Cloud AWS re:Invent Coverage | John Purcell, DoIT

Clips
News
More from Cloud AWS re:Invent Coverage

John Purcell

Chief Product Officer

DoiT International

DoiT International tackles cloud efficiency gaps with intent-driven insights

As businesses adopt cloud optimization strategies, the focus on efficiency and aligning workloads with business goals grows sharper....

play_circle_outline Illusion of efficiency in cloud environments

play_circle_outline Six pillars of cloud: operational excellence, security, reliability, efficiency, cost optimization, sustainability

play_circle_outline Need for human input in conjunction with AI technology

play_circle_outline Importance of combining technology and human expertise

Info
Transcript

John Purcell, DoIT

John Purcell

Chief Product Officer DoiT International

DoiT focuses on optimizing cloud operations in areas such as operational excellence, security, reliability, efficiency, cost optimization, and sustainability. When companies omit one of these, they may struggle to achieve desired outcomes. By combining technology and human expertise, DoiT aims to provide effective solutions. An example is working with a financial services company experiencing issues with their caching cluster. Despite traditional metrics showing optimal utilization, the cache hit rate was low, leading to increased costs. By investigating the ... Read more

explore Keep Exploring

What is the primary purpose of DoiT in business and how do they ensure that companies are effectively harnessing the cloud to support and drive their business growth? add

What are the six areas of concern that companies have identified as important to address in relation to running workloads and businesses in the cloud? add

What is the approach of DoiT towards combining technology and human experience in their work? add

What is the key to building a successful business through deploying in the public cloud, and how can technology be utilized effectively for data gathering and pattern identification in a cloud environment? add

bolt Powered by CUBE AI

John Purcell, DoIT

search

Bob Laliberte

>> Hello, I'm Bob Laliberte, principal analyst of the CUBEResearch, and welcome to our Boston Studio for a special CUBE conversation, which is part of our exclusive CUBE coverage of AWS's annual conference re:Invent. Now, one of the things we're exploring is how to drive real efficiency in cloud environments, and this is really important given the increased use of cloud and multi-cloud environments. So our guest, John Purcell, Chief Product Officer, DoiT, will help us understand this illusion of efficiency challenge that so many organizations face in their cloud environments and how to overcome that by leveraging both technology and human expertise. So welcome, John.

John Purcell

>> Thanks, Bob. Great to be here.

Bob Laliberte

>> Yeah, thanks for joining me today. I think we're going to have a great conversation. So let's get started I think by just looking at this challenge that we talked about, this illusion of efficiency. What are you seeing today in the cloud and what is this illusion of efficiency?

John Purcell

>> Yeah. So DoiT's primary purpose in business is to work with companies to ensure that they're effectively harnessing the cloud in the best way possible, as completely as possibly to support and drive their business and their growth in that business. And so that breaks down into a variety of different practices as you can imagine. But I think to your question, a lot of cloud operators will explore that environment to determine, "Are we operating optimally? Are we operating how we would define efficiently?" And we come across every day this... Which is why we coined the term the illusion of efficiency. If you rely on classic traditional observability systems, observability platforms, they will tell you at a sort of a core metrics basis, "Am I using the technology I've deployed as effectively or as efficiently as I might be?" What they won't tell you is, "Is that infrastructure, are those workloads performing the way we had intended them to operate?" So by understanding the intent of the workload as you're deploying it, then you can use the data that's being fed out of systems to really determine whether you are actually operating efficiently or if it's just an illusion as we would call it.

Bob Laliberte

>> Got it. Yeah, that's fascinating. And it seems like for the most part, organizations are operating continuously under that illusion. They're just leveraging and maybe blindly trusting the tools that they have to tell them how efficiently they're running.

John Purcell

>> Yeah, absolutely. I mean, the traditional questions we ask ourselves are, "Are we ready to scale? Do we have the right infrastructure type deployed? Is our choice of technology supporting our scale, ultimately at the end of the day supporting the experience we're trying to deliver to our customers?" And so that's what we're really trying to help them harness.

Bob Laliberte

>> Absolutely. That sounds great. And I know from our prior conversations that... You talk about the six pillars of the cloud. Maybe you could also explain what those are and why it's important to address all of them.

