S4 Ep1: Overcoming Cloud Challenges: Exploring the Future of Cloud Computing
About
Episode Guests
We’re pleased to share that the "Livin’ on the Edge" podcast is back for Season 4 after our summer hiatus! With it, we have one new host, Jacob Caddy, and myself (Kay James) will also return as a fellow host. This season, we’re still going to be highlighting all the best Kubernetes practices, but we’re also going to be diving deeper into the world of API development, ephemeral test environments, developer tricks, and more.
To kick off episode one, I chatted with developer relations manager at , Kunal Khushwaha, for an in-depth interview on all things cloud computing and the main challenges that organizations are facing as it relates to cloud-native development.
Transcript
00:00.00
ambassadorlabs
You are good to go.
00:02.77
Kay James
All right. Hello and welcome to the Living on the Edge podcast. I'm your host, Kay James. And today today we are joined by Kunal Khushwaha from SIBO. Why don't you take a chance to introduce yourself?
00:16.61
Kunal Kushwaha
Yeah, I won. My name is Kunal. I work as a DevRel Manager at SIBO, a CNC ambassador. And yeah, excited to be here.
00:26.96
Kay James
And we're excited to have you. So today we wanted to talk about the cloud, right? And I have a lot of friends who have nothing to do with tech. So they'll ask me questions like about whatever's happening. They see me on the news or something. And um I guess in that way, if someone were to ask you, like what is the cloud, um how would you explain that to them?
C
00:50.75
Kunal Kushwaha
Yes, someone else's computer. That's how people say it. ah But essentially, to to a non-tech person you mentioned.
01:01.73
Kay James
Yeah.
01:02.65
Kunal Kushwaha
ah so So let's say you're using YouTube, for example. So all these videos are stored somewhere in like a hard disk or something. And these are essentially stored in someone's computer. And you access these videos over the internet. That's like the cloud.
01:20.06
Kunal Kushwaha
um It's not like like like a literal cloud. Believe me or not, like I have some non-tech friends who thought it was a literal cloud, but it just means that it's just someone else's server or computer. And then there are cloud services, meaning sometimes companies just want to let Netflix, for example, they created Netflix and now they want to run Netflix. Obviously, millions of people use Netflix, so they didn't want to they were like, okay, you know what? we don't want to oh We don't want to have our own servers and you know buy a bunch of servers, build things, set it up, manage everything.
01:56.99
Kunal Kushwaha
We don't want to do that. We are not we don't have that much time. So they just hired a different company to do that for them, AWS. So they use AWS, Amazon Web Services.
02:08.06
Kunal Kushwaha
They use their computers. um I'm like dialing it down to like really like a layman terms.
02:11.62
Kay James
so
02:14.27
Kunal Kushwaha
If you want me to be more technical, let me know. But Netflix was like, yeah, we have written the code, but we don't want to manage our own hardware, so we will just use Amazon's hardware. And that's why there are many cloud companies like that, Amazon, Google Cloud, placebo, DigitalOcean, a lot of companies. But that's ah that's essentially what cloud is. yeah Someone else's computer you're accessing over the internet and various, various use cases.
02:41.22
Kay James
Yeah, and this is a technical podcast, of course, so you can get very technical. But just as an introduction ah introduction to the topic, I wanted to you know kind of talk about what is cloud from a high level. um So on that, um you kind of talked about the challenges right of why people wanted to use cloud. um Can you talk a bit more about some of those challenges, like what were you trying to solve when you know when pete when when cloud was introduced? What were some of the challenges we were trying to solve?
03:09.46
Kunal Kushwaha
I think when it first came into the scene, I was seen as a game changer like after the virtual machines and Ira and everything. um Cloud it you know promised to make it easier and and cheaper to access and manage these technologies over the internet. And there was like flexibility to scale it up, scale it down as you need it. Because virtualization was good, but it still wasn't as flexible as containers, for example, in today's time. So these benefits really made cloud like, OK, if you're a business, you must be on the cloud.
