The Invisibility Threshold: Enabling Innovation with Seamless Infrastructure

In this episode of Unleashing Genius, Gabie Boko welcomes Jesse Evans, Principal Solutions Architect at Infor, a global leader in business cloud software specialized by industry, to discuss what happens when infrastructure stops working seamlessly in the background and starts affecting day-to-day operations.

As organizations scale healthcare workloads, deploy cloud-native applications, and manage AI workloads, the need for consistency and reliability becomes paramount. Drawing on Infor's healthcare applications running on Amazon EKS, the conversation examines how data access, storage efficiency, and infrastructure design directly influence performance and user experience.

The discussion focuses on what it takes to build environments that teams can rely on without hesitation. As AI becomes more deeply integrated into healthcare workflows, clean, consistently available data is becoming the foundation for innovation and better business outcomes.

Transcript

SPEAKER_00 0:06

Welcome back to Unleashing Genius. I'm Gaby Boko, the Chief Marketing Officer at NetUp, and this series is a set of executive conversations about what's changing in industries as AI and data and shifting business demands really reshape how organizations are running. In each of the episodes, uh, we're looking at where the pressure is building, what assumptions are we breaking, and what it takes to respond in a way that really holds up over time. A lot of people talk about what's possible. And uh this series is really about the shift from possibility to execution, what it takes to turn new ideas into systems and workflows and anything that actually helps us work and perform at scale. So, welcome to Unleashing Genius again. For a long time, to set the stage for this one, application and data teams were expected to work around infrastructure. Um, if it behaved as it was expected, then they moved forward. If it didn't, then they adjusted their workflows or they might have waited for it to get resolved. But today, the expectations are decidedly different. And the goal is really for data infrastructure to be invisible in some teams, or so seamless and reliable that teams don't really ever have to think about it, manage it, or work around it anymore. So as organizations are moving those critical workloads to the cloud or especially complex healthcare applications, generating high volumes of small files, maintaining that invisibility layer is starting to get a little harder. Um but this creates a different kind of threshold. Not whether or not the infrastructure for data is working, but whether teams are gonna trust that infrastructure, trust that it's gonna stay out of the way without having to think about it. And you know who my guest is today. My guest today is Jesse Evans, the principal solutions architect at Infor. Because Infor was faced exactly this same type of situation during their healthcare cloud migration. Jesse, welcomed Unleashing Genius.

SPEAKER_01 2:10

Thank you so much for having me. And you pretty much frame that up perfectly for our initial journey into the cloud. Yes, ma'am.

SPEAKER_00 2:18

Do me a favor, Jesse, just really quick, give me a 30-second snapshot. Who is Infor?

SPEAKER_01 2:23

Right. So Infor is a it is a actually very large software development, privately held software development company. Um so I'll speak specifically in my clinical vertical. We really focus on clinical message routing. So this is this is really around the message translation layer. So things that I think a lot of people don't really think about, especially with healthcare, is those HL7 and fire messages, things like that. So that is really, it's not where we have ERPs and we have financial um solutions. We have pretty much anything you could possibly think at it in for. Um we have solutions for it. Uh personally, I I'm in that clinical health care vertical and it's it's interesting, it's challenging, and I feel like um, you know, can actually make a difference. So it's exciting for me as well.

SPEAKER_00 3:11

I love that. Uh it's always fun when the job is um personal and you feel like connected to it and it and you feel like you're doing something important. So thank you so much for describing Infor. So let's get into kind of what I set up at the beginning. Um I set up this kind this idea of invisibility of data infrastructure, right? Um, let's start there. What do you think fundamentally changes when the workloads start to push data infrastructure in ways that start to test that reliability you need every single day?

SPEAKER_01 3:44

You know, so what really fundamentally I think changes here is the infrastructure stops being this background assumption, and it actually has to be part of the product conversation. And in our case, you know, our workload was not simply this huge data volume. You know, this wasn't data late. This is very, very high frequency, you know, small, maybe 4K uh messages at times, highly transactional and fan out, you know, very one to many. We might have one transaction coming in and it fans out to 30. And these are all very small, very, you know, extremely sensitive, sensitive to latency. So we actually we we end up dealing with millions to tens of millions of these type of messages a day. And what ends up happening here is that this kind of workload really exposes the inconsistencies very quickly, you know, at that infrastructure layer. So if that storage layer is behaving differently in the cloud, any managed service, we're going to find out about that pretty quickly. And what ends up happening, we end up discovering that by downstream systems having impact. For instance, we have uh ADT messages coming in. Um, so that might be you know admit, discharge, transfer pages coming in. And if if those aren't being you know routed properly, you know it, right? You can't start assigning beds to patients and things like that. So really um that visibility threshold was this inflection point here at in product and architecture and a development team where infrastructure, you know, our infrastructure can really no longer the infrastructure's not successful because it's just out there. It really needed to meet that guaranteed consistency that we require for this type of highly transactional data.