John Purcell

>> Sure. These are sort of six areas that I think really since the cloud sort of really came into its own and started to become a very relevant place to run workloads, to run your business, these are sort of the areas of concern that companies have realized we need to be paying attention to and doing work to improve. So whether it's operational excellence, the way we deploy workloads, develop and deploy workloads to the cloud, whether it's the security posture, once we're there, the reliability, "Are we running? What's our uptime?" For example. Whether it's the efficiency of that operations as we talked about earlier. "Is my traffic passing the way I expect it to? Are we truly set up to operate efficiently?" Whether it's cost optimization, very, very popular and important topic these days. And ultimately then sustainability, which is, of course, a growing concern for many, many companies. So the cloud operators, the hyperscalers themselves have sort of gotten their arms around these as areas of concern, whether we call it the well-architected framework, but in general, more generally, I would say these are the traditional areas that when we work with our customers, they're coming to us with questions, we're engaging with our technology across all six of those areas of concern.

Bob Laliberte

>> Got it. And what do you typically find when an organization omits one of those areas, if they're looking at four out of six, three out of six? I guess that would be a good question. How many times do you find organizations are actually looking at all six? Or is there any common starting ground where they'll start and focus on one or two areas?

John Purcell

>> When we engage with our customers, they're typically maybe more dialed in in one area than the other because that's where they perceive an area of need or that's where they perceive they need assistance. So by deploying our technology, we can sort of focus on all six, but when we're sort of fusing together this idea of machine and human intelligence to really hunt for that true underlying efficiency, if we get past that sort of the illusory effect of efficiency, what we find is that companies tend to be, let's say, more aware of gaps or vulnerabilities in one area versus another. But there's generally we find work to do in all six, quite frankly.

Bob Laliberte

>> Absolutely. Yeah, that makes sense. That makes sense. So one of the things that I find fascinating in talking to you and learning about what you're doing is that there's so much emphasis on technology these days, and especially AI. You can't turn on the news without hearing about AI these days. However, what I found really fascinating is that you believe and DoiT believes it's a combination of both, that technology and human experience. And so I think that's something I find really fascinating. I've always thought, as part of those feedback loops, it's really important to get the wisdom of the experienced people, the resources today, because in 20 years everyone will be relying on AI and not understand the intricacies and the details that maybe go into a lot of these technologies. So I'm wondering if you could shed a little bit more light on how DoiT infuses the technology and that human intelligence and expertise to add value to your offering.

John Purcell

>> So we are technologists at heart. We are a technology company and have been for over a decade. And so we've built up really thousands of years of cumulative experience in the world of cloud operations. How do you build a company or build a business by deploying in the public cloud, whichever cloud is relevant to you? And so we've built up this deep contextual knowledge of to do that well and where to look if something maybe goes wrong. And so our technology as a platform is a very efficient way to gather data, to look for the things you know to look for, maybe even when we're talking about large data sets, for example, billing data that's running through a cloud environment, operational or metrics data that are flowing out of a cloud environment. Technology quite often is really the only effective way to hunt for and identify patterns that might be telling you something or indicating something to you. So of course, our story starts with deploying technology in the right way and then leveraging it to kind of surface insights we believe are relevant. But we have over 30,000 customer interactions per year with the human component of our solution. And quite often they start with the same core question, which is, "What were you trying to achieve when you deployed this workload? What is its purpose in this whole cloud environment, in your business at the end of the day?" And when you take the story the data appears to be telling you or the technology is telling you, and you layer in that contextual knowledge of what the intention was in deploying this, that's where you really find the insight and you can really laser focus on, "Okay, so what kind of remediation steps do we need to take here or do we need to equip you to take as the customer?" So this is how we feel. I mean, generative AI for sure, Bob, is going to help us with that technology component. And of course, we're leveraging it effectively to do so. We still feel like the technology can take you to a point and whether we're feeding the output to you and letting you decide, or you just need to phone a friend and DoiT is the company that we want you to think of.

Bob Laliberte

>> Yeah, no, that's a great point because I mean, even in a lot of the best AI environments today, they're trying to get to about 80% efficacy. I was at a show last year and someone had talked about how they went on... It was a year-long project. They created this model and so forth, and they were at 3% efficacy, which clearly states they've got a lot more runway to go there. They were encouraged by the fact they at least got started and were building in. But when you see some of the more sophisticated ones, the very mature ones at 80%, and they're trying to push up into 90, and you're seeing things like the Agentic AI and RAG and things like that, trying to increase that efficacy and make them a little bit smarter, it's still, as you said, it points to the need for some human relevance there and also that human experience to be able to help drive and improve the technology. So I'm curious, with all the experience that you and your company has, how has that impacted the technology that you are delivering? I know you mentioned generative AI and AVA I think is one of those solutions and so forth.

John Purcell

>> Correct, yeah.

Bob Laliberte

>> So I'm curious as to what that internal feedback loop looks like at DoiT for how you improve the effectiveness of the technology that you're helping to deploy to your customer.