03:45.52
Kunal Kushwaha
um But as time went on, you know some of the issues started to show up, costs began to rise, billing became more complicated, and then there's a bunch of other things like data privacy as we see on the internet, um who really controls the information. you know So it's clear that, like don't get me wrong, the cloud has delivered on some of the promises, but I think it also brought newer challenges that need to be figured out.
04:17.11
Kay James
Can you talk a bit about a multi-cloud approach? Is that kind of one of those solutions for some of these challenges that come with the cloud?
04:25.53
Kunal Kushwaha
Yeah, I mean, ah if you don't know already, multi-cloud is when a company uses services from, like I mentioned, Amazon, Google, Microsoft, digital version. So they use multiple cloud providers rather than relying on just one. And the idea behind is simple. It's it's just that you're not putting all your eggs in one basket. You're spreading your resources with multiple cloud providers. You're reducing the risk of downtime because let's say one of the provider has an issue.
04:55.53
Kunal Kushwaha
then you can rely on another provider. It also helps prevent ah vendor lock-in, meaning that you're not stuck with one provider's pricing or limitations. um And it gives you the flexibility to choose the best services that every provider has to offer. Sometimes you know an AI solution would be better on one provider,
05:16.01
Kunal Kushwaha
a platform solution on another provider, cost of storage would be cheaper on someone else. So it can address some of these challenges of high costs and dependency on a single provider because you're getting more more more and more options and control to choose from.
05:33.63
Kunal Kushwaha
But I think vendor lock-in is one of the main main reasons why people go for a multi-cloud approach, meaning not putting all your eggs in one basket. and yeah Because for the like a big three hyperscalers, for example, if you are really into their ecosystem, moving away from that is going to cost you a lot of money.
05:55.96
Kunal Kushwaha
ah it's not It's not free. I mean, the EU is now regulating it. So they're like, if you ah if a customer wants to move away from cloud because they're not happy, they should be free to do so. um But there's like ah there's There's like a little bit of a challenge with that as well, because with Google, how they do it is they're like, if you want to move away from Google Cloud for free, you have to move away, like move all your services. You just can't move Google Kubernetes or just databases. So not not relying on one provider is a good thing. I mean, we have seen it with like VMware. Now people are, after the acquisition with Broadcom, smaller companies are facing challenges. So it's always good to have a backup plan.
06:40.62
Kunal Kushwaha
you
06:42.09
Kay James
And you kind of jumped to something I wanted to bring up a little later, but the VMware getting acquisition, right? um What are your thoughts on that? Do you have any opinions about that?
06:54.68
Kunal Kushwaha
Yeah, I just want to make money and that's fine. Raise the shareholders value. ah For those who don't know, I think I don't know if it's the biggest acquisition.
07:05.80
Kunal Kushwaha
My my my memory is terrible. But they Broadcom acquired VMware for 69 billion dollars last year in November. and the main shift that they made was from perpetual licenses and now they have a subscription-based model. So that's one thing and now many of the products that they offered are also being killed off um completely and some of the products that remain you cannot avail those as stand-alones.
07:34.73
Kunal Kushwaha
Instead, they have like bundles. So the VMware made like two primary bundles. So you can either get the VMware Cloud Foundation or the VMware vSphere Foundation. So previously, you could get like, OK, I don't want to use all the services that you're offering. I just want to use this one thing. Now, I think it's bundled together. I mean, check the website. You will get a more clearer understanding. But what has happened is because of this new pricing change, it has I think the small businesses are just, they just can't, can't pay that much. So for for example, there's a UK university, you can see this post on Reddit, ah their VMware support cost, it was costing them 40,000 pounds per year. And guess how much it's costing after the acquisition? Half a million pounds. So that's like 1200% increase.
