Distributed Cloud – Digital CXO Podcast EP86

Hello and welcome to the Digital six. So podcast. I'm Amanda Razani, and with me today is Ari Weil.

He is the Cloud evangelist for Akamai. How are you doing today? I'm doing great, Amanda.

Thanks for having me. Happy to have you on our show. So can you first talk a little bit about Akamai and what services does the company provide?

Absolutely. So Akamai is a, uh, cybersecurity and cloud company that really focuses on providing enterprises with the best platform to build, deliver and secure all of their applications anywhere that their users might come online. Wonderful.

And then share a little bit about your background for our audience. Yeah, I'm happy to. So I've been, uh, in the technology industry for about 25 years.

Um, I've held roles from development to solutions engineering, product management, um, product marketing, and, uh, overall running marketing teams all focused on conceiving of and bringing products to market and then scaling 'em throughout their life cycle. Uh, my favorite thing about my job is always working with clients and customers to understand their problems and to help 'em solve them with breaking technology. Wonderful.

Well, I know that, um, pretty recently y'all commissioned a slash data survey of, let's see, 700 professional cloud developers and 38% cited improved user experience due to reduced latency among the top five benefits of distributed cloud. So let's talk about distributed cloud. Uh, it seems like, um, modern workloads, they're getting more and more and businesses are looking toward distributed cloud solutions.

So what can you share about that? How does that work and how is it a better, uh, answer for businesses? Yeah, absolutely.

I'm happy to. So if we take a, a very abbreviated sort of journey over the last, call it 10 years, businesses have gotten very comfortable and have started to use cloud technologies to accelerate their time to market, to, to save, uh, on the cost of bringing a product to market and to scale it to meet the demands of their customers. And we've seen the adoption and the widespread growth of cloud native architectures, which are basically going from monolithic applications that were built to be run on premises and then scaled out through delivery networks, uh, and other delivery partners to their customers, to using cloud technologies in order to create, uh, a services based architecture.

So little microservices connected through APIs that can help them both bring a product to market faster, scale and advance certain elements of the product as the needs arise, as the scalability needs increase, maybe because they have a promotion or they're seeing a lot of interest from a given customer segment, or maybe they need to process more data. Um, and then the API based architecture also allows them to take advantage of things like open source software versus commercial software to use multiple clouds or a hybrid infrastructure where you have data in a data center on premises and then also some of your infrastructure in a cloud. But that flexibility and the architecture of the the cloud native approach also comes with some challenges.

Uh, it's hard for the IT teams and the security teams that need to manage those applications to understand all of the component parts of them or how workflow actually operates across all of these different services and through API so that they can apply the right, uh, resiliency, the right high availability controls, or even the security controls that would ensure that only the right people have access to the data, that they can identify malicious traffic or threats to their application, or they can avoid misconfigurations that might allow things like ransomware to spread, uh, throughout the organization and ultimately do some of the things that we've seen, which would compromise a business or impact end users when data becomes either encrypted or stolen. Uh, and so this shift to cloud from from cloud native to edge native, which is doing all of these cloud native things with composable architectures, with the microservices and the APIs is moving out to the edge, the edge being required because people want experiences to be faster. You know, when you and I started talking today on this video conference, we expected everything to work right away.

We wanted to have our video delivered in high fidelity. We want our audio to be synced up with what we're saying. It needs to be optimized for the different devices that we're using.

And that's similar to any other sort of an end user experience that's powered by technology. If people are flying, they expect, you know, what they see in an airport to be synced with what they see on their devices to make it really easy for them to check in and get onboard to their flight, to even consist of all the media entertainment that they would consume while they're in flight, you know, from one destination to another. People who consume retail experiences wanna have a very consistent experience that suggests that their retailer knows who they are.

And enabling those sorts of things requires that move from a centralized cloud to the edge. Now, in order to do that, historically you would've used these clouds like AWS or Google or Azure, and then a delivery partner at an Edge network like Akamai to take the results of your compute and then to create content that would be shipped out and cached closer to the user. And then platforms like Akamai would add some edge computing to apply personalization or some dynamic content.

