Integration As An Accelerator for Innovation

Jason Andersen: Hi, I’m Jason Andersen, Vice President and Principal Analyst at Moor Insights & Strategy. I’d like to welcome you to this episode of Six Five On The Road. Today, we’re at AWS re:Invent, here in Las Vegas, and our topic is actually integration. And we’re being joined today by Justin Callison and Bill Boora, and excited to have you here. We’re really excited to talk about this topic. And why don’t we just start with each of you giving an introduction on what you do at AWS?

Justin Callison: Sure, I’ll start. So I’m the Director of Application Integration, and so it’s a group I’m responsible for building a set of services that are critical to integration on AWS. And I also work closely with a lot of other service teams. So I’m a part of the team that builds these capabilities that enable customers.

Jason Andersen: Okay.

Bill Boora: And so I lead Go-To-Market Worldwide for our application modernization and integration services. And so I work closely with our customers to really understand what their modernization goals are and where integration can play a role in that, and loop that back into the underlying services that AWS provides.

Jason Andersen: Perfect, perfect. So, why don’t we talk about the customer side of this in terms of, how are organizations thinking about integration today, particularly around modernization of their existing applications and moving them into the cloud?

Justin Callison: Yeah, so customers are thinking more and more about integration in a strategic manner. So integration is something that you can think about tactically to solve specific problems, but what we’re finding is for a lot of companies, integration is becoming a strategic imperative. And the reason behind that is really the challenges that these customers are facing. So, the challenges they’re facing are much the same as they have for many years, but those challenges are growing and they’re accelerating. So, one is the pressure to modernize is increasing. As time moves on, these legacy systems that have formed the basis for these organizations and their technology before, they’re aging. They’re getting farther along, and there’s a greater and greater need to move those to cloud to modernize them into new architectures. This is building over time, and so that’s creating additional pressure.

Jason Andersen: Yeah, I bet you the skills are also a problem.

Justin Callison: Exactly.

Jason Andersen: It’s not just the technology, right? It’s also the people.

Justin Callison: Yeah, it’s the people, it’s the infrastructure, it’s all of these things. The second one that’s accelerating is the need to innovate. Every organization is under tremendous pressure to innovate in order to keep up with their competitors and keep up with their customer expectations. You have a lot of these organizations that have these more legacy systems that they’ve built and have driven their past success, but now they have to compete against new competitors that are maybe born in the cloud, building modern architectures from the start. And we add in the advances in AI over the last while that are allowing new folks to come in and challenge incumbents. So the need to innovate is greater than ever.

Jason Andersen: Okay.

Justin Callison: Now with that, one of the other things we’re finding with our customers is they’re recognizing the need to embrace diversity. So in the past or in other times, they have thought, I need to modernize and I need to take all of this technology, these rich technology assets and move them all at once. And what they’re recognizing more and more is they need to embrace the diversity. Some of those things are better to stay where they are, some of them are better to move parts of them, some of them are moving wholesale. They also recognize they want to go to SaaS providers or cloud providers like AWS to get components, and those are going to be diverse and different. And this is what’s driving the importance of integration. They need to think about integration from the start so that they can take advantage of all of these diverse components.

Jason Andersen: Okay, okay.

Bill Boora: Yeah, and more and more as we speak to customers and particularly the technical audience, we’re starting to see a shift between the technical teams providing a service or a bit of a cost center for the organization, to really becoming an enabler for the business and becoming a profit driver. And so some things that Justin talked about, pressure to get applications out quicker to move through the release cycles faster, so instead of six or 12 months, to monthly or weekly. We’re also seeing a lot more from a customer experience perspective, where events are really starting to derive the outcomes of these solutions, right? So if you think about an airline industry, for example, there’s a weather storm. What impact does that have upon the flight schedules, the hotel bookings, insurance? It’s the same thing. So there’s a lot more of that event type stuff and integration to these SaaS applications that Justin spoke about.

