Sanjog Aul [00:00:10]:
Hello and welcome to the CTN Virtual Roundtable. The topic for today is decoding complexities and forging ahead in a hybrid and multi cloud era. It’s made possible by Zones. And I’m Sanjog Aul, your host, and I will be helping facilitate a rather candid, collaborative and conversational dialogue here because this is a very important topic. So, coming from media, we have been talking to leaders around the world about the top challenges they are facing when they are trying to scale their efforts. And to that end, what we have learned is there have been challenges related to the sheer complexity that it brings when you’ve got multiple cloud vendors or the number of things you’re trying to do with cloud. That’s one.

Sanjog Aul [00:01:01]:
Then we have seen a lot of spend management issues people have brought up, people have spoken about security challenges, and then now the AI is in the mix. They’re also talking about the privacy and governance related ethical concerns and frankly, sustainability. It’s surprising, but it’s for the right reason. Sustainability is also coming up as a very key topic because while we are trying to do best with our business, we also need to take care of our planet. Now, these challenges exist, and in some form or fashion I’m sure all of us might be facing. But I’d like to dig into specific details of what is your top challenge as you’re trying to make the most out of the cloud investments. And to that end, what are you doing to solve the problem? So, to discuss this, I’ve got a great panel here. So I have Jonathan Bagnall, head of product security with Fresenius Medical Care.

Sanjog Aul [00:01:53]:
And then Sushmitha Nalamothu, director of software DevOps at Northern Trust. We got Satish Kumar Chala, head of digital transformation office with Societe Generale. Riaz Yusuff, chief information officer with Harper College. Anshuman Mandal, vice president, practice head with cloud and infrastructure and DC, which is data center, at Zones. Brian Gardner, chief technology and information security officer with the City of Dallas. Arjun Puri, director of cloud data engineering and analytics at the Depository Trust and Clearing Corporation, also known as DTCC. And Tony Giampaolo, head of cloud and data center connectivity at Zurich North America. So welcome everyone. It is a great pleasure to have all of you. So let’s jump right in.

Sanjog Aul [00:02:41]:
And what I would like to do is to request each of you, in the sequence that I request you to share and contribute, share the top challenge. The top challenge with the backstory of why. What is that top challenge which is holding you back from exploiting cloud, hybrid cloud and multi cloud to the hilt? Now, along with that, what I want you to do is to answer this question. If the cloud environment you have was your pet animal, which animal would it be and why? So this answer should come along with your top challenge. So with that said, Jonathan, please go ahead. Jonathan Bagnall, please share your top challenge and don’t forget to answer that question.

Jonathan Bagnall [00:03:27]:
Thank you very much for inviting me. A great lineup of panelists here. So the biggest challenges that we face from my business background, and that is healthcare, is the ability to scale and the ability to implement holistically across the globe. And that challenge relates directly to regulations. And I would also say the challenge relates to the healthcare industry as a whole. We are faced with a legacy environment, we are faced with legacy products, and it does prevent our ability to scale at the pace that we would like to.

Jonathan Bagnall [00:04:21]:
Do I get to pick my animal?

Sanjog Aul [00:04:23]:
Yes. That is what I’m waiting for. That’s an important question.

Jonathan Bagnall [00:04:27]:
I didn’t know whether I should go large, big.

Sanjog Aul [00:04:32]:
Hey, I do not know what you want to have as a pet animal, so go for it.

Jonathan Bagnall [00:04:35]:
So I guess I would choose an animal that was fast and agile. So with that, I would pick the leopard. And when I think about cloud and hybrid environments and the ability to adapt and scale, that’s what I would want to align to.

Sanjog Aul [00:04:56]:
Great. Thank you so much, Sushmitha.

Sushmitha Nalamothu [00:04:58]:
Thanks, Sanjog, for this opportunity. First of all, it’s great to meet all these leaders and hear out their mitigation plan and challenges. It’s a pretty diverse group. Thanks for that. So having one cloud versus multi cloud is a little different in an organization. But it is quite natural and it is the usual norm that multi cloud is part of our strategy. But when we come down into certain business units within the organization, maybe they’re following one cloud. But when cross functional BUs are working with multi cloud, that’s where the lack of visibility and control across cloud environments is the challenge.

Sushmitha Nalamothu [00:05:42]:
Mainly I am seeing, working between different organizations, how we are struggling with that is the visibility and control over their environments, and multi cloud deployments and automating those pipelines to get into cloud and decentralization management, et cetera, et cetera. I can go deeper and talk about details, but I want to keep it very high level. So the policies that we were trying to look at is to ensure getting to a common ground between all these different providers, and also trying to have a common platform in terms of centralization of dashboards and monitoring and managing those resources between all these multi cloud providers, and also trying to have the security and compliance and regulatory match across these different providers. Basically everybody pretty much are trying to be in a regulatory zone, but not that we have a centralized place where we can automate one shot. It’s like different providers have different automation policies as part of code, or whatever you call it as. So this has been very volatile and we are trying to bring up a governance between multi cloud environments.

Sushmitha Nalamothu [00:07:08]:
Now, Satish? And you did not let me pick my animal.

Sanjog Aul [00:07:10]:
Yes, I think, yes, Satish, you got to wait. This is an important question. Go ahead, Sushmitha.

Sushmitha Nalamothu [00:07:16]:
I pick rabbit. It’s cute and fast. So cloud seems cute. But we want to run fast and be there with on prem legacy and all that. So it’s a race.

Sanjog Aul [00:07:27]:
All right. So you picked the right animal for the race. Absolutely awesome. Thank you, Satish.

Satish Kumar Chala [00:07:32]:
Hi, everyone. Great to be part of this conversation. So my top item I would like to pick in terms of cloud strategy would be sustainability. And when I say sustainability, the backdrop for that I would give is, as we all know right now, ESG is one of the hot topics in all of the industries in general. And we see that it could be potentially very soon one of the risk taxonomies. So when we say a potential risk taxonomy, none of the organizations so far had ESG as one of their metrics to begin with, in terms of reporting. The minute it becomes a metric, all of the adoption of the cutting edge technologies, all of the major initiatives within the organization have to be aligned to that metrics. And you have to revisit your strategy, your programs.

