service provider I will be locked with it. So, let’s say if a particular cloud service provider, it’s obviously sold saying we have been reducing cost year on year, so seven years from now you will be paying much less than what you are today. But then when you put the finance angle to it, when you shift from a pure capex model, where you’re able to put this very nicely in your balance sheets, but when you move into a pure-play opex model, this dramatically changes.

So in a hybrid cloud scenario, it’s the conversation which starts off in the beginning, like, should we use one cloud service provider, should we use many, because we are not locked to a particular cloud service provider. But as you keep exploring, you realize the investment needed to move to a single cloud service provider is slightly lower, because one, you need a specific set of skills. Though, let’s say, a GCP or Azure or an AWS have similarities, but there are skills or niche in the market in terms of the architecture, in terms of some of the embedded cloud-native solutions which are very, very specific to a cloud service provider. You need skills which have to cater to all, so the investment in terms of skills dramatically increases if you start choosing multi-cloud.

Second, which is again very, very complex in this ecosystem, is how do you make your application multi-cloud ready? It’s easier to say, okay, I adopt, I know this is Azure, I use the whole ecosystem, Azure DevOps, maybe some functions on Azure. But then, if you’re using Azure Kubernetes Service, maybe you do not have an exact replica. If you have built everything for your Kubernetes service, is it as simple as taking this and putting it on the AWS Kubernetes service, is it an instant lift and shift? It’s not true. So there’s a lot of work that needs to happen to make it, one, multi-cloud ready, two, cloud-native, and then it’s not just the application, it’s the skills, it’s the database, it’s the connectivity from your data center, how many pipes are you going to put from your infrastructure to reduce latency, and maybe in some of these setups this is not really required, but when you get to a low-latency kind of compute needed, then a lot of these things come into picture, saying, do I dedicate focus on a provider and go full stream on it?

But again, companies who have successfully been able to do it are, one, who have built their own platform, which is basically a layer above multiple clouds, and say this is our way to deploy onto multi-cloud, so scaling across different platforms. The second, where people have been able to completely make it cloud-native and have a good exit strategy — it’s not an entry strategy that’s important, they need to have a good exit strategy. Have they automated it fully, can they completely demolish the instance on one cloud service provider and bring it up on some other cloud service provider in a fully automated fashion? Obviously, which means they have put a lot of investments behind the scenes. If they have done these, they are successful in a hybrid cloud scenario. Otherwise it’s basically, I use service A on one cloud service provider, service B on the second cloud service provider. It’s not like a hybrid DR scenario, but it’s like I have used it spread across different things for the best of breed.

Yeah, are these — you mentioned skills quite a bit in that and previous descriptions. Are you able to, are you hiring and able to find this type of talent externally, or are you having to train people up to meet these needs?

I mean, again, strategy is very key, Kin. This is my third company where I’ve been playing the role on the cloud side, every company has a different strategy depending on where they are located and depending on obviously the availability in the market. Taking this simple example, in my prior setup we were getting very good people in the market who are focused on a single cloud service provider. They would say — which is how things were in the past, I’m a Java architect, I’m a C++ architect — now we are seeing segments of saying, I’m an AWS specialist, or I’m an Azure specialist. But people who have taken a multi-cloud journey are unfortunately one level higher, I mean they’re more superficial, in terms of, I know all services, I know how to move to the cloud, but when you get there they are not really able to build platforms which could scale across two things. So as a combination of what I said, companies are retraining, so there is a lot of investment going into helping people scale up on multi-cloud scenarios.

It’s not just the developer that you need to train. In some cases you need to train the customer, because you need to tell what really this brings, how are, let’s say, the billing model going to change, because maybe we were saying you need to pay X amount to us for a transaction, in cloud maybe it’s a different thing, we say okay, when you scale up this is what you need to pay. So it’s skills beyond developer which is there on the cloud, and with respect to beyond developer, it’s not just the customer you need to train, the QA, because the QA has to test in different environments, you need to retrain your infrastructure people, because managing AKS is different to OpenShift, being very, very different. So it’s multi-dimensional skills which are needed in the market right now.

Yeah, it’s really investing in the people side of this, and investing in being able to properly set up, tear down this, and do it architecturally. But you mentioned the cost part, and you did this earlier on too, because I remember the early promise of Amazon was, it’s going to be cheaper. I remember my data center, it was costly for me to do things in my data center, make changes, so the low-cost promise of Amazon was a big selling point in the early days, and then they stopped using that, because I think once the secret got out that your bill wasn’t going to be cheaper and it actually could change, and then Azure has a different approach. So do you need someone just paying attention to the bill? Are there ways that you’re having to help enterprise organizations manage costs at this level?

