Ep#150 Cloud Migration Challenges & Success with Joe Conway

Host: Jon

Hi everybody and welcome to the Jon Myer podcast. I'm your host, Jon Myer. Today's topic is cloud migration challenges and success, and our guest today is Joe Conway, who's the Chief Technology Officer at ADDX. Joe has spent his career solving data management and software engineering challenges for public sector customers. Please join me in welcoming Joe to the show. Joe, thanks for joining me.

Guest: Joe

Hey, Jon, great to be with you.

Host: Jon

All right, Joe, our main topic is cloud migration, and I have to tell you that I've been through several cloud migrations and I think they're still going on. Let's talk about some of those challenges is

Guest: Joe

Yeah, us too. So I heard somebody say recently that cloud migration is a journey that never ends. And I think that's pretty close to being true. Unfortunately, I think it's a disservice to our customers to say that primarily because cloud migrations are a means to modernization, but it's a project that has to have a beginning and an end and a cost associated with it and a schedule and risks associated with it. Someone, someone likened, the cloud migration to a journey. And you think about it, when in your life have you ever taken a journey where you didn't have an itinerary? So if you and I said, Hey, let's go to Paris, you would normally say, okay, well when are we leaving? How are we getting there? Are we going on by plane once we're there, what are we doing? Are we going to the Eiffel Tower or the Louvre? Where are we staying? Are we going by bus or Uber or whatever? When are we coming home and oh yeah, how much is this going to cost me? Right? And if you can't answer those questions, the journey becomes undoable, right? Unmanageable. And I think that's where cloud migrations have ended up. Frankly. I

Host: Jon

Agree with you because with the cloud migrations, how many of them have got to the 80% and stopped, never able to achieve the final one? Roles change, responsibility. People who are managing the project are no longer there, and engineers change. You can't get over the 20%. They're either too complicated, you don't know how to modernize the application. And you know what? This project has gone long enough and is a hundred times over budget and you're like, screw it, I'm out of here.

Guest: Joe

And customers of all shapes and sizes are concerned about it. Flexera publishes an annual state of the cloud report, and their most recent one, which just came out in February about is a survey of customers of small, medium, and large US and international, and cloud migrations was a top concern for two out of three customers. And you would think, well, we've been doing this for quite a while now. Why is it still a concern? And I think the answer is that it is this journey that people go on and they had no idea what the cost or the schedule or the required skillset or who makes the decisions along the way would be. I think that's probably why it's still a major concern.

Host: Jon

So cloud migration, there are several challenges going into it. Cost is always one. How are we getting there, completing this journey? What are some of the challenges that you see from your customers?

Guest: Joe

Yeah, so I was talking to one of our customers, a lieutenant colonel the other day, and if you spend any time around military folks, they can be very direct. And he looked at me and said, okay, so for this cloud migration, when are we starting? I have an operational system and we're going to move this thing to the cloud and I want it to operate after it gets there. So when will it be back in operation in the cloud? How much is it going to cost? What's the schedule and what are my risks? And what happens if some of those risks come to fruition? If you can't answer those questions for me, we have nothing to talk about and I have other things to do that that sort of candid assessment from Lieutenant Colonel is kind of where most customers are. I think Jon, there's this sense that I'm signing up for something and it's somewhat unbounded. It's more art than science. And companies will come in promising migrations and they take too long and they cost too much. And there's an army of people involved in making it happen. And sometimes they just run out of time and patience and money and they say, okay, fine, it's at 80%. Great, move on. I think that's where a lot of folks are at.

Host: Jon

One of my biggest challenges with cloud migrations is that they take too long. We've got a 2, 5, 10-year plan to get there. We've been doing this for a while, does it take that long or can we break this up into chunks and tackle this in tandem to get it done in parallel to get the migration done? And then the other pet peeve I have is the cost. You're thinking, oh, you know what? Okay, two years I can live with two years. We've got a lot of business applications that do it and it's very hard to come up with an exact cost and relatively, is it ever even close to what it does cost? Yeah,

Guest: Joe

Yeah. And I spend a good bit of time in the federal space and the incentive structure I think is a little bit out of whack. You have large companies that come in and they're incented to draw these things out as long as possible because they're getting paid for their time and effort. And as a taxpayer, I think that's repulsive, but that's kind of the way it works. If you get into a migration of an operational system and that thing has to be up working, you're going to throw all kinds of money at it to make sure that it happens, even if it's not planned well, even if the roles and responsibilities are identified, the decision-making authority isn't crystal clear where the documentation isn't precise. If a government customer, you'll spend more money on that because you're kind of stuck and I think there's a better way to do it.