Satish Kumar Chala [00:08:29]:
And my main backdrop, or my main reason why I want this front and center, is that future is coming very soon. It could be around the same year, if not probably in a couple of years from now. I want to be prepared in terms of understanding what it means for us to track for sustainability, to be prepared from an ecosystem which is not impacted by other business demands. So that would be my top priority in terms of having sustainability as the ask from this panel. And going with the animal pick, I would go with swan. I would say the reason why I would go with a swan is I think it would go with the grace on how it would swim, but you don’t know the paddling which is happening under the water. So my program of sustainability is like you gracefully go into the transition mode in terms of the program, while you are really juggling with multiple priorities and still be able to have a smooth transition for your stakeholders and your clients.

Sanjog Aul [00:09:49]:
Satish, that’s an amazing analogy. Awesome. Awesome. So, Riaz.

Riaz Yusuff [00:09:56]:
Thank you, Sanjog. My name is Riaz Yusuff. I am a chief information officer at Harper College in the suburbs of Chicago area. The top challenge I have is, I think having a multi cloud solution is inevitable at this time. And we are no different. And we do have multiple cloud solutions. And additionally, we also still are keeping some of the on prem solutions as well.

Riaz Yusuff [00:10:25]:
And as a result, what this is introducing us is a challenge in terms of collecting the data from all these different cloud applications and on prem applications to one place where we can do some quality reporting and do some quality data analytics. And this, as we know, is quite different because the cloud and on prem are quite different animals. In cloud, you don’t have database access, but in on prem, you have full access to the database. In cloud, you can only do reporting for the data that you want to pull out. So that introduces a new level of challenges when we try to combine the data, marry the data between the two environments in a meaningful way. So to solve the problem, we are still talking to multiple vendors, trying to figure out what solution would be a better fit for our situation. And that’s where we are.

Riaz Yusuff [00:11:28]:
To answer Sanjog’s question on what would be my pet animal that’s comparable to cloud, thanks to a quick online search, looks like red kangaroo is the fastest growing animal and the second fastest land animal. In cloud, I don’t think any of us can avoid going to cloud, so it’s a faster move to the cloud. And at the same time, the price is also fast growing, just like red kangaroo, the way it grows fast in a fast pace. So that’s what I would pick as my pet animal.

Sanjog Aul [00:12:03]:
Yeah. We will get a lot of learning, and especially the learning about the different flavors of animals you’re comparing it to cloud. I mean, that’s amazing. So, Anshuman, you’ve been around the block, and of course, you have dealt with different organizations, scenarios, etcetera. What do you see? If you had to pick a top challenge or a set of top challenges you’ve seen, which are kind of a theme that you see, what would those be? And of course, you have to tell about the animal, too.

Anshuman Mandal [00:12:31]:
I’m Anshuman. I’m the vice president and practice head at Zones. My favorite animal, I would pick my labrador I have at home. And there are reasons to it, I’ll come back to that. From a challenge standpoint, what I have seen and observed as a service provider, as well as interacting with various customers across different verticals, so to speak, predominantly it is cost. So when cloud is full scale, cost is number one. How do you contain it and how do you manage it judiciously? The second thing is, after cloud adoption, how do you re architect or adapt your applications to the transformed cloud technology if you have not done it before? That comes from the scalability standpoint. Then finally it is…

Anshuman Mandal [00:13:24]:
Obviously it ties back to FinOps again. But how do you streamline processes and provisioning and decentralizing if you have a multi cloud environment? How do you manage it? Which Sushmitha said, she has a multi cloud situation. And how do you orchestrate similar things, applications and workloads in different cloud zones? So how do you bring all the three, all the couple of clouds, if it is AWS, Microsoft, GCP, whatever it is, how do you bring them together? So orchestration is very important.

Sanjog Aul [00:13:56]:
Brian.

Brian Gardner [00:13:58]:
Hey, thank you, everybody. Thank you for the invite. Doctor Brian Gardner. I’m the CTO for the City of Dallas. So our greatest challenge, I’m going to back up a little bit. It goes to some history around 2017, 2018, when the city first brought on cloud. So the challenge was, hey, we’re going to move to the cloud. We’re going to be cloud first.

Brian Gardner [00:14:22]:
So we moved some, migrated a bunch of stuff to the cloud and found out that, hey, there’s a pretty substantial cost to the cloud. Oh, my gosh, we haven’t budgeted, we haven’t planned. It’s not in anywhere that we can cover this cost. We had to rob Peter to pay Paul, and then we made the challenge to move it all back. So, in that we lost a significant amount of data that was fairly critical because somebody pushed a button that they probably shouldn’t have pushed, and we lost all this data that was very impactful to the city. So my greatest challenge is, what does the city need to do? We know we’re going to have to be a hybrid. We know we’re going to have some data on premise that we’re going to have to maintain here, and that’s just because of some of the regulations and some of the sensitivity around the data and availability. But what do we move to the cloud? So we’re really looking at what that should look like.

Brian Gardner [00:15:18]:
And initially it was a single cloud. I’m really looking at it from a multi cloud perspective. So a hybrid multi cloud is where I want to head, but how do I get there? How do I know what to move where and what is going to be the most cost effective, sustainable, and from a governance perspective and security perspective, to be which cloud where. So that’s probably the city’s greatest challenge. And for me, I would say this animal is a lion that needs to be tamed so we can get it to, we can have, I don’t want to say circus, but that circus act running smoothly.

Sanjog Aul [00:15:56]:
I was waiting for somebody to call it a beast, but I found, like, people calling it, okay, it’s tamable, which is a good thing, right? At least we have the right feeling that, yes, we can tame it. Beautiful, Arjun.

Arjun Puri [00:16:08]:
So this is Arjun Puri. I’m with DTCC, responsible for cloud data engineering and analytics. What that really means is all our data platforms, BI tools, machine learning platforms, et cetera. So we started our cloud journey pretty early, 2018, like exactly the words that Brian used. Our CIO at that time adopted a cloud first. So if there’s any new workloads, anything new, let’s look at the cloud first.

Arjun Puri [00:16:36]:
We successfully migrated and it’s been extremely successful. And we are running some critical, critical workloads from a regulatory compliance or SEC purposes, et cetera. What we did find, though, is that, going back to the point that Anshuman made, and I’m not sure how prevalent that is, it was a little bit of a lift and shift. But once you’re on the cloud, really to leverage the cloud and be successful in the cloud, everything needs to be modernized, automated, from your account vending to any process. So the whole stack has to be thought through and modernized. And so we’re going back and revisiting some of that. Security is a big challenge for us. And that, again, ties back to the automation component because you need to reduce your blast radius because things could eventually happen. We do hear about all kinds of nightmare ransomware stories.