Oh, absolutely, I think this is, Kin, I would say one of the most important aspects. You’re right, Amazon did have the approach of saying we reduce costs on a monthly basis or a yearly basis, and one of the graphs that I remember is saying we have reduced AWS cost 70 times, which is quite interesting when senior leadership looks at it, saying okay, this is diminishing cost, so it’s great service. But one of the things that I’m noticing across all setups is, neither did Azure or Amazon tell any lies. They really did say the truth in terms of saying yes, the cost will reduce, but then if you do things right. If you behave like a data center on the cloud, you’re going to burn money. You need to redesign for the cloud, you need to think cloud-first, and then obviously the cost benefits are coming.

But stepping back onto the topic that you mentioned, one of the things that’s really in focus these days — I think this was in the state of cloud report as well — the number of companies investing on a team which focuses on cloud cost is dramatically increasing. When people move to the cloud, the first thing that they thought, okay, this is something now it’s 100 percent opex, I will look at some dashboards and I’ll make the cloud service provider manage this, was wrong. So they are investing on either bringing in a FinOps team, cloudonomics, there are various names teams adopt, or they even call it the cloud ops team or the cloud finance team, and this team focuses on a proper monitoring, dashboarding, budget management, reporting back to the CFOs, the finance structure, in terms of how the old capex has now transformed into opex.

And last but not the least, which is something that I really enjoyed in one of the organizations that I work, is pure-play gamification of this. You would pitch one team against the other, saying okay look, your cloud cost was X, this team’s cloud cost was Y, and month on month it’s a race who helps reduce the cost the most. Reduction of cost comes in by saying, have I put the right practices in place, have I put the right measures in place. Our product dealing with ten thousand customers will certainly have more cost than a product dealing with five thousand customers, but do we have the right measures to say what is the cost per customer here, cost per customer here, cost per transaction, cost per transaction, how do you measure these two and gamify and reward them? It’s a funny thing, you name them, shame them, report them, penalize them, however you want to call it. It works beautifully in companies where they take this positively and challenge each other to say, okay, my architect is doing better in terms of reorganizing my teams for betterment in terms of cost.

And, Kin, I want to touch upon one important factor, which I found very, very interesting. I will not name the cloud service provider, but this cloud service provider, we had an engagement with them, and very interestingly when you sign up with them they give you two people. One person is an architect, one person is an efficiency person. The objective of the cloud service provider with this person is to help reduce cost for you, and the second person is to increase cost for you, in the sense, increase cost is by adopting new services. So even with the cloud service providers, they put in these — genuinely try to help the customer in helping the customer move forward by putting these sort of gamification factors in those ways, and even those people are rewarded based on that. So if the first person, he’s more like the technical account manager, comes and says I have increased the footprint of 10 services to 20 services, he gets rewarded for it. But the architect, or the solution person, he is designed to say, how can I reduce the cost for this customer, by saying these people were using for this product 10 C4 instances, now they have gone into 8 of some other kind of instances, and it’s helping them reduce cost but better efficiency. This is something that I really love where that cloud service provider provides.

Yeah, those two roles are opposite of each other and pushing on each other to find that right balance, and help the customer strike a balance in the cost and efficiency. This really feels like a lot of what I’ve seen in DevOps, when it comes to the optimization for quality and productivity when it comes to development teams being able to move faster, release velocity, all these things, but you’re doing it from an overall architecture and cost perspective. That’s powerful.

Yep, cool, that’s true.

So one of the other — you mentioned earlier data residency, being closer to the customer, but within data residency there’s, I’m hearing a lot more packed into this, and it could be under data sovereignty, I’ve even seen data nationalism used, regulatory mandates that you have to operate within certain regions or zones, and I know that the cloud providers feed into this in various ways. So is regulation a big driver of the cloud strategy you see across enterprises?

Absolutely. Again, this is one of the key factors that comes in. Let’s say a US-based company, if we say we are developing largely in, let’s say, India or Philippines, and their customers are Europe, it’s always a debate to say the company originated in certain locations, so why don’t we host this data there? But then, with regard to being closer to the customer, you would say this data needs to be closer to maybe Europe or wherever the customer is, or the developer, because the developer spends most of the time doing those evolutions, working with the cloud — where should these be? So in terms of enterprises, in terms of what domain they are in, if it’s obviously a banking domain, they want to make sure data residency is very, very, very important, all the policies, governance etc. is highly important. If you are in healthcare, then obviously this becomes very heavy personal data, then you’re looking at it in a factor saying, is it even safe to put this data on the cloud?