08:24.79
Kunal Kushwaha
And I think for smaller organizations, it's a very big financial strain. I can say that because I work at SIBO, we're a cloud provider, so we are getting a lot of messages like, okay, we want to move away from VA, how can you help us? And not just SIBO, many other companies have their own like solutions now. So I think with Broadcom, what their whole strategic plan is to maximize revenue from the VMware product stack without making it without focusing too much on sales, I would say. And they're like bundling these products available list um and pushing customers into the bundles wherever possible. But that's, I think, essentially why they're doing it, just maximizing revenue in the next next few years. hi I know if it's ah if it was me. I don't know if the bigger companies would stay in the next five years with VMware, with the current pricing models.
09:25.18
Kunal Kushwaha
um yeah I don't know. i why Why not just move to managed Kubernetes? That's a question for the audience. like I have no idea why you would not move to managed Kubernetes in today's time.
09:36.80
Kunal Kushwaha
no maa yeah Focus on the applications, you know not the infrastructures.
09:37.57
Kay James
Yeah.
09:42.64
Kay James
You know, it it sounds like the cost of things is kind of the the crux of everything here, like is is it like cloud itself, like the benefits of using it as like a functional thing are clear and still the same, but it's like the cost is really the big issue here. That's the, you know, for smaller companies, it's harder or bigger companies, just way more expensive and you know
10:06.38
Kunal Kushwaha
yeah
10:06.48
Kay James
That's kind of the battle that we're at right now.
10:09.13
Kunal Kushwaha
Yeah, and that's that's what that's what the initial challenge of cloud the initial promise of cloud was and why it failed to deliver. you know um In 2012, Andy just used the managing director of Amazon.
10:24.08
Kunal Kushwaha
talked about at their conference, how the cloud is going to save like money by ah passing the benefits to the customers. So I quote, he said, AWS is able to take its very large scale and pass on those savings to customers in the form of lower pricing. That's what it was mentioned. Fast forward like now 10 years later, um we are seeing rapid cloud increases,
10:52.24
Kunal Kushwaha
There was a survey by Dark Matter that said that 82% of the organizations are struggling with the cloud costs. And it's not just that, it's it's that the power with cloud providers, the big three, Azure AWS, Google Cloud, they now hold so much power that's creating concerns about costs and complexity and just getting logged into one provider. so so So some of the challenges I think were solved, but I think with with rising costs and non-transparent pricing.
11:30.50
Kunal Kushwaha
I think there there are more more and more challenges. And not just that, I mean, there's a report by BBC, which says that the data center power use is expected to surge around six fold in the next 10 years. So there are no environmental effects as well, especially with the whole large language models, training, GPUs, it's not good for the environment. So we have to think about environment as well. I mean, I get it.
11:56.05
Kunal Kushwaha
you know I come from Delhi. so Delhi is like one of the most polluted cities in the world. It's just funny because there's so much advancements. A new iPhone came out, a new AI tool came out, a new AI robot came out, a new Dyson came out. and Then and I just step out of my house and I so look around and I'm just like, okay I can't really breathe good air. So what's the point of all of these advancements when people are not having basic like clean air and water and, you know, other things. So that's just, yeah, something to think about.
12:38.35
Kunal Kushwaha
so
12:39.66
Kay James
Do you have any ideas on how, I guess on both sides, how the companies who use the cloud and then the companies who are providing the cloud, how we can do this more ethically, like on that note of like, you know, consuming too much energy or something? Like, is do you have any ideas on the ethics of this and how we can be better?
13:01.17
Kunal Kushwaha
Yeah, it's a good question. um Thanks for sharing. And being ethical, like obviously there's, you see like a lot of company CEOs and stuff in the parliament, for some reason or another, ah they're answering questions, paying fines or whatnot. So it's, it's I mean, and that's why I'm starting my answer with like data privacy and security. And I think
13:26.10
Kunal Kushwaha
being open and clear about your cloud practices is essential. So essentially disclosing where the data is stored, how it's processed, who has access to it. So you know transparency, because in the end, it it just builds trust and allows users to make more informed decisions about about their data. And as I mentioned already, the cloud has significant environmental impact due to the energy consumption of data centers. So as a company, you can adopt more sustainable practices by choosing providers that use renewable energy. And trust me, like we are a cloud providers, so I know that there are like almost every country and government, they are doing some sort of work with like renewable but renewable energy. A lot of investments are being made in those areas as well. So yeah, picking cloud providers that that use renewable renewable energy, I think that's one of the key steps that you can take, because that's only going to help those initiatives furthermore.