SPEAKER_00 5:38

And let's go back just like a hot second because you you mentioned kind of what you were trying to accomplish. This this cloud migration that you're talking about, especially in healthcare that you were undergoing, what was your primary goal? Like what were you trying to accomplish with it from the outset? Because the way you've just described what you were trying to do with it really started to say, I need, you know, the end result is how do I take these small messages, which could magnify into larger messages in terms of volume? But you also said something really interesting is at the end of the day, I need to be able to assign beds to patients. What was the goal of what you were trying to accomplish and why was the cloud so important in this?

SPEAKER_01 6:20

Right. So honestly, um, one of the biggest goals was to get infrastructure out of the product roadmap. So getting into managed services in the cloud, getting to a point where myself, to be frank, myself as an architect and as a product director now, really don't I don't have to think about that infrastructure layer. I I really need, you know, myself, my team focused on our clinical, you know, domain, our business logic on to make value-based care even better. That was really the biggest driver there is you know, let's take this off the plates of our users. They don't, and that's reality, right? Like we heard this from our customers too. It it costs a lot of money to host all of the hardware, the infrastructure, everything on-prem, and then they have to have all of those experts there, all the infrastructure experts, right? So moving it into the cloud, embracing these managed services, it was really a way for us to say to our customers, you know, we can do better. We can provide, they loved it the way it was, but we can even provide better and better solutions for our end users, provide better clinical ROI, things like that. That's really was my driver.

SPEAKER_00 7:32

Yeah, so that means that you're you were really looking at the workloads. You were looking at the day-to-days and saying, this actually looks like it might be better affected over here. How do you think your peers, just to kind of break out of you a little bit, how do you think your peers responded to that when you were examining these workflow, these workloads, these these kind of application areas that you needed to change and address and thought you could do better, did it push infrastructure in ways that maybe others weren't expecting it to be affected?

SPEAKER_01 8:06

Yeah, let's I take one little step back. Yeah, so it absolutely impacted, you know, the team as a whole, um, the customers, even the industry, really. It's it's very interesting because they when we get out of ERP and we get into this type of clinical transactional interoperability where we really are truly connecting multiple systems, you know, it's many to many. We're anything from finance to you know admissions, and that needs to go out to maybe 15 different places, right? So I I do want to mention that you know, going to the cloud isn't always what everybody thinks it's gonna be, right? Like, so there's not an easy button, right? There's no easy button that you click and manage services just work. That's not quite how it works. So we really, you know, we really had to focus on what the goals were and what were we really trying to do for our customers and the solutions. So whenever we get into you know this migration and now we're into the cloud, there was things where the infrastructure where it was, you know, on paper, it's it's expected to work. That's really where we saw and we started identifying in other team members. You know, you start to identify other things kind of bubbling up, like, hey, there's better ways to address this, there's better ways to kind of utilize these managed services and these other teams like NetApp that can fit these, you know, kind of fit these domain logic, these specific infrastructure places that we need so we really can focus. So it affected a lot of folks actually, um, and not negatively, it really is allowing them to focus on on their core um business logic and the and their goals.

SPEAKER_00 9:51

You know, it's really interesting when you say that because what you're talking about is that the infrastructure is actually starting to become more visible. Um which is honestly, I think applications, enterprise applications have always had the luxury of being the most visible inside an organization. And I think now when you're talking about infrastructure and kind of data infrastructure, it's it's really solving the data problems. So and bringing that forward, which is why it's so important to focus on those outcomes, like you just said. So let's go to the next kind of thinking area for you. When you're focused on those outcomes and and that data infrastructure is kind of coming out of the background and becoming just as important as anything else in your stack because of those outcomes, what's the first sign in that project line that something has kind of started to shift? Right. What what's the what's the first indication that that you're on the right track or maybe you need to adjust?