Arjun Puri [00:17:39]:
You need to reduce your blast radius and manage that very effectively. And then I mentioned machine learning. So that opens up to all these packages that our users want to use, but how do we want to govern and manage and secure those and make sure that they’re actually free from any security holes and loopholes? So that’s some of the key challenges that we’re working through. In terms of the animal, it’s probably more of a mammal. I’m not sure if it’s an animal, but I think of the cloud as a dolphin. It’s a friendly, smart mammal, animal, fish, that really we can be our friend. We just need to make sure we train it, and it learns and goes to, as somebody mentioned, the word circus with us and performs very nice.

Sanjog Aul [00:18:35]:
Tony.

Tony Giampaolo [00:18:36]:
Good afternoon and thanks for having me as well. In my role, on the network side of things, it’s more migration sequencing is where the challenges lie. And this is around having the appropriate application baselines to understand what kind of throughput is needed to migrate, the sequencing in terms of what goes first, and then the BU contention. In a very highly matrixed environment, you naturally would have variation of requirements, expectations, and really just getting that alignment and sequencing to not disrupt the business and be able to adequately migrate applications is a little bit of a challenge in our environment. As far as pet animal, I’m going to stay with the fickle felines as well, and support folks on the phone, but I’m going to go to the common house cat, where the common house cat will have high expectations. It purrs when it’s fed and you take care of it well.

Tony Giampaolo [00:19:38]:
But it’s sometimes hard to grab. It’s not like a dog that’s going to necessarily come to you. So it’s fickle in that nature and it can hiss at you if you’re not careful.

Sanjog Aul [00:19:46]:
We should have a separate discussion on all these flavors of animals that you just spoke about when it comes to cloud. It’s so beautifully said by all of you. Beautiful. Thank you so much. Now, coming back to you, Jonathan. So you gave your specific challenge, and what we want to do now is very clearly help us understand what you specifically do. Let’s get to very specific areas or the interventions you tried and what worked and did not work. And what’s that snag you’re hitting? And as soon as you share that, would love the rest of the group to jump in to see if we can help Jonathan walk away more enlightened and get something actionable that he can use.

Jonathan Bagnall [00:20:22]:
Well, I’d like to first say I validate everyone else’s challenges. It’s how deep do you want to go? So in my profession, I am the head of product security. So that’s both medical devices and non medical devices, but all of which is regulated because in healthcare we’re dealing with a lot of PHI. I would say that when we talk about challenges and what we need to overcome, I heard a lot of challenges. So the matrix of large enterprises having to either sell or develop a standard to multiple business units that have different priorities, and their priorities typically are to get product to market and start to generate revenue or start to please a customer. I stated that one of the biggest challenges is regulations, and the regulatory and laws continue to evolve rapidly. And what I’m seeing changing is that the laws and the regulations, as they come into effect, they are not giving any grandfather options, nor are they giving you a lead time to comply.

Jonathan Bagnall [00:21:48]:
So that’s very challenging. The cloud can help in that area, or it can be detrimental in that area. Detrimental, meaning that you have aging technology you have to deal with, or you have to prioritize it internally to be able to move forward. One of the things that rang a bell listening to some of my peers here, some are looking to migrate to the cloud, some are in a mid level shift. It rang a bell in my mind that back in 2005, when we started to understand what the cloud was and trying to understand how we could utilize it. When we think about a highly matrix environment, we had individuals going out to public clouds and initiating cloud instances, and starting to develop on those instances. I remember back then, it was really out of control because we didn’t know who was doing what or where they were doing it.

Jonathan Bagnall [00:22:53]:
And the only way eventually that we could stop it is to turn off the ability to link to those domains and as well as to develop policies. So in my thought process, the usage of the cloud developed almost by need, through employees and departments going out and creating their own cloud environments, public cloud environments, to do what they needed.

Sanjog Aul [00:23:20]:
So, question is, if you had to get an answer which will help you do something new, more or different, going back to your desk tomorrow morning, what answer would you look for? Because people are standing by and they’re all very accomplished with a lot of feathers on their cap and bruises, both.

Jonathan Bagnall [00:23:36]:
Well, the answer that I would look for is how do we, globally or regionally, I guess, come up with standards within cloud environments that will help us operate and grow more effectively, efficiently, and also contain the cost? And that would be, I think initially would be in relation to regulations. So in the United States, we have state level regulations, we have federal regulations. They are a long list of them, and it’s very complex to be able to navigate in those environments and have any level of assurances that you are in compliance at all times.

Sanjog Aul [00:24:22]:
So are you looking for a specific playbook of sort or a blueprint or one standard like a stencil, which you could utilize and make most of? I’m just trying to get to that.

Jonathan Bagnall [00:24:33]:
Yeah, I think I would move to cloud environments that have certain requirements, regulatory and legal requirements. How do we develop that standard approach and how do we do that holistically? So you go to cloud environments. Well, a lot of them have a menu and if you click off a menu that I need this level of compliance, your price sometimes doubles and triples. There are those that would be, I guess, motivated not to click off that menu.

Sanjog Aul [00:25:09]:
Sure. So, Sushmitha, I’m sure you are kind of living the same because you are from a regulated industry. But please feel free to go ahead and share what you are suggesting as a way forward for Jonathan to be able to use it.

Sushmitha Nalamothu [00:25:24]:
So what we are doing is, or I did in my previous roles, was we have a center of excellence. And first, as stated, it was more of anybody wants to do anything, and then there was a lot of redundancy and cost that was part of all these conversations with different cloud providers. So our organization created a COE which is called cloud center of excellence. And then we are determining certain standards for every single application that wants to go into cloud. We are assessing their architecture, we are assessing their service based architecture, re architecting as needed and not letting lift and shift as much. That is one thing. My role is expanded to compliance across my BU right now. So I am super excited to add some of the things I started to do. I’m looking at entire suite of, and it’s very hard, I get it, the compliance regulatory risk, entire IT control standards is a huge, huge area and a black box for most of the organizations or applications, I would say not organizations, for application owners.