But developers have said, well, that creates a really fragile complicated architecture for me. And IT teams say, it's really hard for me to orchestrate and troubleshoot when I have so many different things happening in my application where I'm computing here, I'm caching there, I'm applying transformations at the edge. And then somehow security has to keep track of all of these things.

So what distributed cloud is doing is it's saying, let's have the opportunity on one platform to have my core computing for things that are, are heavier weight today. We would think about it like a lead large language model that would power ai. Then if I need to take some lighter weight compute and move it closer to where my users are, I can do that on the same platform, but I'm distributing that compute.

So it's still actively something that my developers are managing. It's the infrastructure that they, they build upon, but it's closer to where the users are and I can scale it now where I see demand. And so instead of having to scale up in a centralized data center in anticipation of demand somewhere else, I can be very responsive to demand where the users generate it.

And then beyond that, if I wanna apply even more lightweight computing at the edge, then that distributed platform will also provide me with the option to have my compute extend to the edge closer to where the users are coming online. And then I've got a continuum of compute resources that as a developer I can take advantage of and really architect my application for that scale and for that low latency use case without having to think about a very complex multifaceted architecture where, you know, anytime I make a change, I worry that I might break some aspect of my, my application, I have more control over it end to end, and I have more flexibility to build the features that my customers demand of me using all of the skills that as a developer I possess because I know how to build technology in the cloud. Hmm.

And so with this distributed data, how do business leaders handle the security concerns with the data? So when you're creating an architecture like, like we do, like we enable on a distributed cloud, we're also providing you with the ability to, to apply consistent security controls on that architecture, because now I have visibility into the entire lifecycle and the entire continuum of my compute and that workload as it's flowing through my application. And so companies like Akamai that have an edge network and the ability to apply security closer to where, you know, attacks might originate, but also where the users are coming online, and then we can follow, uh, the entire architecture back to the, the core cloud or an on-prem data center.

We can apply one consistent set of controls that would be delivered and and managed at the edge. But that applies to the entire stack and the entire continuum of the application. Now, it's also important because businesses will continue to use multi-cloud architectures and multiple cloud partners because there's a problem around concentration risk that exists today.

You know, you could have a security vulnerability on one cloud, and you need the ability to then make sure that your applications are continuing to function while you address that vulnerability. But you may need to take certain systems, not necessarily offline, but shut them off from the external internet Ari Weil them off from the external internet. And you need a way that you can both identify where those problems are and then have the ability to maintain continuity of your application even when one of your cloud partners has that issue.

And so in distributed cloud, we are providing the option for portability and multi-cloud, but applying security holistically across that entire application infrastructure. And so what you'll see basically is one set of controls a consistent way to apply them on one cloud or multiple clouds, and then a single control point so that you can make the adjustments that you need at the front end of the application where the data is flowing from machine to machine or for the infrastructure that is actually powering the application itself. Okay.

And so while they're moving some of this from these older legacy systems, are those, uh, just end up as technical debt, or are they still using Both Companies are gonna continue to use a mix of technology and, and really it becomes both a security consideration and a business consideration as far as how much you're gonna modernize certain parts of your, your architecture. Um, years ago there was this concept that people started to talk about of repatriation where when the industry got really excited about moving things to the cloud for scalability and time to value and the added security, uh, controls that were inherent in the cloud, some businesses started to realize that not every part of their architecture needed to live in the cloud. There was no cost benefit.

It actually might have exposed them from a, a compliance or a regulatory perspective, or even from an IP preservation perspective where they were afraid, you know, they might lose control of their core IP assets. And so people started to move some of that infrastructure and some of those applications back to on-prem today, what we realize is that mix is going to be dictated by business necessity, whether it's the ability to reach certain users within the latency that you require, whether it's data sovereignty considerations that mean that some of your data and decisions need to be managed inside of the country, that it originates or where the people's data needs to be preserved. Um, and so that mix of old and new technology of on-prem and cloud is going to continue into the future.