SPEAKER_01 10:48

So I think that it's usually not a some dramatic outage. Um, it's a hesitation, right? It's if for me and for us, it it was this, you know, environments technically running, but that behavior wasn't as consistent as I saw on-prem, and it was not consistent enough for our customers. Um, you know, there's there's some very great services out there, file-based services, you know, that we've used in many use cases. You know, the workload patterns are completely different. And that's really what it came down to was that workload pattern. Um, you know, with writing these millions and millions, tens of millions of small files across AZs, round trips, you know, through availability zones and regions, it introduced so much latency that instead of that guaranteed consistency, it was eventual consistency. Um and our engineers really had to, you know, that that honestly, like that is whenever the infrastructure became visible, right? That is truly whenever you start to realize that, okay, that yes, this works. But are now our engineers really have to start thinking about that storage behavior before green lighting any GA, right? We have to have very, very detailed um processes to make sure that we can scale this up and above. We actually are following our customers' you know, workloads, you know, and in clinical, it's early morning, evening. It's very it can be very unpredictable. So that that's really where we were. You know, we had to ask our um whether the throughput would hold, you know, whether we uh have an embedded database would stay responsive and the architecture could actually meet those performance expectations of our customers. And I want to say, like, those file systems weren't bad. It just was not the right fit for our embedded database and that high frequency small right. We really needed to get into that block level.

SPEAKER_00 12:38

I, you know, I think that's really relevant because what you're suggesting is, and correct me if I'm wrong, you're suggesting that when you're addressing those issues for building that foundation, you're also addressing the process issues. Like you're not dragging the legacy process along with you. You're actually saying, I'm gonna just I'm gonna disrupt and stop this because the foundation needs to be this and it's different. How do you how do you remain true to that that future state of building the right kind of foundation and at the same time understand that when do you decide to maybe stop legacy thinking on process and say, you know what, this deserves a different kind of approach as I rebuild this foundation based on the behavior and the outcomes that I'm trying to achieve?

SPEAKER_01 13:21

So that is an excellent question. And honestly, I I think it's a learning lesson for a lot of us. Um, it is really something that you have to go in open-minded and be ready to kind of pivot early. For for lack of a better term, if you anchor it in that legacy technology and this has worked, right? This has worked for so long, and and our customers love this, you get in that mindset of okay, then we need to stick with this. Yeah. But you know, as the industry changes and technology changes and services changes, and it it even the use cases, you know, the way that our our clover leaf and this message routing solutions have changed over the past few decades, it we really it's important to stop and take that pause and really kind of re rethink the architecture. And that's not always a bad thing. And I think that's something we need to point out. You know, a lot of times whenever we we get into these type of discussions, um, folks will look at that negatively because they see like something was done wrong and it's not, right? It's it's just a matter of is there a place for improvement for improvement? And we definitely found that here.

SPEAKER_00 14:29

So let's let's make this a little bit more personal then. Um we've talked about it from a process and a behavior and an outcome. What do you think it takes from a leadership perspective, right? So how do maybe you lead in that? How do how do you help leaders around you lead in that? Because what you're essentially saying is that there's uh so many things at stake, and you're the change management could be dense and and tricky. So what does it take from leadership to change that perspective on on what that foundation, that data infrastructure foundation could be?

SPEAKER_01 15:04

Uh one of the first things, honestly, is patience, right? I think patience is is crucial with leadership. And they, you know, they have to be willing to allow the product and engineering the time to do what's best. And if that means that you know, we go in with an architecture that you know everybody truly thinks that this is the correct direction, even if it is somewhat still carrying some legacy, and you know, we we see that we now have this infrastructure that is highly visible that you know, product and engineering have to think about all the time. Um, from that leadership point of view, that is really where the leaders need to come in and say, hey, it's okay to pause. It is okay to stop and think about what we need to do correctly. Um, not that it was done incorrectly in the past, but let's look at what can be done and give the teams, you know, the the pathway to do it correctly. Honestly, I think that's one of leadership's biggest biggest accomplishments, right? Is empowering the people to innovate and not just have to constantly rely on that legacy standards to keep going.

SPEAKER_00 16:13

So it's technical leadership, but that you'd have to know what you want to change, how to change it, and at the same time, it's also personal and and and process-based.

SPEAKER_01 16:23

Very much so.

SPEAKER_00 16:24

So let's talk about that. You've got a foundation now. It's amazing. I'm just gonna say that it's amazing for you. Um, and that you did a great job of change management. But what are you thinking about the next, right? And the next is always, you know, that this question, the two little letters, AI, right? So how do you think that this infrastructure and and kind of what you went through is gonna evolve um as organizations? It could be yours, could be other organizations. Feel free to comment on both. But how does that start to affect this this new way of thinking on on infrastructure and your data that you've created for these environments?