Sushmitha Nalamothu [00:26:41]:
So we are creating a centralized set of compliance regulatory risk security checklist for on prem, SaaS, vended and cloud. The easiest is cloud. It’s just because it’s much more matured at this point rather than on prem legacy and vended and SaaS products. I’ll be very honest. So we are trying to automate and prioritize that automation policies as part of code. And again, it’s easy for cloud and not as easy for on prem, vended and SaaS products. So I’ll take a pass. I think I went through a lot.

Sanjog Aul [00:27:20]:
No, absolutely. Satish.

Satish Kumar Chala [00:27:22]:
So yeah, I agree with what Sushmitha has stated. And just to add further on the compliance side, which I found very useful, is having a robust regulatory development program, regulatory watch program. I would say this is more of a business side initiative, involving the legal team and of course the cyber security team. But then primarily you are trying to have a channel through all the regulators based off the applicability of your business lines and the products you have, and get the latest and greatest developments from all the regulators. Maintain that repository and see how it is applicable for your underlying system. So in a way, it is a mapping exercise where you will get to the checklist which Sushmitha was referring to. That’s one. The second is also to, again, as the landscape is quite challenging, you may want to optimize your continuous risk assessments in terms of the IT assets exposure and see how you could be on top of it.

Satish Kumar Chala [00:28:33]:
So that’s one additional thing which I found very useful, where the optimization is taken care in terms of relevancy of digital assets that needs to be assessed. Primarily, you could have multiple feeding factors to it. You could go as cutting edge as using a machine learning model out there on top of it. But then if you are doing it manually, I would say a simple example could be based off your changes which happened, or your major releases which happened for a given set of applications, or it could be exposure in terms of APIs, so on and so forth. But here I pause with those two major examples.

Sanjog Aul [00:29:15]:
So, Jonathan, do you want to just…

Jonathan Bagnall [00:29:19]:
I did, and I probably should have given an example. So the EU NIS2 and Cyber Resiliency Act is very aligned to cloud environments and the protection of data and as well as AI. And it’s very complex in a large enterprise to motivate and understand where it is that you may comply and where it is that you need to, I guess, enhance or understand what that environment is, whether it complies or not. Now we’re talking about environments, but the Cyber Resiliency Act goes beyond that. I agree with everything said. It’s just challenging sometimes. It always seems to fall back on security when it’s asked, do we comply? And a lot of these laws and compliance requirements go beyond security in many contexts.

Jonathan Bagnall [00:30:20]:
And if I took that one again as an example, it’s still evolving. And we’re getting feedback that countries within the EU are going to have specific requirements to NIS2 and the CRA. So it becomes, it’s just very complex.

Sanjog Aul [00:30:39]:
Sure thing. Sure thing. Riaz, you have a thought?

Riaz Yusuff [00:30:43]:
Yeah, I do. In the higher education space, we use a tool called HECVAT. It stands for higher education community vendor assessment tool. And this tool helps us to assess the risks of the vendor by asking, it’s a questionnaire, basically asking a number of questions. We use that. And in my organization, we also leverage our external legal team in order to involve them to make sure that we are not missing anything out as needed. And we also have a subscription to an IT research firm that provides analysts. So we, time to time, we ask them to review our contracts before we sign our contracts to cover our bases.

Riaz Yusuff [00:31:26]:
So those are the things we do. It’s not 100% bulletproof, but it at least covers some of the key points to keep us in the safe zone.

Jonathan Bagnall [00:31:36]:
Yeah, I would agree that, I’m sorry to interrupt, but my challenges are wide ranging and it’s very evident with the alignment of the input here. So I thank you very much.

Sanjog Aul [00:31:47]:
Okay, I’ll take one last comment, Arjun.

Arjun Puri [00:31:50]:
Yeah, sure. So somebody mentioned machine learning. In the financial services, I’m sure same case as for us, and this predates machine learning because we have quantitative models. Anytime we talk about AI and models, it comes with the baggage of model governance. So how do you know that’s accurate, your models are making the right assumptions and all that? And that’s a very critical and somewhat something that we’re grappling with before anything goes into production that’s AI related, and that needs to go through this thorough model governance process. And the regulators will also be asking us about that.

Sanjog Aul [00:32:27]:
Great. Okay, over to you, Sushmitha, about the challenge that you did share. But then now let’s talk about what you did to crack that problem. And where do you need answers so the team can come and try help.

Sushmitha Nalamothu [00:32:38]:
So cloud governance is a vast area. I think I have challenges across security, compliance, cost, but I want to touch on cost a lot more. So when we promised our business that we are moving to cloud, the first, or maybe one of the important points that we stated, was that we are going to optimize the cost. And that did not happen for the first time. And now we are getting better at looking at dashboards and reflecting on where we are using our resources on a cloud. We have our own checks and balances, and then what is not utilized, we are trying to have a gate of disappearing it and stuff like that.

Sushmitha Nalamothu [00:33:26]:
I want to hear from others what are the strategies that they are implementing in order to control the cloud cost.

Sanjog Aul [00:33:33]:
Okay, so that must be a very interesting subject. So yeah, team, who’s going to take it first? All right. Satish.

Satish Kumar Chala [00:33:38]:
Somebody else was volunteering. Fine.

Anshuman Mandal [00:33:38]:
No, go ahead. Go ahead, Satish, I’ll come after you.

Satish Kumar Chala [00:33:38]:
Well, yeah, so in terms of cost, I think we were late adopters and of course we burned the fingers or learned the lessons the hard way to begin with. And it’s been the story for quite a few industry players in general, where cloud first was the main strategy, and later, when cost really skyrocketed, that’s when it hit a hard realization. Now, in terms of containing the cost, at least I have seen that what has been really beneficial is to understand the business priorities in terms of what is our forward strategy, in terms of the organization, from the plan, in terms of how we are trying to grow, in terms of our revenue stream models and all of that, and aligning that in terms of scalability. Second, in terms of identifying the vendors, the partner vendors, and again at the same time mitigating the risk of not having a single vendor so that you are not pigeonholed in terms of the pricing game and also in terms of regulatory pressures, so backlashes, in terms of disaster recoveries and all those standpoints. So long story short, in terms of cost benefit analysis, vendor consolidation, and having a forward looking map in terms of regional product mix, that has been very beneficial. When I say regional product mix based off your organization structure, let’s say you have a high turnover, 24 by seven kind of availability kind of requirement in AMER region, and in APAC as a business model, it is not of that importance. Probably you need not really have cloud first kind of strategy in there, or you need to really define your product mix in terms of your cloud adoption in each of these jurisdictions.