Justin Callison: And something that Bill just said, he talked, Bill talked about events. So another thing that we’re seeing as a really strong movement in the marketplace is embracing event-driven architectures. So event-driven architectures are a pattern that’s been around for a long time, but with the advent of the cloud, advent of some of these modern architectures, is really allowing those to accelerate and move more quickly. And so one of the things that a event-driven architectures help when you’re thinking about integration is they allow individual participants to remain decoupled where they can take… Maybe you have a legacy application that’s built on a relational database. You can source events from that on important things that are happening and then make those available in the cloud. And then you can allow that innovation to work where the producer here doesn’t need to change and these other pieces can move at a different pace. This follows the diversity that I talked about and allows the organization to move in a very agile way.

Jason Andersen: Okay. It’s interesting just because of how it’s evolved. It used to be one of those things that was like an inhibitor to moving forward with innovation, right? When people were talking about way back with service oriented architectures, things like that, where it was like, oh, we got to do that before we get to the cool stuff, type of thing. Now it’s good that it’s integrated, it’s a much more integrated and also it’s seen more strategically.

Justin Callison: Exactly, enabler of innovation. And some of the offerings that we have in AWS are a big part of what’s changed that equation. Where previously as you said, these things would be an inhibitor because they required organizations to invest a lot upfront in oftentimes licensing, oftentimes in expertise, in order to get started. With the cloud, they can take this and move forward immediately.

Jason Andersen: So can we double click a little bit on AWS’s approach to integration then? That’s a good jumping off point.

Justin Callison: Yeah, so maybe I’ll just say a couple words about AWS in general. So one of the things when we look at AWS and one of the things, I’ve been here now for about six years and I’ve learned a lot about how to drive innovation and continue at scale is one, we’re disproportionately customer obsessed. We don’t build things that we think are going to be cool or sound cool. They’re based on concrete customer problems, that’s the first thing. And the second thing is we don’t believe that one size fits all. There’s such diversity in needs in customers that we really want to focus on providing purpose-built solutions that our customers can use and apply. And that’s why you may find, for example, a data streaming service that people use in integration like Kinesis, that’s fully AWS native, but we also provide managed services for Kafka. So similar types of patterns that you will see, and it’s about providing the right tool for the right job.

Bill Boora: So generally speaking, when we talk to our customers, whether it’s the business side or the technical side, we are hearing that they want to have cloud-based solutions. So when Justin talked about, they may be at different stages of their maturity so they may have some stuff on-prem, they may have stuff in the cloud, they’re building microservices. As they move into AWS, they’re looking for solutions that are also cloud-based so that they can take advantage of the speed, the agility, and all the benefits that come with cloud sets. That’s number one that we’re hearing from them. Secondly, Steve, you talked about specialized skill sets, right? And so the ability for them to leverage the AWS skills that they’ve got, and then they’ve got these diverse use cases. So being able to have purpose-built solutions is important. And then finally, licensing has become, maybe not… A big challenge for some of these organizations. So not only getting away from those expensive upfront licenses, but having lower cost operations and longer-term benefits there.

And so our approach is to really, number one, our solutions are, we follow the serverless model, which means that we take care of that operational overhead that customers are typically dealing with. We provide purpose-built solutions for various application file and data integration use cases. So if we think about where these applications are, they might need to integrate with APIs. With events, they might have some complex workflows that they’re running, so, EDI use cases as well. And then also some of their data might be on mainframes, and it’s really important in terms of getting that into the cloud so they can make use of that, but also keeping things synchronized. And so the ability for us to give these purpose-built solutions that they can now build for their own business use case is really important. And then finally, on top of that, it’s all a pay per use model, right? So there’s no upfront licensing associated with it. Pay per use helps them scale up and scale down, it’s only paying for what they consume. And then in the long run, this model gives them the lowest total cost of ownership as well.

Jason Andersen: There also probably has to be a bit of a support angle to it, because a lot of these integration platforms that are in the legacy space, as they end their useful life, the support costs can go up fairly dramatically. So it’s not just licensing, it can also be maintenance and servicing cost for the software.

Bill Boora: Yeah, exactly. And in this case, they can leverage the resources on AWS at least, that they’ve trained up, do whatever certification programs and whatnot.

Justin Callison: And what Bill was talking about and a lot of that comes to the skill set of developers.

Jason Andersen: Actually, that’s a huge point. So, let’s talk about that one a little bit more, in terms of developer empowerment. Is it a big skill leap to go from the on-prem legacy idea towards serverless and those types of things? First off, there’s the empowerment, but there’s also the, we’ll call it change management associated with that too. So, how does that kind of fit into the picture?