But very clearly, Kin, a simple answer to your question is yes, all of these play a part. But even higher is the thought process in the organization, the culture in the organization, in terms of saying what do they want to protect, what are their customers asking you to protect, and obviously what kind of relations you want to keep, and how close you want this data to be, how fast this data needs to reach them. So a combination of all of this. Again, all the GDPR, the OFAC licenses, PCI kind of content, all these play a role. And again, the bit we touched upon earlier was a conversation as a cloud evangelist with the CFO, in terms of convincing how cloud is cheaper and effective, but this is with the CISO. So you will have to — the cloud evangelist’s job is not just convincing a developer, but it’s convincing a CISO, convincing a chief financial officer, convincing the CEO in terms of being closer to the customer. Everybody plays a role, and all of them have factors to play in before this decision has been made.

Yeah, I was reading through a cybersecurity report from the US federal government, and they were providing a calculation for cost of data breaches, and it was based upon, is it PCI payment information, is it PII personal information, or is it PHI personal health information, and then which one of those it is, and the number of records, so scope — are we talking thousands, are we talking millions — and then the price goes up for the responsibility or the cost of the data breach around that. And then that feeds into, you mentioned GDPR, CCPA, so can you actually find those records, do you know the scope of the data that you’re storing? So security is always top of mind, you always hear people paying attention to it, but it’s really hard to quantify in this cloud-native, new Kubernetes, this hyperscale world. What is security from your vantage point, what should organizations be paying attention to and investing in?

Okay, it’s a tricky area, Kin. The bit that I mentioned, it’s where your data — again, all aspects. How secure the feeling you had that your data center was secure, and that perception to move from this to a factor of putting this on the cloud. Again, this is a very well-known factor: the number of accidents — you’re more prone to an accident sitting in a car versus sitting in an airplane, but everybody fears sitting in an airplane and not sitting in a car, because you have less control of things in the cloud, and that’s a perception. The security is no different. When you’re thinking of security on your data center, it’s that homely feeling to say it’s 100 percent secure, so you focus on putting the right firewalls, you’re putting the right controls, and then obviously there are certain segment of people who have all the access on that data center, so it’s the perception of being secure. But when it moves to the cloud, obviously we are looking at a larger surface area of attack, because it’s not a one-time move that you’ve completely shut down your data center and you go to the cloud. So for a period of time you’re transitioning from your data center to the cloud, so you’ve sort of doubled your surface area of attack on a security viewpoint, and that phase is obviously quite complicated because you’re managing two setups.

But once you go to the cloud, the thought process always is, the cost of managing a very basic level of security is quite low. If you want the basic security principles which the cloud service provider gives, if you do things right in terms of setting up the right access control, the logic of least-privileged access, all of those, if you do it, I would say cloud is fairly secure. But the moment you get into the next level of security, trying to bring in — I think it’s Azure Armor, AWS has something similar, high-scale DDoS protection etc. — the cost blows up very, very heavily. But again, it’s not just the cost, it’s also about knowing what kind of attacks are possible on the cloud and how do you contain them. So I think these are factors playing above and beyond just the data residency. What kind of attacks do we see, are we seeing new-age kind of attacks?

But there is one segment of people who are quite happy with this structure: people who are really, really automating, thinking DR, thinking high availability even before, has a higher priority on security. If you can completely rebuild your infrastructure in a matter of seconds through your automation script, completely codified applications, they are less fearing the whole security thing. Obviously the security on data loss continues to exist, and that’s the segment of people who are trying to look at, let’s say, the encryption during transit, encryption during rest, and all of those kind of stuff, but again, very heavily focused on putting the data in your data center and hoping that it’s safe. So that’s, I would say, a very safe approach, but the logic of security is multi-fold in terms of skills, in terms of you investing heavily on the cloud service provider, beyond just data.