And what we'll see, I I'm sure are some of these regulated industries like, you know, financial services, whether it's consumer banking or trading or insurance, healthcare providers, government entities, they'll have some of these legacy systems that simply don't need to be modernized using some of the, the latest technology because they function perfectly well and there's no business or practical advantage to, to move them. But the customer engagement elements, the things that need to be dynamic, that need to be fast and customized and take advantage of things like artificial intelligence, those will continue to evolve and those will leverage both cloud and distributed cloud much more. Okay.

So when it comes to this change management process and, and they're distributing to the cloud, are there any roadblocks that you see business leaders face or any place where they have hiccups that you can share with us? And then, uh, solutions? So one of the benefits of distributed cloud is actually that it simplifies, uh, the orchestration and the scalability considerations of an architecture.

When you are moving from on-prem to the cloud, if you're using a hybrid infrastructure, there's a lot of orchestration that has to occur for high availability, for scalability, for disaster recovery because you're thinking about, you know, uh, availability zones that are geographically distributed in case there's a natural disaster or there's a power outage, or you know, a data center is compromised for another reason. Uh, you're also thinking about how do I need to scale the individual components of my architecture? And so teams need to put a lot of thought and a lot of time and effort into monitoring, into orchestration, whether it's, you know, process wise or code-based.

And those things can really take up a lot of the, uh, in intelligent capital that you manage. So the people in your organization, it can also make it very difficult for you to find and address problems. And so what distributed cloud is, is bringing to the market is the ability to scale automatically based on demand.

So we're taking the benefits at Akamai of 25 years of orchestrating workloads at the edge and being able to scale and distribute data and decisions automatically for our customers. And now applying that to cloud computing in the distributed tier. So instead of you having to think about retraining, re-architecting and coming up with new processes to move from core cloud to distributed cloud, we wanna say take all of the advantages that you've developed over time with your familiarity with cloud technology and let us apply the automation and the, uh, auto scaling and the reliability of the edge and marry those two together to make it much more seamless for you to take advantage of distributed cloud.

Now, what source of challenges does that present to an enterprise? It is a newer concept. It's growing in, in, and we're in the early days of adoption and understanding of it.

And so there's a perceived complexity there where people are saying, you know, you're telling me to go from one model to another model. And so the, the responsibility that we have as we start to, uh, de deliver distributed cloud to our customers is to provide them with, you know, enough documentation, process, work and services so that we can help them with that migration and help them with that adoption so that they realize that it is not as complicated as they think that it is. And we're, we're really engaging now with media and entertainment companies, gaming companies and SaaS companies to show them just how easy this can be at scale.

Uh, and what we're seeing is people are really delighted by the outcomes because they're able to address, you know, new problems that are emerging with technology, with security considerations, and also with customer demands, but with the existing people and the existing skills that they already have. Okay. Well, if there was one key takeaway from all this that you could leave our audience with today, what would that be?

Really the, the biggest takeaway is as we are looking to see how can we capitalize on, you know, the, the growing trend with generative AI and AI powered applications, things like computer vision and AI inference, that's really creating these next generation applications, what we call edge native applications. Uh, distributed cloud is going to be the architecture that will power that next wave of next generation applications. Um, and so it's really important for technologists and, uh, business leaders to familiarize themselves with the concept of distributed cloud.

And to start considering how would you architect your application today if the complexity of cloud and some of the constraints that you've been under for, you know, lock-in or concentration risk or even skyrocketing costs, were not a consideration. If you have a platform that is open that provides you with, with easy to use low cost services, with integrated delivery and security as part of them. So basically like if you had your perfect architecture, then what would you build and how would you engage your audiences differently?

And you should look into how distributed cloud will enable you to meet, uh, those needs. Wonderful. Well, I wanna thank you for coming on our show and sharing your insights with our audience today.

Sounds great. Thank you so much. I appreciate it.

And I wanna thank our audience. Stay tuned. There's more.