Justin Callison: Yeah, so at AWS we’ve embraced developers for… it’s really the core of what we’ve always done, really embracing developers and providing them the capabilities that they want. And so this has been an approach we’ve taken for a long time, and what we find in the integration space is that developers are key stakeholders, whether they are using another team’s offering in order to do integration, or doing it themselves. And so integration tends to be a very core aspect of what all developers are doing. And then it really fits with some other trends that we’re seeing going on. The first one I talk about is platform teams. So I talked about diversity and I talked about being purpose-built. We provide those purpose-built solutions at AWS. We have a lot of organizations when faced with common challenges, maybe ones that have specific security requirements where getting things wrong is expensive or many teams need to do it, they’re employing platform teams, and those are typically internal teams, including developers who want to build that solution that’s purpose-built for the organization. And we’re seeing a lot of this, and the advantages of that is they are not limited.

One, they can make it bespoke for the needs of the organization that they have. They’re also not limited by scaling or licensing that could cause impediments with some other approaches. So those platform teams are developers, but they want to have the components and the building blocks that AWS provides in order to build it. So that’s one really big trend we’re seeing. The other one we’re seeing is generative AI. Generative AI is creating more developers. In the past, being a developer and taking a developer-driven approach to solve problems would require a skillset using a general purpose programming language in order to work with these things. The advancement, things like Amazon Q, are making a huge difference in helping people to gain that skillset of being a developer, get access to this expertise, and learn this language in order to solve problems. And so we see that happening and reinforcing developers driving integration in the future.

Jason Andersen: Just let’s ask a follow-up to that one. There’s been a lot of talk here at re:Invent about agentic programming, right? It’s a topic we’ve been researching a lot at Moor. It sounds like it’s an enabler to people having to not know SOAP and all this other old stuff, in terms of getting it done. So, but is there some level of integration between agentic applications and the integration services themselves, or is it actually pretty straightforward? Just out of curiosity.

Justin Callison: I think it’s something that’s really emerging and it’s going to be really exciting to see what happens over the next while. But I see these really going sort of hand in glove in moving things forward. So when we see agents being applied, we can see them in the software development space of helping those people building one can imagine, someone who needs to… maybe they’re in a role where their responsibility is to integrate two systems together, and agents being able to help them to apply best practices, to apply validation in what they’re doing. But we also expect to see agents used for driving integration, for being the things that are doing the integration maybe between a conversation that’s happening with others. So we see that happening a lot, I think it’s a really exciting and emerging space.

Jason Andersen: I also think it has a big role, especially with integration where performance comes into the fold, right? Making sure that you’re getting the right throughput and that the right data is getting to the right place at the right time. So I could see an observability angle being a pretty big play there as well.

Justin Callison: Yeah.

Jason Andersen: So, why don’t we talk about the cost and operational benefits that organizations see when building AWS integration solutions.

Bill Boora: Yeah, and so at Amazon, we think of it not so much from a usage perspective or a licensing perspective, but really, that total cost of ownership. And so when you look at the total cost of ownership approach there, as I mentioned earlier on the serverless side of things, the operational overhead goes away. And so the organization can focus in on what’s going to drive business value for them. The other piece of it is there’s no licenses associated with our solution. So your upfront and your ongoing costs go down. Our integration services are all paper use or paper consumption, so they’re only paying for what they use and no more. So those are some of the overall benefits there.

Justin Callison: Yeah, and that total cost of ownership, it is really key because it’s not just about doing these things, it’s about doing them well. When you’re talking about integration, the durability, the reliability are of paramount importance.

Bill Boora: Absolutely.

Justin Callison: And at AWS we invest a tremendous amount. It’s really our bread and butter and our expertise. And so being able to take advantage of that foundation that we’ve provided. The serverless mindset is really about allowing our customers to focus just on their business logic, small amount of code focused on the things that are core to what they do, and let us provide the scalable, cost-efficient support.

Jason Andersen: Cool.

Bill Boora: Yeah, and so when Justin talked earlier about how gen AI can be applied, this also keys into your total cost of ownership as well, right? So as we get assistance helping you optimize your code, you’re getting to market faster. Developers, it’s easier for them to develop so that the investment that needs to be made, in terms of that skill set goes down. So it’s really that whole package of things when we talk about the agility and the velocity of getting things out, as well as that cost.