Yeah, I really like your response, and I apologize for putting you on the spot there regarding security, it wasn’t part of our original questions, and I know that’s not your main focus, but you brought it home beautifully. Because from my vantage point in the world of APIs, a lot of people like to talk about penetration testing, authentication, all the obvious encryption, as you said, but really the number one security problem with APIs is enterprise organizations just don’t know where all their APIs are. They don’t have control, they don’t have discoverability, they don’t even know where they are. So it’s not a matter of, oh, I’m sure there’s vulnerabilities in there, but you don’t even know what’s powering your web apps, and you don’t have that topology control that you talked about. Because if you have that level of control to tear something down, put it, stand it up, you know all the moving parts, you know where all your APIs are, you understand your infrastructure, you’re using Kubernetes. So just having that level of control over your infrastructure is going to help make your world much more secure, and then you can start addressing all of those actual known knowns when it comes to security, because the unknown unknowns are a lot less.

True. I think one of the points that you brought about which is very important, which we miss, is, in the new-age world, when people are developing fully backward-compatible kind of applications, they have not just the code but, be it APIs, be microservices, toggle-based development, all of those, the amount of code that gets unused, or things that sort of become the functionality that starts getting unused, becomes quite large, and people lose focus. They just switch off the toggle and say, oh, this feature is switched off, but then do you really know that that part of the code, which is not being accessed and not available from a UI standpoint, is that vulnerable? This heavily happens, and when I say code, obviously these mean APIs as well. You have some redundant APIs which was catering to a customer, and that customer no longer exists, but you just switched off an API, but then the code really exists behind the scene, and that could be an open loophole, which is very difficult to invest and keep investing, saying okay, I have all these, how do I keep fixing it. And associated to that is documentation, right? What have you put in that documentation to say what are you giving up, or in terms of what can people external to the company look into and then see it as a potential area of attack? So beyond code, it’s got to be visible.

Yeah, it’s got to be visible, observable, documentation’s a piece of that. But you describe many of the high-profile breaches I’ve heard of. The Equifax data breach was stood up for a contest, it was an API, and the feature was turned off, as you said, but this API still lived out there, it didn’t get the patch, the Java Struts patch I believe on it, it didn’t, and left that vulnerability open. It didn’t have the right policies, so once someone got access they were able to get access to a large number of records. There was just many layers there, and it’s just based upon lack of visibility there.

So as you’re evolving, as you’re doing what you do on a daily basis, are your challenges mostly technical and dealing with architecture and processes, is it more business and getting the funding and making sure these things are properly invested in, or would you say it’s more political, and politics and people and culture type changes?

I think, again, all of them come in phases. The first one is when you’re working on the buy-in in terms of saying move to the cloud, or move to have a big digital or cultural transformation, there is a huge set of resistance. There is a resistance in terms of people, largely people, having people at the developer level saying, oh, I don’t want to learn something new, there will be resistance at a management level to say, oh, this is too much of an impact, is the same the focus for us, because functionality is more important, moving to the cloud is not. And at a management level, which is, we just start talking business, they’re saying, oh, I have very clear visibility on my data center, I have control over my data, I have control over my people, why do you want to shake the tree? So it is a combination of all three, largely political to begin with, because there is resistance.

Then it becomes a people challenge, because you have to move early adopters, you have to give them things that are really, really fast for them, because you need them to keep churning out something new and become an evangelist within the organization. For the laggards, you need to help them give right business cases and help them move, so it becomes more of a people challenge next. And the third, which is a technical challenge, because in this whole storyline you’ve probably oversold, or you have underestimated the change, because you have to make — and we discussed all these points — you need to relook at architecture, how cloud-native do you want to be, do you want to just lift and shift? There are companies who would have over-promised, or you take the whole infrastructure as is, put it on the cloud, lift and shift, absolutely you’re not going to see any cost difference, in fact over a period of time you’ll start losing money, but that’s like the fast decision, long-term cost impact. But if you want to take high impact at the beginning, it means a lot of technical evolution, you need to have the skills, you need to know what needs to be done, learn new things. Again, cloud is really finding a needle in the haystack, everything is available but what is the right thing for your need. So a good partnership with a cloud service provider, working well with the architects from the cloud service provider, all of that helps in the third phase when you go deep down tech.

Yeah, it’s definitely all-hands process, and the entire org’s got to be at this, this has to be top-down, bottom-up to do right. And based on the industries you’ve worked in, is this something all enterprises should be thinking about, investment at this level, or is it just if you’re a heavy tech-focused enterprise organization, or is this pretty much any enterprise should be thinking at this level to do business in today’s climate?