Satish Kumar Chala [00:35:55]:
That really helped us save a lot of cost. That’s one. Second is in terms of identifying the real business sensitivity around adoption of these use cases. Earlier it was everybody who were trying to get any newer workloads, they were onto cloud. But now we have a very useful and in fact a vetted questionnaire which helps us enable to understand the use case and also see the two to three year roadmap of what that looks like and prepare based off that particular assessment, rather than you just onboard based off the initial use case. And if there’s no sustainable aspect of it, we don’t really entertain those. And the third important thing which we have noticed is in terms of the scalability aspect of the use case, what is the future use case going to be? The what if case analysis of that use case. When I say that, is this going to be a GenAI adoptable component? I’m just trying to throw in the cutting edge.

Satish Kumar Chala [00:37:00]:
But this was not an equation which we used to have five years ago. But the way at which we want to really exploit technology has also to be brought into the equation. That helps us save that architectural move and design moves and reduce the shifts for the end users and again for the infrastructure teams. So these were the three main things which helped us reduce the cost in long term.

Sanjog Aul [00:37:26]:
Thank you. Anshuman, you had a thought.

Anshuman Mandal [00:37:29]:
Yes. And it actually ties back to Jonathan’s comment as well. And what Sushmitha mentioned a minute ago. There’s a part of the previous question. She talked about COE.

Anshuman Mandal [00:37:40]:
COE is a great thing to do to maintain the regulations, governance on cloud. And what I’ve seen, as a matter of fact, with one of my customers in the recent past, they are about 90, 95% on the cloud. Only a few legacy systems are there in their small data centers that they have. But they were bleeding through their noses. So we looked at it. We went as an advisor, we looked at their cloud consumption and cloud spend and their commits with all the hyperscalers.

Anshuman Mandal [00:38:11]:
So your discounts are dictated by the commits that you make with hyperscalers. And today, at some point right now, if you can have the right forecasting for your first one to three year spend, as whether you want to go with a reserve instance or anything, that is one basic control where I’ve seen a lot of pitfalls right now. The other thing is from a governance standpoint, Sushmitha, you mentioned. Yes, obviously the financial compliance, there is security, there is service compliance. You can have a technology governance as well as what is coming down the pipe and all that. But on top of that, typically from the other side of the fence, from a services side of it, if your consumption in cloud, I’m just saying on the table, if it is a million dollar, the benchmark for your operational and transformation cost should not be more than 16% to 18% of your consumption model.

Anshuman Mandal [00:39:09]:
So that is a guideline that I have seen and I strictly adhere to. In addition to that, if you suspect that you’re overspending in your cloud consumption, there are partner programs driven by SIs. You might want to explore because it’s a gain share model. For instance, again, if you’re spending a million dollar a month on Azure, there will be commitments. They will do a quick assessment. They will look at your workloads, what are idle, what kind of usage you have, what scalability you’re using and all that. And they will give you a gain share model in terms of committing to 30, 40% savings, and that funding can be pumped back into other initiatives and the partner cost itself. So those are few things that I’ve seen and has gone very well actually.

Sanjog Aul [00:40:01]:
Great. Awesome, Tony, thank you.

Tony Giampaolo [00:40:05]:
I really like what you were saying about the relationship with the CSP. I think that a lot of folks don’t really tap into that. And the other part that I’m hearing again is just consistency in terms of cloud first. In organizations that I’ve been in, even in consultancy roles, I always say this is a, it’s not a leap then look. And a lot of folks, they get in a cloud, they leap and then they look, when the bills start coming, opposed to look, then leap. And some of the ways that I’ve developed a better working relationship around cost management, if you look at like some of the base cloud governance disciplines, is around cost management and resource consistency. I feel passionate about this subject because I really enjoy it. I think from a resource consistency standpoint, that’s one way to really tighten the reins on what your cloud spend is going to be, because the nirvana state is to get a level of OpEx predictability, and anybody with a credit card could go ahead and say, hey, we’re going to go ahead and set these cloud services and then they expense it, whatever, like that. That’s just not good behavior overall when you start thinking about the desired business outcome.

Tony Giampaolo [00:41:26]:
So, and then to go back to the CSP relationship, that’s where your TAM is key. And that’s where, if you could explore reserved instances, or you know that you just have x amount of flavors within your resources that you’re going to take from a compute standpoint, or even your storage classes based on where they’re aligned in terms of hot, cold in Azure, or is it going to be long term storage like Glacier or standard in AWS. It really just goes back to what problems are you trying to solve for, put some parameters on the technology use case and then managing it through. And then I’ll say that one of the most invaluable things that had helped me is my relationship with finance. Even when there was an absence of a cloud business office in terms of maturity, just aligning to that level of FinOps practice when you think about crawl, walk, run, and then really leveraging your financial partners. Because depending on how complex your cost allocation model is, when the BUs have to, if it’s in a general catch all bucket, it’s a little hard. But you can encourage, and I’ll stop short of saying force behaviors, but you can encourage behaviors with the right cost allocation model where, if you’re not following the resource consistency baselines, then it’s a difficult conversation up their leadership chain on why they didn’t follow relative standards. So cost allocation, relationship with finance, first couple of disciplines of cloud governance around consistency and cost management, and then really leveraging your TAM relationship.

Sanjog Aul [00:43:11]:
All right, guys, I got a number of different comments. This seems to be a favorite subject. A lot of people are giving input. We have to make sure that each challenge gets addressed. So let’s be very quick. I’ll start with you, Jonathan.

Jonathan Bagnall [00:43:24]:
Just very quickly, just to add to some of what was said. We are in the process of consolidating standardization. And one of the things that we’ve embraced is architecture review and architecture review boards in standardizing on an architecture. So when we have new development, new products coming into our cloud environments, they have to meet a requirement at best they can, and they have to identify why they can. That addresses cost containment too. But cost containment, the relationship with your CSPs and also governance of that relationship. We’ve seen a lot of cost containment in which we just didn’t have our eyes on the glass. You have 100 licenses. Well, we only use two.

Sanjog Aul [00:44:14]:
Usage monitoring is what you’re referring to.

Jonathan Bagnall [00:44:16]:
Yes, absolutely.