Jason Andersen: So why don’t we talk about a couple of the newer capabilities, right? So maybe we start with step functions.

Bill Boora: Sure.

Jason Andersen: If we can talk a little bit about what that’s all about.

Justin Callison: Sure. So step functions is a workflow service, a visual workflow service. So what is a workflow? It’s basically a set of steps that you want to enable a system to execute reliably for you. So reliable execution of multi-step processes, and these are fairly key in integration because you have different pieces where you need to take an action. We also need to compensate where maybe that system that you’re talking to may not be available. So you now need to be resilient to that outage, or you may get a response and say, based on this, I need to take some other action. And so it’s really critical for our customers to be able to describe in a workflow a reliable set of steps in order to manage their integrations. And so step functions plays a really key role.

The other thing about step functions, it’s a visual service. So the way it presents workflows is what we call a state machine. So even like a flow chart, do this, and then when this is done, go do that. After that, maybe I go here. And it’s really helpful because it allows people to visualize what they’re doing, and it provides a visual interface called Workflow Studio that you can use to compose these things, a drive and snap. And it’s really, really helpful for people to visualize the steps that they’re going to do when they’re building and carry that all the way through to when they’re operating this in production. Step functions itself integrates with every AWS service, it integrates with any public API through such as SaaS providers, and some new capabilities for doing with private APIs as well.

Jason Andersen: So how about EventBridge? You talked about event-driven architecture, so maybe we talk a little on that particular capability.

Justin Callison: Sure, yeah. So EventBridge is cloud-based event router, and so it’s really about connecting sources and destinations of any form. So when we think about, EventBridge, it has a few key features. The first one is what we call an event bus. The concept of event bus is pretty well understood within technology, right? And so event bus has enabled many to many communications. So you can have many producers that will send events to the bus, and then you have many consumers around this. So this pattern is really critical in integration. The second feature that I’ll talk about is called pipes. And pipes is, as the name implies, a way to connect one source to one destination. And sometimes you need that sort of one-to-one type of integration. And then we also have a scheduler involved because very oftentimes you need to run these things at a specific time. And so we see a lot of customers embracing EventBridge as the basis for their event-driven architecture and integration on top of AWS.

Jason Andersen: You mentioned APIs a couple times. The world is now API-driven when it comes to integration. So, there’s also some new capabilities on private APIs and what does that ultimately mean to the customer, right? Because that’s becoming a big deal, especially for customers who want to deploy those in the cloud.

Justin Callison: Yeah, so we’re really excited about this. We just launched it this week. So I talked earlier about how both EventBridge and step functions can communicate with other APIs. So prior to this launch, we were able to do those, but those APIs needed to be publicly accessible. So, really great for when you’re integrating with SaaS providers, etc, but very often, we talked about, if I go back to where we started, these organizations have different applications. They have thousands of applications, and many of those are running in private network environments, And they want to be able to have those and help them compose.

So now with this capability, the organization can say, “This specific API from this network environment, I want to make it selectively available,” and then we can have EventBridge and step functions, both connected to those events. What that allows them to do is to bridge across these different network environments and really make this whole integration movement a lot simpler, a lot easier, a lot less need to invest in development to help assure security, assure compliance, assure performance. We can take care of all of that for you. It’s a really great accelerant to these modernization efforts.

Jason Andersen: That stuff always sounds great. New features are awesome. Why don’t we switch and kind of close out with some customer stories? So I know you both have some to share, so we’d love to hear them.

Bill Boora: Yeah, and this is actually part of one of the things I enjoy most about my role is that I get to talk to customers about what they’re doing. We’ve got customers across the spectrum that use integrations for, it could be data processing workloads or real-time analysis of things, all the way up to, they need to go and replace their integrations platform. Some of the most novel or innovative things that happened were back in the pandemic, for example. Taco Bell is an interesting example of this because during the pandemic they had to close down their in-store operations and moved to an online model, and then they partnered with a delivery app service. So initially, they had to swivel chair where they would input an order into their POS system, and then they’d have to do that on the delivery app. So within the course of a couple of weeks, they were able to create a powerful integration to do that for them automatically with the food service delivery plan.