14:25.15
Kunal Kushwaha
and optimizing your own cloud usage to reduce waste, seeking um more and more ways by which you can minimize carbon footprint. So a good example for this is a company called Deep Green here in the UK. So what they do is they they take the heat generated from your data centers and servers, and they use it to heat the swimming pools.
14:48.85
Kunal Kushwaha
and houses and stuff. So all these companies are like, we're trying to be carbon neutral, and they're like, we're being carbon like negative. So um that's just that's pretty cool, you know that that heat is being used utilized somewhere, and it's not just going in the environment.
15:06.94
Kunal Kushwaha
One last thing that I think made one of the biggest x that I have in today's time is just cloud costs. It's not just about rising costs. It's about transparent pricing. So you should as a company should have transparent pricing model.
15:22.38
Kunal Kushwaha
and not exploit smaller businesses or customers customers. So avoiding practices that lead to vendor lock-in, which can limit the customer's choices and also flexibility. So as a company, the approach that you should follow is that the customer should stay with us because they really like us and not because we are forcing them to do it. So I think that's the transparent pricing is very, it's one of my main, yeah, as a company that you should Yeah, follow.
15:52.62
Kay James
Yeah.
15:53.05
Kunal Kushwaha
But the big bigger companies don't follow that. that's They want to focus on like maximizing shareholder value and whatnot. Like Google, it's a monopoly, but they don't like to come off as a monopoly, for example.
16:06.43
Kunal Kushwaha
Yeah, transparent and pricing wherever you can.
16:10.63
Kay James
Yeah, I like that too. And I have a similar story of um there's I live in New York and there's a bath house out here that they heat their pools with like Bitcoin mining, something like that.
16:20.81
Kunal Kushwaha
Cool.
16:22.29
Kay James
And it's like, OK, like, you know, people have their opinions about Bitcoin, but at least they're like kind of doing something positive with the energy that they're creating or wasting or however you want to look at it.
16:33.14
Kay James
um But yeah, I guess continue on the idea of ethics and stuff.
16:33.32
Kunal Kushwaha
Yeah.
16:37.00
Kay James
um Open source, is there any connection to how that can play into a more ethical or positive cloud experience?
16:49.02
Kay James
or I guess as open source is evolving too, like does that play into it at all?
16:53.25
Kunal Kushwaha
Yeah. I mean, when we talk about AI and cloud computing specifically, there's a research by Gartner. I mean, maybe you can link in the, the in the podcast, uh, the machine learning research. It mentioned that 85 was the, of the machine learning projects failed to deliver and only 50% or something. I don't know the exact number, but maybe it's 54, 55 or 53. Uh, it's somewhere in the fifties. Uh, they,
17:21.43
Kunal Kushwaha
progress from the prototype to the production stage. So only like 50% of those. And there are two main problems. So like yeah that the first one being the project failure stage. So the the harsh reality is, now this is interesting that in the industry,
17:38.04
Kunal Kushwaha
um The harsh reality is that with machine learning projects, they are very expensive expensive to set up and run, both in terms of like the upfront cost and time as well. Because before even developers, they can start to like generate machine learning insights. um They have to configure let's say various parts of a complex.
17:58.77
Kunal Kushwaha
infrastructure where their model is going to run. They have to manage machine resources. They have to set up like monitoring and for machine learning models, things like feature extraction and all these other things. So after they do all of these things, only then a small part of the system is actually usable machine learning code.
18:16.89
Kunal Kushwaha
And in that survey, you will also find that 11% of the people said that they have never abandoned the project. And that's like 11% people said they have never abandoned a project. So 89% people have abandoned projects. And the second big issue is when we talk about AI and cloud computing is time.