Sushmitha Nalamothu [00:44:18]:
Sushmitha, one thing I want to add is we touched about CSPs and all that. One of the things we did when we were going cloud first in different organizations, I am a multiplier. Instead of bringing in consultants and charging about $250, $300, $350 per hour, we started training our own resources from a cost optimization perspective. So we had a pilot, like first a hundred, for whatever that three month, six month or nine month plan was. And then that actually, when you aggregate it, the organization was able to train their own resources and put them from on prem to cloud engineers. So that actually saved a lot of cost. I just wanted to add this.

Sanjog Aul [00:45:05]:
No, no, absolutely. This is a great point that you made. Money saved is money earned. That’s also a good way to look at this. Yes, absolutely. Quickly, Arjun, thoughts?

Arjun Puri [00:45:13]:
Yeah, so I think we’ve talked a lot about tools, technologies, cloud, etcetera. The one thing that I didn’t hear, want to call out, which we’re having to do, is taking care of our human capital, our people. So for example, everything in the cloud we heard automation. That means you need to train in everybody, whether it’s Terraform or other ways of doing things. Same way DevSecOps have to be modernized and refreshed, in case you use GitLab then everybody. Likewise, if you’re using a security vault, standing that up, making sure everybody’s trained how to use it, providing them the APIs for the apps to easily use them. So basically, net net, taking care of the people and modernizing and refreshing and giving them the right tools also should be at the forefront of any move to the cloud.

Sanjog Aul [00:46:08]:
Great. We’ll go very crisp going forward. So, Satish, your top issue where you would like to get answers, and of course preface it with what you tried already so you don’t get something which you’ve already tried and it hasn’t worked for you.

Satish Kumar Chala [00:46:22]:
No, sure, definitely. And I have a hard stop in like twelve more minutes. So primarily, in terms of the expectation, I’m talking about sustainability as my topic. So sustainability from an ESG alignment perspective. So currently, all of the programs which we have, we measure in terms of the KPIs and the KRIs perspective. And when we throw in ESG as one of our risk taxonomy, we’ll have to report our programs effectiveness and adoption in terms of cloud, in terms of measuring aligned to that ESG component. Now, having said that, what we have tried is to build a framework. We are not there yet and it is not yet official in terms of risk taxonomy, and trying to actually absorb the shock with a seamless transition when we have officialized that ESG is going to be a risk taxonomy to begin with. So the preparation work currently is I’m trying to first reach out to all the industry peers to see how they are aligned and what are they trying to ensure in terms of sustainability, because they may not really have ESG to begin with.

Satish Kumar Chala [00:47:38]:
So I’m just trying to see from a sustainability standpoint what are the key benchmarking frameworks which are adopted. And that’s where I have seen some metrics in terms of cost optimization, which we discussed, of course, earlier. The second is in terms of the carbon footprint and measure, which is again a concept which is being computed, but we again have not yet determined which is going to be a standard to begin with, which we can adopt and which will not help us create newer problems. And that’s another metric which we were munching upon. But these are the two things which I have currently to share and I would love to hear if there are other ideas in this space which are tried or are at least being thought of.

Sanjog Aul [00:48:33]:
Anyone who’s looking at making it a better planet while making a ton of money for the business?

Anshuman Mandal [00:48:40]:
I would like to weigh in quickly on this, Satish.

Anshuman Mandal [00:48:42]:
And it is not from your industry, it’s mostly from the retail industry. But sustainability is big there. What I have seen as well, right now it is not specific to cloud, but to a total business. For instance.

Anshuman Mandal [00:48:56]:
From a carbon footprint standpoint, if you have any legacy infrastructure from a data center standpoint, that is your first place to start with. How you can bring in a circular economy, in terms of if you have assets that are being retired, how you can work with your OEMs to make them go through an ITAD process. So that is big.

Anshuman Mandal [00:49:23]:
And it’s not only data center assets, but end user assets as well. And we fail to recognize, like when we talk about 300,000 employees or even 20,000 employees, there is a lot of carbon footprint that can be reduced through that. Now, in the retail, QSR, travel and hospitality industry, I’ve seen they have leveraged a lot of IoT solutions for their remote monitoring to reduce the truck roll.

Anshuman Mandal [00:49:47]:
For instance, if there is a dispatch that is going, how can we reduce that? How you can reduce that dispatch? If there is a quarterly health check that is needed for a remote office, if you deploy any IoT solutions, how can you reduce those cycles so that you can reduce the truck rolls to those locations? And finally, the optimization of infrastructure as well as your cloud resources plays back. Now, when it comes to your carbon footprint measurements, I don’t know if you have any reporting back to any agencies or anything. There are frameworks where you can measure your carbon footprint and set your organization internal goals if it’s not being reported. So these are the few things I have seen. Just wanted to share with you.

Satish Kumar Chala [00:50:34]:
Thanks. Thanks. Just to give you a feedback on that carbon footprint. Yes, we do have access to those computation which we are currently doing, including using your emails in a year, how much you have saved and all.

Sanjog Aul [00:50:53]:
So, Riaz, specifically what you’ve tried to do to fix problem and ask very precise question if you could, to get a precise answer from the team.

Riaz Yusuff [00:51:02]:
Sure. Where do you want help in order to fix our data warehouse problem? Because we have multi cloud and on prem, we have been talking to multiple vendors. So we had an RFP published. The vendors who responded to that is the ones we are talking to at this time. We haven’t still identified that one perfect solution yet. So I’m curious, whoever has a solution, I’m assuming some of you have this situation of having multi cloud and possibly on prem still available in your environments. How are you doing your, or fulfilling your, reporting and data analytics needs?

Sanjog Aul [00:51:50]:
Okay, anyone? Any takers?

Arjun Puri [00:51:54]:
Yeah, so that’s in my wheelhouse. We of course run large warehouse and data lakes, etcetera. Of course, we use a particular, we use Snowflake. Been very successful with that. That meets all the security guidelines as well as DR, etcetera. But it is a pure cloud solution. So there is no on prem.

Arjun Puri [00:52:21]:
So those on prem workloads would push data out to your warehouse and then you can do the reporting on it. But that’s been very successful for us.

Riaz Yusuff [00:52:31]:
So do you push your on prem data to that cloud software and get all the data in one place?

Arjun Puri [00:52:37]:
That’s correct. Yeah. A lot of our source systems are still, critical systems are still on prem and we push data over. I’m sure there are some network experts here, but it goes over Direct Connect and PrivateLink to the cloud. Our traffic is off the public internet.