Justin Callison: And building on what Bill is saying there, in a couple of weeks. This is the sort of thing that people are able to do because they can take advantage of these capabilities that are available in the cloud, enables agility that just would’ve been unthinkable before.

Jason Andersen: Gotcha.

Bill Boora: So that’s one of the ones where the innovation really came out. One of my favorite ones, and I’m Canadian and I fly Air Canada a lot. So about a year or two ago, Air Canada completely re-envisioned their Aeroplan app. So Aeroplan is where traditionally a passenger could go in and you could redeem your rewards miles. They moved from an off-the-shelf type solution to a microservice-based architecture using event-driven architectures and serverless primitives, to create an application where I a consumer can now go in, I can check out my flights, I can book flights, I can book hotels, I can redeem rewards points as well. They’ve integrated that with their reservation systems, so it’s a full solution. And now they’ve started to extend that using the composability of the solution to other partners. So if I go to a gas station and fill up, I can use my points there, I can use, if I use my bank card, I get those points. So really, they’ve built this ecosystem within the application using these composable modules.

Jason Andersen: And you’re not doing that in a private data center, right? I mean, it’s just too hard, really expensive to do. So the cloud really opens that up to connect to other companies

Bill Boora: Yeah, and taking that one step further. So coming out of the pandemic. And in Canada, a lot of the carriers were taking heat because bags were being lost, things were being damaged, people were complaining, they resorted to buying air tags and putting them everywhere. And so Air Canada, again Air Canada, they used event-driven architectures and our integration services to create a baggage tracking app. So now when you check in, your bag is scanned, and then they give updates at when it’s put on the carousel, when it gets on the plane, when it’s offloaded. And what was interesting about that approach is that they had to work with airports around the world where there’s already Baggage events that are happening, and I think it’s between 10 and 15 events per bag times, hundreds of thousands of bags.

Jason Andersen: Oh, wow.

Bill Boora: So in their system, they ingest the data, they take a look at what they care about, they give updates on their mobile app as well as their website. And from a modular perspective, they were able to then expand that to, you could subscribe for push notifications. So again, an example of where these integration services allow you to focus on a business problem, solve that and rapidly get that to market, and then start to scale up as you get different use cases coming in and build that up.

Jason Andersen: For sure.

Justin Callison: So I’m going to share one of my favorite customer stories, which is Nationwide Children’s Hospital. So Nationwide Children’s Hospital is one of the largest pediatric hospitals in the US and they have a really interesting partnership with it. They have a genomics institute as well, where they do a lot of genomics research, and so they end up treating sick kids. And so the genomics research, and they match the research with the clinical application, what they do is a lot of work where they can look at the, they have a sick kid that comes in fighting cancer often, and they can get blood samples and they can start looking at this child’s genomics and understand exactly the nature of the disease that they’re facing, and then they can use that in order to provide targeted treatment.

So they’ve been working on this for a while in this great partnership between the hospital and the research. And as you can imagine, a lot of medical expertise, a lot of research expertise and a lot of data, and they wanted to make this move faster, because this would take, it was in the order of weeks from getting a sample until getting it in the hands of the practitioner. So they looked to moving to the cloud to do this, but they also took an integration first perspective and they built it in an event-driven way with these services, step functions, and EventBridge. And what that allowed them to do was to take these processes that the researchers were using, model of them using step functions, and then have those tied together with EventBridge, so that would have one and the other.

Then as they went to change and evolve these, they were able to expand it more easily because they were taking this event-driven approach and using all these services across AWS. In the end, they got this thing that was in the order of weeks down to the order of a day. And so within a day, the sickest kids could have information in the hands of doctors to make sure that they get the best treatment. So it’s a really amazing outcome to see how applying this approach for integration can drive real world benefit.

Jason Andersen: That’s great, great. So Justin, Bill, thank you for your time.

Bill Boora: Thank you.

Jason Andersen: Thank you for the great customer stories as well. I appreciate it. Thank you all for joining us at this episode for Six Five On The Road, at AWS re:Invent in Las Vegas. This is Jason Andersen, and if you liked the video, please click on like, subscribe, share with your friends, and tune into our website, SixFiveMedia.com. Thanks.