Host: Jon

Joe, what are some of the successes that you see out there for cloud migrations? What is the approach to make it successful?

Guest: Joe

So in a word, Jon, planning and complete realistic planning.

Host: Jon

Is there ever too much planning though? I think sometimes we're having meetings to have meetings, we're planning this. Is there ever a point that you are just like, we have enough planning, we have spent a good amount, we need to get this done?

Guest: Joe

Yeah, I think where there is a good plan, a completed plan that you can execute, but there's an old saying, I think it's attributed to Abraham Lincoln, if I had six hours to chop down a tree, I'm going to spend five hours sharpening my Axe. I think that's the case here with cloud migrations. You need to take the time to complete the planning. And when I say planning, I don't just mean a list of tasks, but the resources required, the executive buy-in that's required. The amount of effort and time that's going to go into if based on a set of applications in your portfolio, identifying what those are, what's going, what's not going, what are the priorities, which one goes first, what are the interdependencies? Which one, if it moves, then affects others? All those questions that don't get asked or answered have to go into the plan.

Guest: Joe

If it doesn't, you get surprised. You go, oh crap, there's a database that nobody knew about and it's not in the migration plan. Do we have anybody here that knows how to migrate a database? Oh, okay, do that. Go figure that out. So that's where I think they happen, where the planning breaks down. There's also an old saying that no military plan survives first contact with the enemy. And I think that's kind of true with cloud migrations as well. They change as soon as you start the process because you haven't planned well. And some things happen. In my experience though, technology is usually not the problem. The technology usually works. It's the non-technical aspects of it, the people, the processes, the decision-making, the roles, responsibilities, those types of things that are really where the problems are. And frankly, I think the other thing is that some people see cloud migrations as a CIO function. It's the IT folks' job to do this well, cloud migrations affect the entire organization from the CIO and the CEO to the financial folks and the contracts folks and the users and all of those folks have a vested interest in the success. Many times they don't have a voice in the whole process. And so that's the other reason they can struggle and fail.

Host: Jon

Joe, do you feel that having a solid plan in the beginning ultimately attributes to a successful cloud migration of getting over that 20% of like a hundred percent migration?

Guest: Joe

Yeah, I do. And I've seen it by the way, just full disclosure, we've migrated ourselves. We have some of our company business applications and processes running in the cloud, and we moved them. We picked 'em up and moved one of our contracts systems into the cloud. That was our first experience. That was about five years ago. It was traumatic, I have to say. There were just all these surprises that we experienced, but we did get it done. We did get successful. It is running happily in a w s today, but it was pretty traumatic. We thought we knew what we were doing, and then reality kind of struck. We needed some additional resources we didn't know that we needed in advance. So our planning was decent but not great. And so I think it, Jon, boils down to the set of questions that have to be asked and answered. And for example, I know a W SS has this thing called cloud readiness assessment, which asks about 80 questions, and I think that's a great tool. It's a great start, but there are hundreds of questions that have to be asked on a large portfolio, including what the portfolio of applications that are moving, and failing to answer those questions means that there's something you don't know and worse, something you don't know. You don't know that you need to be successful.

Host: Jon

Hi everybody. My name's Jon Myer. You're watching the Jon Myer podcast and we're talking about cloud migration challenges and success with Joe Conway, who's the Chief Technology Officer at ADDX. Joe, what does a typical engagement or success of a cloud migration journey look like for you and your customers? What is the process behind it?

Guest: Joe

So we have developed a repeatable, complete cloud migration process that asks all those important questions. It goes from highly abstract roles and responsibilities down to a readiness assessment, which is looking at things like what are the skills that are required to pull this thing off. And we get to a portfolio. These are the applications, databases, storage, et cetera, that are going to migrate. And then we get to a detailed set of testing and migration scripts that become part of the documentation. And then finally, we have a completed migration test plan that we put in front of a customer that says, okay, here's the schedule. Here's the budget, and here are the required skills that you're going to need. Here are the decisions that have to be made. And hopefully, by that time we have the people who make the decisions and are accountable for the decisions, and we hand them the plan and say, here you go.