SPEAKER_01 17:02

Right. So I think AI really it raises the bar even further. So uh I'll just kind of I'll speak specifically with my clinical healthcare, you know, industry right now. It is it is so important that it is consistent, right? And it's stable and it's available. When you add AI into that, the the number of interactions with that data makes even more inconsistencies show up faster. Um so AI, obviously, you know, if there's inconsistencies in your data or the the actual rights of your data, if if it's slow to write and AI tries to make a call, you you might miss that. So the context is completely off. So, really, in in healthcare, AI-driven workflows really depend on that clean, consistent available data. You know, it is what's coming out of that inference engine is only as good as what's going into that. Um, so we need, we very much need that clean, that quality data, and we need a consistent path to get it there, to guarantee that consistency. So whether we're talking about the message routing, transformation, you know, fire normalization, uh, operational automation, AI layer really, it's only as reliable as the data foundation underneath it. And that that data persistence layer that NetApp was able to provide us with FSX OnTAP literally allows us to continue to not have to stress about is that data layer, is that persistence layer uh solid? Is it consistent? Absolutely. Um, it is above and beyond anything I could have hoped for. And the teams agree, everything has been exceptional. Um so it what's great about this is I really have such a wonderful trust in the product, in the service, in the data quality, the consistency, the availability, that that variability is not such a scary thing for me now that we are jumping into AI. And and we are. Um, you know, we we've actually just, I think last fall released our first uh MCP AI server for cloverleaf. Yeah, so it's exciting, and it's but it really does boil down to I think if you don't have that consistency to start with, you're not going to have a good experience whenever you start to introduce the AI. It's it's it's a learning process.

SPEAKER_00 19:25

AI is nothing without the right data, and it's nothing without clean data. You're absolutely right. You know, this title of this podcast is Unleashing Genius, and and the thinking is that once you get certain layers of your data infrastructure leaning into the intelligence that it can bring, that it empowers the genius that comes through with the human part. What are you most excited about, the human outcome that you can drive next?

SPEAKER_01 19:51

Yeah, I was just gonna say so what's what excites me so much about AI and especially clinical health care is I I am really starting to see a short path to increasing our care, our our actual care, right? So the way that you and I interact with clinicians, the way that you know value-based care and and if you have to do prior authorizations, anything like that. Like, you know, I I have um personal experience with with family members you know with with chronic diseases, that they spend so much time trying to get prior off and trying to get you know these things figured out. AI is going to be able to close gaps in care with fire. That's something I'm very passionate about. So now, you know, taking that gaps in care where I have a feedback loop to say, hey, we identified the gap in the data. We notified the clinician, and the clinician was able to close that gap. And now we updated our fire um living persisting out of net uh NetApp. Now I can really use AI to even start in making more inferences, readmissions risk rating, things like that. And that is all, I think, just going to drive value-based care in the right direction. I really see healthcare getting better and better because of what we're doing. And it's exciting.

SPEAKER_00 21:16

That is um that's the best kind of outcome, right? Uh making life better, right?

SPEAKER_01 21:22

A little better, right? Incrementally better for each generation is what I'm shooting for.

SPEAKER_00 21:26

You gotta leave it better than you found it, right?

SPEAKER_01 21:28

Yes, ma'am.

SPEAKER_00 21:29

Thank you, Jesse, because this conversation just really highlights that as you've called it out, the standards have changed, and that the the infrastructure and the data infrastructure or the data behind it, that those basic functions just aren't good enough, maybe in some of the legacy processes, that the behaviors and the data infrastructure has to change and to sync into a consistent and connected and integrated operating system so that it's making your data reliable and stable, and that the infrastructure is operating at a level where your teams don't even know that it's there, but they are confident, they create a level of confidence. And and I think that that's what turns your ambitious, I love the your ambitious plans to modernize that and and those new capabilities on changing uh patient care and outcomes for the healthcare. I think that that's the kind of genius that we want to change the world with. So, Jesse, Jesse, thank you. Thank you so much for being here and talking to us about this amazing work that you're doing, and uh keep doing it. We're really proud of you.

SPEAKER_01 22:34

Thank you for having me. It's very exciting. Thank you so much.

SPEAKER_00 22:37

And thank you all for turning into unleashing genius. I'm Gaby Boko, the CMO of NetUp. Stay tuned for our next one. You'll never know what kind of genius is around the corner.