Sanjog Aul [00:52:54]:
All right, Jonathan, quickly.

Jonathan Bagnall [00:52:56]:
I would just add that the question is very complex within the healthcare industry and within any industry that is building and will rely on AI technologies going forward, which is many of us. That complexity is the integrity of the data. So it’s not just about data warehousing your own data, but understanding how your third parties or how the interoperable and how your data is interoperating with other data sources. The integrity of that data is going to be very prominent within regulatory and regulations and laws going forward. So where did the data come from? Who touched it? How old is it and what regions? I mean, it’s huge. So I think it’s presenting itself as a larger complexity and a bigger problem as AI grows.

Sanjog Aul [00:53:59]:
All right, awesome. So, Anshuman, you come from, again, having a wider lens so as to dealing with so many people. Any specific question that you may have, based on what the leaders here on this forum have shared, or a specific tough but fair question you would like to get answers for so you can sleep well at night?

Anshuman Mandal [00:54:20]:
Most of the questions I had, it was well thought out and we beat them to death. Specifically on FinOps, from a Zones standpoint, our go to market strategy, obviously looking at FinOps, that is one big thing. And then the sustainability that we talked about. So these are the two very niche advisory offerings that we are going after right now. My question to all the leaders here, as you guys are navigating through all these hurdles, you talk about compliance regulations, transformation, how do you save cost? It’s not a question, it’s more of a statement, I would put it that way.

Anshuman Mandal [00:55:04]:
The cost to your cloud, cost to your consumption is absolutely, the bulk part of it is how you consume those in terms of cloud workloads, resources from the hyperscalers and others, and you get a bill for that at the end of the month. But there is a lot of opportunity that we can derive as how you manage those workloads. I’m sure a lot of you guys are doing it in house, something could be outsourced and all that. So keeping a close eye, which I’ve seen, if you guys are doing that, that sanity as well, like operations and technology both has to go hand in hand in order to have agile cloud environment. Actually that’s my statement as well as questions to all of you.

Sanjog Aul [00:55:53]:
So any thoughts that you would wish to get from the leaders here or is this something that you are sharing your observation?

Anshuman Mandal [00:55:59]:
It’s more of an observation. My only, because everyone brings in different perspective from the organization, my only intriguing question here is from a spend standpoint, manage your cloud spend. How you guys are benchmarking yourself specifically on the operations part? Is it by resources, is it by consumption, or is it by internal resource, which is a fixed price anyways, be it internal or external? So what is the majority cost model that you internally account for, for operations of cloud?

Sanjog Aul [00:56:36]:
All right, Sushmitha, you are muted.

Sushmitha Nalamothu [00:56:44]:
So we have created a cloud operations team. So the one who’s going to be center of excellence, and then there is a cloud implementation team and then cloud operations team. I work for different organizations so it’s a mixed bag. But in one of the clients that I worked with was, back to your statement, Anshuman, about when we’re talking about FinOps, like how we are gonna estimate our cloud spend, like when we start versus when we are adopters and then when we are matured. The operations cost need to go down or up depending upon our spend. So we have a ballpark of where we want to be from an operation spend standpoint, and we are doing initiative of cloud implementation, run and change, as well as ongoing operational cost for cloud. So that is the strategy we are doing right now.

Arjun Puri [00:57:50]:
Yes. So in our space, and of course my specific, similar to what Sushmitha mentioned, in the cloud area they have a chargeback model. But in the analytical space we were able to develop a very detailed chargeback model, which is rare at DTCC. So both labor and software was very easy to charge back to the consumers. And so we developed that. That was developed about maybe two years back and it took some work from IT finance and others helping us with that. So we implemented that specifically in my space.

Anshuman Mandal [00:58:30]:
Just one thing, Arjun, when you said you are doing the chargebacks, so you’re charging it back to your internal cost centers, correct? Based on the usage and all.

Arjun Puri [00:58:38]:
Yes, yes, exactly.

Sanjog Aul [00:58:39]:
Yeah. All right. So did you raise your hand again? Did you have something else to add?

Sushmitha Nalamothu [00:58:43]:
No, I’m good. Sorry.

Sanjog Aul [00:58:44]:
All right. Awesome. Okay, so Brian, coming to you, your top challenge where you feel the leaders here in the room could help you, specific thing where you could get some answers.

Brian Gardner [00:58:55]:
So thank you for the question. What I would say is I feel like I’m the last guy running this race after this discussion here, because I’m listening to everything that they’re talking about and these are questions I’m going to have. My top question and what we decided was, how do we decide what that multi cloud looks like? So we’ve taken an approach where we are engaging different cloud providers and doing assessments between them and coming up with, this is how we’re going to put our cloud strategy together. So I guess my question would be, what were the lessons you’ve all learned as you’ve moved in this multi cloud? And I have to drop as well here very shortly. So thank you all again.

Sanjog Aul [00:59:38]:
So, Sushmitha, go ahead.

Sushmitha Nalamothu [00:59:40]:
So at one of the organizations I worked was they brought in all the cloud providers just so that to ensure that we are not doing big bang dependency on one versus the other. But inside the organization we had multiple different views and, depending upon the application structure and the adaptability, we decided which cloud provider we wanted to go. We predominantly picked one because of cost between two others. And I don’t want to take names, but it comes with a challenge of maturity. But then right now, pretty much all three or four cloud providers have pretty much the same. But the cost efficiency is one of the things that we picked.

Sanjog Aul [01:00:28]:
All right, awesome, Jonathan.

Jonathan Bagnall [01:00:31]:
Just some quick areas that we look at to identify and to manage cost. We look at areas of right sizing resources as it relates to the business teams. We also work with the CSPs and we try to identify where we can reserve instances and develop saving plans with them. I’d also say that we have integrated usage monitoring and alerts. We also look at where we can optimize storage, where there’s a huge cost there. And I could go other ones. But another priority area is data transfer costs. So monitoring them and understanding what your requirements are around them, you will be able to identify what your true cost is and areas which you can continue to improve them.

Jonathan Bagnall [01:01:26]:
I liked what Sushmitha said about adoption and optimization. I will say that I have experience where in the adoption mode, that there isn’t a high enough investment to get to the optimization mode. You always seem to be in this adoption and nobody knows anything. So I would recommend that if you are in the adoption category, that you allocate enough spend so that you can get past adoption and move to optimization. And you can actually project downward cost trends based on industry analysis that’s out there.