Guest: Joe

Here is your itinerary to Paris. This is when you're leaving, this is when you arrive. This is when your system will be operational. Again, this is how much it's going to cost you. These are all the people that are going along, if you will for the ride and how much they're going to cost. Here are your risks. Here are the things that can go wrong on this journey, and here are the ways that we can mitigate those risks. A lot of the risks revolve around skills, required skills. So for example, if you're moving a PeopleSoft application running on Oracle, you need PeopleSoft expertise and you need Oracle database expertise, and you need storage and you need networking, you need all these things. And so having that plan for it's tailored to a specific application is key to getting that done. That's what we're seeing with customers if you, most customers, at least in the public sector, they've been told to get to the cloud, right?

Guest: Joe

They're mandates. And the federal government, it's called either cloud-first or cloud smart, depending on which administration is in charge, and they've been told to get to the cloud. So it's not like they have a choice where they have to be convinced whether or not to get to the cloud. What they have to be convinced of is that if they take this journey to the cloud, their failure will not show up on the Wall Street Journal or Washington Post front page. That's really what they're concerned about because that is a career-ending kind of thing. So they want to know how they want the itinerary and they want to understand how it's going to happen, what they need, how they budget for it, and when they can report up the chain that this operational system that's probably mission critical to their business is going to be available again. And failing to answer those questions either sets them up for failure or sets the cloud companies up to take the heat because of the old finger-pointing thing, right? Who's responsible? Well, it's not me, it's somebody else that's going to happen. And so it's a pain for everybody. Everybody loses in that sense. And so we want to set up for success for our customers, for success, for our cloud partners' success.

Host: Jon

Joe, I think you touched on some of the challenges of migrating to the cloud of really having the right skills for it, identifying those skills, and having a solid migration plan. But if you don't have the skills, you're not getting there. And that's one of the failures I see when you're trying to migrate to the cloud is that you assume you have the right people in place for it. Hey, I have this expertise. They can help do it, but not identifying and make sure that they have the right skills to do the work ahead of it.

Guest: Joe

No, absolutely. And that's one of the things that comes out of our cloud migration navigator. Our process early on, just after we get to beyond the charter is what we call the readiness assessment. And it's think of it as the AWS Cloud Migration Readiness Assessment on steroids. That's really what it's, we expand to a whole bunch of areas, non-technical areas around staffing. One of the things is, do you have people who know how to budget for cloud services, which is very different from running a data center? And typically the answer is no. It's like, okay, well, that is a skill that you need to have to be successful. Do you have folks that know how to optimize the spending so you're not turning on services that you don't need very expensive services that run the bill up, but don't accomplish anything? Do you have those kinds of skills? And do you have the security folks that know how to do a cloud-based cybersecurity compliance process very different from a data center-based one? So all those skills have to be identified upfront. They have to be put into the plan, they have to be scheduled. You may not need them for the entire time, and they need to be documented and presented such that somebody can make an informed decision about whether they have the skills or when they will have them, and when they can begin the journey.

Host: Jon

I agree with you, Joe. I'm going to ask you two following questions, and these are very key to the challenges of cloud migration. We've already talked about what is challenging for them, but what does success look like to you and your customers for cloud migration?

Guest: Joe

Yeah, so again, I'll talk about the public sector in the federal government, either civilian or Department of Defense, you can't just throw something up and have it and start working on it. You have to get something called an authority to operate an A To. And there's a process that's near and dear to the heart of anybody that works in the government called the Risk Management Framework. It's based on a set of NIST security controls. And to get a system through the At O process, you have to identify the necessary controls. And when I say necessary, the NIST, the NIST 800 dash 37 I think is the publication. There are somewhere between 900 and 1200 individual controls. Okay? So just think about that for a second. Let's say I'm good and I can write the controls and test them and document them in an hour each.

Guest: Joe

Okay, that's seven to eight months just to write the plan for those security controls. That's with a human. We have a different way of doing it that's much faster with our process, but that's where the time is spent. That's where the difficulty becomes in getting through that cybersecurity compliance thing. But it's essential. If I have a government customer, which we do, we have several that have an operational system that has an ATO that has the authority to operate. They want to know, okay, when will this operational system that I have in my data center be operational again in the cloud? And that's not a technical question. That's typically a cyber process and documentation question. And sometimes the migration could take six months or a year, but the ATO process could take 18 months or two years.