I would say the fastest approach, Kin, is to find those opportunities within your organization which can move fast. So let’s say a greenfield project, you’re building a completely brand new project, brand new customer, take that plunge. Across any enterprise I think this should work, even if it’s a banking domain which is heavily regulated. I think one of the easiest things is to look at that greenfield project which you can start building afresh on cloud, use these success stories and sell it within the organization to say, look, this is how a successful model will work.

One of the very interesting factors — I think this is when I was talking to somebody from Netflix or Spotify, I don’t remember — but one of the key things that they said is, if you copy us you will fail. So every company, every enterprise will have to figure out their way of working in the cloud, for everybody it’s different. So the best approach is greenfield, the second best approach is, certainly not lift and shift, but re-architect to the cloud, take some applications where customers, people, all of them align to that goal to say yes, cloud is a good thing to do. And only when you have good evangelists within the organization, try making it a large-scale process within the setup, otherwise it’s going to blow up in your face, in terms of saying, oh, we took a huge plunge to the cloud but then we are struggling. So you have to work in that approach. And again, this is something which I spoke about from a banking industry standpoint. For e-commerce it’s very different, primarily because they are more B2C, and for them the first-mover advantage really, really matters, so using the latest and greatest flexibility that they get in the cloud is always a good thing to adopt and move.

Yeah, I think that B2B or B2C is really going to dictate, because you said it early on, you have to explain this to your customer how things are going, you have to keep your operations in alignment with your customers and what their world looks like, otherwise you’re going to fail. So this is a producer and a consumer relationship that has to move forward in equal time. So as part of that, the world’s shifted in the last two years pretty substantially in various ways, in different industries differently. How has COVID impacted your world?

Oh, I mean, again I’ll take the same people-customer view. On the people side, on the skill side, most of the organizations have gone remote. So when we look at people working in different locations, they have demands from different locations, so if you have one data center, if you’re looking at developer efficiency, having one data center having all these dev infrastructure sitting in one location has its own problems, you’re probably not as effective. But when you look at the customer side, there is one large segment which is, everybody wants to be digital-first. The kind of transformation people saw, people expect it to happen in the next 10 years, they want it all crunched in the next one, saying, oh, I want this now, because my customer has become more demanding, my customer wants fast-paced development, they want to be more agile. So there is a huge demand coming up. I’m excluding all the negativities that come with COVID-19, but all the positives that it comes with is faster attraction towards transformation, faster traction towards adopting or thinking in a new way, where people said, oh, DR is not really required, now really, no, DR is required. And with potential, not directly with COVID, but also I think a lot of security attacks have increased, all that makes a lot of difference in terms of upskilling your people for the new demands that are coming in. So I would say COVID has done an impact, in the short term negative impact, but very positive impact in the long term.

How do you — we talked about upskilling, and you just talked about staying in tune with what’s happening — how do you stay sharp in knowing what’s coming next, and keep your skills relevant to the companies you work for?

I would say hiring the best talent, they keep you on your toes. If you hire them, if you get good people, they push you up along with themselves, but that’s certainly the mantra for being in touch. But then also being, I would say, connected, building your network. Building your network in the new world is very, very different, because you’re not attending conferences in person, you’re not personally meeting people as much as we did in the past, we’re not even networking with people as much within the organization. So maybe we restart some of those, but I think conferences, listening to great podcasts, reading through a lot of content. And again, culturally different, Kin, is how much are you contributing back to the society? The more you give, the more you get. I think that’s one of the big things that I push my people to say, go speak in conferences, it’s not about attending, but go speak in conferences, run your own meetup groups, do your workshops, so you share more data, you get more in that process which you don’t realize, but if you go in the hunt of learning more, you’re limited by opportunities.

I like it, and I would add, join people on podcasts like you’re doing with me.

Absolutely.

Because this — I would say Breaking Changes has become my networking opportunity and meeting smart folks and learning from smart folks like you. So this has been great, I learned a lot, I really like the financial, you called it earlier FinOps, being able to manage the money side of this. This is something I would love to explore more, and I think I might ping you and have you back and have a deeper conversation on that. We’re — this will go out as a regular episode, but we’re doing a lot of other segments derived from the conversations we have, leading up to a conference next year as well, so I’ll be in touch further to see how we can get you involved. But I really appreciate your time today, it’s been pretty enlightening.

Thanks a lot, Kin, it’s a pleasure speaking to you as well. You did dig into corners of my brain, so thanks for it, and happy to contribute more.

Thank you so much, appreciate it.