John Purcell

>> Yeah, absolutely. So AVA is the name we ascribe to our generative AI-based technology. And look, these are inference models at the end of the day. And so we believe the advantage we have in our ability to leverage this and increasing that efficacy score that you mentioned is simply because when we work with a customer, we are gathering and processing so much information about the environment that you've deployed, which is essentially what you've asked us to do. And then the model can look for connections. It can form inferences based on, "A thing happened. What could that thing have implied? What may happen?" So there's a predictive component to what the technology can tell you. Obviously we have an ongoing research investment into how is the evolution of GenAI, if I could put it that way, best harnessed in a way to really deliver on this core mission that DoiT is sort of pursuing here? And so what we're finding to this day is there is still a point at which the next step is guesswork. The next step is... And whether that's intelligent inference or that's truly a guess, it is still... It's implied. So our customers consistently rely on being able to talk through, "Jeez, we deployed this tech. It looks like it's well architected. It looks like it's performing well, the resources look like they're adequately consumed, and yet it's still not scaling effectively." What were you trying to achieve? What does success look like here for you? And to us, we have not been able to make that connection directly. We have not seen the market be able to make that connection directly purely on the basis of technology.

Bob Laliberte

>> Yeah, no, that makes a lot of sense. Obviously there's still work to do and fortunately humans need to do that work, although we are seeing more of the Agentic AI taking hold to help with that. But again, I look at those as efficiency pieces. How do we make that...? Because there is so much data that has to be gathered, there is so much that needs to be interpreted. So having the machines do more of the rote collecting, correlating, et cetera, and letting a lot of the interpretation still-

John Purcell

>> I think, Bob, that is at the core of why we think our approach to customer engagement is innovative. At the end of the day, we are leveraging technology, whether that's sort of what we would call traditional tech or whether it's generative AI enhanced technology, and we're leveraging that to do a lot of the work that we feel just tech should handle for you. It should just be behind the scenes, it should be doing the work and then surfacing to you things that you tell it are relevant so that if and when a human-to-human interaction is required, that A, we starting from a much further progressed stage. We don't have to do all of the manual, menial sort of troubleshooting tasks. We can get right to the heart of the issue and really get you to efficiency on a much shorter timescale.

Bob Laliberte

>> So speaking of that, can you share any customer examples of where you were able to go in and have an impact and deliver better outcomes for your customer?

John Purcell

>> Yeah. We've been working with our customers now... We have over 4,000 customers in our base we're fortunate enough to call our customers. We've been building this business now for over 12 years. And so these are the customers that contribute to that 30,000 interaction number I described to you each year. And so this is exactly the model of the engagement we have with all of them or with most of them. One of our favorite examples that we use to kind of illustrate this blended innovative service delivery sort of concept, it's a simplistic example, but I think it illustrates it quite nicely. So if the audience is familiar at all with the idea of caching as a technical sort of construct, where instead of retrieving an object in cold storage or warm storage, we can actually just look it up in memory. It's a quicker, cheaper way to retrieve something that's commonly retrieved. So we worked with a financial services company that had deployed in an environment, had deployed infrastructure essentially that had caching as one of its core components. And they came to us and said, "There's something going wrong here. It's not scaling effectively. Our costs are going out of control and we can't figure out why." So one of the core things we discovered when we worked with them, again, by understanding, "What was your intention here?" You had deployed a caching cluster that by all of the traditional observability metrics, that cluster was being "well utilized." It just looked like it was doing what you asked it to. But when we really got under the hood there, what we noticed was the cache hit rate was quite low. So although your cache was full, so to speak, meaning your resources were fully utilized, you were still hitting the database 70, 80% of the time when you went to retrieve an object. So now you're paying for a cache cluster at full capacity and you're still paying for that database lookup, so this concept of a double whammy, right?

Bob Laliberte

>> Yeah.

John Purcell

>> So we were able to work with them to sort of unwind or fine tune a little bit some of the business logic associated with what's cached, what isn't. Now, we didn't necessarily do that work for them. It's really just understanding, "Here's what we can identify for you to restore that path to efficiency." And again, the technology identifies, generates the metrics to tell you one part of the story. Humans are understanding the second part, and that's what that path to efficiency for that particular customer looked like. Very effective.

Bob Laliberte

>> That's awesome. That sounds great. Fortunately, that's all the time we have. So I want to thank you very much for joining me today.

John Purcell

>> Thank you, Bob. It's been a pleasure.

Bob Laliberte

>> Yeah, absolutely. And thank you all for watching this special CUBE conversation from our Boston studio. Again, part of our exclusive CUBE coverage of AWS's annual conference, re:Invent. For more information on DoiT and its innovative approach to cloud efficiency, please make sure you check them out at their booth on the show floor.