Sanjog Aul [01:02:19]:
Thank you. Anshuman.

Anshuman Mandal [01:02:20]:
Yeah. One last thing I would like to add. Jonathan, well covered. We talked about all the nuances. Only one thing to Brian is, Brian, do not forget to pit providers against each other, even though some workloads are best suited for a particular hyperscaler. Because as you navigate this, you will try to take advantage of a lot of partner funding, as well as not only from an assessment standpoint, but from a migration standpoint as well. That will give you a good foundation as you go through the assessment, as you define which workload goes where, what you have to retain if it is hybrid cloud environment. But definitely take advantage of the partner funding and all the other benefits that hyperscalers come with. I’m sure whoever you’re working with, they will be talking to you about that. Just a quick thought.

Sanjog Aul [01:03:12]:
All right, so I see three people wanting to speak on this subject. Seems to be an interesting one. Riaz, we’ll start with you.

Riaz Yusuff [01:03:19]:
Sure. I’ll be brief. Over the years, I have come to the realization that the cloud solution is not for cost optimization. It’s for other reasons, primarily. So that’s a key thing to keep in mind, and when you’re talking to your business, that’s how you’re better off selling that way, rather than selling the cost first. The other thing is, I think it’s important for all of us to have principles in house. Principles, what I mean by that, we have a technology strategic plan, whatever you call it, you should have some guidelines in regards to how you bring in new solutions. One of our technology strategies is to consolidate solutions.

Riaz Yusuff [01:04:00]:
Instead of having multiple solutions, try to consolidate as much as possible. So when a new solution is proposed, we point our business, and whoever is requesting that, to that technology strategy, and try to convince them to leverage the other solution that we already have in house. Although it may not be the cheapest option, because it has other benefits.

Sanjog Aul [01:04:22]:
Let’s look at Sushmitha. You had a thought.

Sushmitha Nalamothu [01:04:25]:
So to Riaz’s point on the cloud is not for cost optimization. One of the things we sell to businesses, it’s not sell, but unfortunately, that’s how it works. Being in a technology role, I need to get funding. So one of the things is like spinning of the environments as fast as possible. That is one of the greatest and biggest advantage I see as a technology person, wherein if business wants some product to be done, if they don’t have an environment, and we could spin that very fast, that is the topmost of my selling point. And they get sold very well. That’s one thing.

Sushmitha Nalamothu [01:05:21]:
But going back to Anshuman on pitting them against each other, one of the things I did in one of the organizations was the adoption. We were actually trying to train our resources on cloud product, like one versus the other. And we were spending a lot on training, and I can’t imagine how many millions that we were planning to spend from each application or BU or portfolio level. I challenged the provider to train my resources, and they were not used to train more than 20 or 30, but the challenge was so much that they trained 600 people in like a week. And I’m not saying that that’s a great way, but they did persona train. So challenging in some of the ways which are not norm also helps us to aggregate the cost optimization. Just wanted to call it out.

Jonathan Bagnall [01:06:00]:
So, Jonathan, just a quick one. Where you work and govern a municipality, I would look at similar municipalities and see what they’re doing and see if you can align and partner on contracts, if you have like solutions, so on and so forth, for some price gain or for some pricing. That’s what I’d be looking at.

Sanjog Aul [01:06:31]:
So, Anshuman, since you’ve been listening to them and you wear that wider lens of view, well, I just keep coming back to that. You’ve heard all of them. What would you say are the top takeaways which leaders today can benefit from so that this decoding the complexity and forging ahead with the hybrid and multi cloud, the topic that we started with, what does this all distill down to, in your view?

Anshuman Mandal [01:06:59]:
No, absolutely. And I’m very actually intrigued. The real challenges that the leaders said, it’s not going back to the drawing board, but in a quick summary, obviously, everybody is invested in multi cloud strategy. It’s not just one cloud. You have not put your eggs all in one basket.

Anshuman Mandal [01:07:18]:
So there has to be appropriate cloud management platform. Doesn’t matter where it comes from. It could be CloudHealth, it could be, for instance, ServiceNow CMP, whatever it is. You should have a place where you can orchestrate and provision and do whatever you needed to do from one single place.

Anshuman Mandal [01:07:36]:
So cloud management becomes very good from a visibility standpoint as such, for some of the organizations dealing with multi cloud strategy, that is number one. The other thing is in terms of governance, be it your cost control, FinOps optimization, data governance, all this with regulatory compliance, I believe everybody will benefit if there is a centralized or even decentralized cloud service management office. We call it cloud SMO, and I’m a big proponent of that. I have seen it doing wonders. Be it standardization across different global tenants, be it strategies, naming standards, smaller things, service management functions. So it brings in a lot of agility in your environment. You can have it in house, you can have your own people do it, but it is just so important when you’re having multi cloud. And last but not the least, obviously, operations as agile as nimble as you can be.

Anshuman Mandal [01:08:36]:
Previously, our L2, L3 teams used to be big, but my take is we should eliminate first, automate, monitor, and your L2, L3 should be as lean as possible. And finally, once you are in the cloud, you guys have all traversed the journey from adoption of cloud native technologies, be it any application transformation, Docker, Kubernetes, whatever, serverless architecture or even PaaS will give you a lot of mileage in terms of your cost, in terms of benefits, actually elasticity as well. So those were the few things like I can summarize with all the discussion and not to mention the sustainability part of it, I’m a big fan of it. And it is going to take root in every enterprise regardless of the domain, at some point. So these are the five things. Sanjog, thank you.

Sanjog Aul [01:09:28]:
And what a great conversation today. A lot of nuggets. So special thanks to all our panelists, Jonathan Bagnall, Sushmitha Nalamothu, Riaz Yusuff, Brian Gardner, Anshuman Mandal, Arjun Puri, Tony Giampaolo and Satish Kumar Chala. Thank you so much for your contribution to this forum. I’m sure there are nuggets which will really help individuals in different stages of cloud adoption to make sure that you exploit it, get the most value and don’t lose your sleep in the process. So thank you so much again, Zones, for being our sponsor, for hosting the series of virtual roundtables that we have been hosting for many months now and more to come. And have a lovely day ahead, all of you, and God bless. Thank you.