Guest: Joe

That's how difficult it is. That's the lift that's currently in place. Now with some of the things that we're doing with ai. We can dramatically reduce that time. A lot of the documentation, a lot of the processing, and a lot of the security controls can be done with AI and generated in minutes, not months, and at a fraction of the cost of what is typically involved in an ATO process, not for the government. So I think there's a tremendous value that's there for our government customers. And it removes this concern that my lieutenant colonel friend expressed about when will this be operational again for me. We can give him an answer to that. And you don't have to send an army of cyber people in and pay for an army of cyber people to generate the security plan and the security controls and the tests and the, what is called a poem, plan of action and milestones that come out of that to give to somebody to say, okay, here's your risk of operating this system.

Guest: Joe

Is that acceptable or not? And put it back into operation. So we think there's a tremendous value, speed to value that our process brings to prospective customers. And by the way, we built it primarily because we do a lot of cloud work with federal government customers, but there's nothing specific about what we've done that excludes other customers. We think the process, and the questions that you ask are the same. The answers may be a little different depending if you're a commercial customer or a public sector one. But the process is the same and the process works and it's complete and it's repeatable and it's tailored. And there's an old saying that cloud migrations are like snowflakes, right? No two are alike. That's probably true. But all snowflakes are made out of water, and they only exist at a certain temperature. And so there are some similarities there as well. And I think that's where we add value, is those similarities. There are similar questions that have to be asked and answered for every cloud migration. It doesn't matter which one or who the customer is.

Host: Jon

Joe, you're saying that you're using AI to help facilitate this. Can I just do this myself and just, all right, I'm just going to write the framework, write the things myself? Yeah, you know what? I got this. I can handle this. What is the value of that

Guest: Joe

If you've worked with AI at all? There are two extremes. The extreme one is the Terminator view of the world where AI is going to take over the world. And then there's the, I'll say, a different kind of pessimism, which is basically that AI is going to put me out of a job. Maybe it won't destroy the world, but it may destroy my world. And I am on either extreme, I'm somewhere in the middle, which is AI is a tool. It's a technological advancement that is in line with what I've called the relentless march of technology. Through time, technology continues to advance. AI is nothing different. It's a tool that helps it. It's a tremendous digital assistant. We've seen some studies from folks that we've worked with that say that AI can increase the productivity of one person between 20 and 35 times. That one person can do the work of 20 or 35 people.

Guest: Joe

And when it comes to cyber stuff, where it's documentation primarily, that's not a bad thing. That's a good thing because it's complete. People are error-prone. AI can hallucinate a little bit, so you have to kind of watch out for that. But the big thing with ai, especially with generative ai, is you have to ask the right question. There's this whole new set of jobs around what's called prompt engineering. You have to ask the right question. You have to train it with the right data. If you don't, you get bad, very authoritative, but bad, incomplete, or sometimes just plain out wrong. AI is notorious for making things up and making it look like it's very real. And so you have to be able to do it. So the two things are you have to ask it the right questions, and then when you get an answer, you can't be naive and just say, oh, it came out of the air, it's the right, it's the right answer.

Guest: Joe

You have to be able to go in and customize it or correct it. A lot of AI that we're seeing somewhere between 92 and 96% is accurate, but it says 4% of inaccuracies. That could be devastating, right? So yes. Could you do it yourself? Sure. We have a set of prompts and a way of training the AI models on the particular climate migration that's happening based on the artifacts that come out of our process. And it'll get smarter over time, but again, it's about 80 to 90% complete. And then you need some people to augment the AI and make it usable for a customer. But Could you do it yourself? Yeah. I don't recommend it because you have to ask the right questions, and then you have to kind of know a little bit about the answer that you get and whether the AI is making stuff up, which sometimes it does.

Host: Jon

Joe, asking the right questions is key, and I want to thank everybody for watching and listening to our podcast about cloud migration challenges and success with Joe Conway. Joe, thank you so much for joining me.

Guest: Joe

Thanks, Jon. Pleasure to be

Host: Jon

With as always. My name is Jon Myer. You've been watching the Jon Myer podcast. Don't forget to hit that, like subscribe and notify, because guess what? We're out of here.