18:37.53
Kunal Kushwaha
Because, you know, if you're if you're using let's say AWS SageMaker, that is completely different. If you're setting up AWS SageMaker, it's completely different than setting up the Azure Machine Learning platform. Okay, so it's not just that the setup that's complicated, but, you know, building the needed infrastructure that also takes a lot of time as well.
18:56.35
Kunal Kushwaha
Google did a research on this that for every eight hours spent on machine learning engineering, like if you're a company and you want to spend eight hours on machine learning engineering, your company has to put in 24 hours for feature engineering, 32 hours for data engineering,
19:16.53
Kunal Kushwaha
And 96 hours for infrastructure engineering. So that's like 160 hours of work. And only 5% of this is actual machine learning engineering. So if you're a bigger organization, that's not a problem, cost-wise, investment-wise. But if you're a smaller company, then and that's an issue. Because um you know you you can't afford to make that investment.
19:40.46
Kunal Kushwaha
So to answer your question on open source, why I brought up this point is that in the survey, it was said that two thirds, which is around 66% of organizations use open source tools for deploying machine learning applications and models. And this has significantly improved ah time savings.
20:02.06
Kunal Kushwaha
And um there are projects like, I mean, it's it's it's not just because it's open source, you're not just limited to expertise from your own company. Kubeflow, for example, the version 1.7, it was developed by 250 people from various organizations. So many many organizations don't have dedicated machine learning teams. um with that With that in mind, I think open source models make ah make sense for how machine learning is being been deployed today. So I think,
20:31.72
Kunal Kushwaha
Yeah, plays a huge role for companies. um Yeah, it's just making it more accessible to a broader audience and leading to more problem solving issues.
20:36.60
Kay James
Mm hmm.
20:42.57
Kunal Kushwaha
ah Doing a lot of social good as well. I mean, the projects are out there. you can You can advocate for projects. You can advocate for ethical practices across the industry. You can support communities, ah contribute to social good and all these other things.
20:55.33
Kunal Kushwaha
um But yeah, I think especially with AI and cloud computing, the project abandonment issues and time time issues are being solved by open source. ye
21:07.01
Kay James
So what are some of the things that are happening right now in the cloud computing world that you're most excited about or most interested in?
21:21.24
Kunal Kushwaha
I'm always excited about, well,
21:28.50
Kunal Kushwaha
In cloud computing, well, AI, obviously. I don't want to talk about AI too much, because everyone is doing it. But yeah, specifically specifically specifically with cloud, how AI is integrated in in in various use cases. For example, there are some companies who are utilizing with cost monitoring. um Security is a big use case. So figuring out the security vulnerabilities before they reach production, I think that's a good use case.
21:56.08
Kunal Kushwaha
I would personally love to see a use case where AI can help me in like troubleshooting Kubernetes ah clusters. I think that's one one area. there's There's a tool like K8s, GPT. I haven't played around with it much, so I can't say much, but people are definitely doing doing a lot of things. Platform engineering, that's another big buzz around recently. i don't think it's I don't think there's, in in terms of like the trends, I don't think, I think there's one consistent trend that has been going on for the past decade, which is abstraction. Like cloud did it for virtual machines, Kubernetes did it for containers, and now platforms are doing it for applications. So I think in the future, there's going to be more, yeah more more abstraction teams would be smaller. And yeah, um um personally, I'm very, very much looking forward to the a
22:49.17
Kunal Kushwaha
to the the environmental initiatives by by cloud companies. Because I work at SIBO, we're a cloud provider. And that's personally what i what I'm really excited about.
23:02.65
Kay James
So you brought up Kubernetes. um i guess how Can you draw how talk more about that abstraction part, how like you know Kubernetes was the the thing? you know Do you have any more thoughts on that?
23:20.08
Kunal Kushwaha
Yeah, I mean, um if you talk about, so it acts as a, if you talk about Kubernetes, if you talk about, let's say a multi cloud approach, one of the challenges is that it can get a little complex. Okay, okay I'm using GCP, I'm using Azure. So Kubernetes, it acts as a unifying layer that abstracts away the underlying complex complexities of your individual cloud providers.