4#2 - Jonah Andersson - Journey to the Cloud: Cloud Migration, Edge AI, Data as a Service (Eng)

«Don’t go over to the cloud without truly understanding what you are getting into.»

Unlock the secrets of cloud migration with industry expert Jonah Andersson, a senior Azure consultant and Microsoft MVP from Sweden. Learn how to seamlessly transition your data systems to the cloud. Jonah shares her knowledge on cloud infrastructure, AI integration, and the balance between Edge AI and Cloud AI, providing a comprehensive guide to building resilient cloud systems.

Explore the intersection of IT consulting, Data Governance, and AI in cloud computing, with a specific focus on security and agile workflows. Understand the critical impact of GDPR on data management and the essential collaboration between IT consultants and data governance experts. Jonah and I delve into the growing trend of edge AI, driven by security and latency concerns, and discuss responsible AI usage, emphasizing security and privacy. Learn how to navigate the complexities of multi-cloud strategies and manage technical debt effectively within your organization.

Jonah offers tips on avoiding common migration mistakes and highlights the significance of using tools like Azure's Cloud Adoption Framework. Whether you're modernizing outdated systems, merging companies, or transitioning to a new cloud provider, this episode equips you with the essential knowledge and resources to ensure a successful and strategic cloud migration journey. Join us for a deep dive into the future of cloud computing with an industry leader.

Here are my key takeaways:

• Azure services can be tailored to use cases and service needs. But you need to understand your requirements and needs.
• Once you understand what you need to do, you need to gain perspective in the how - what methods and processes are supported?
• Think security at every step.
• Security with integrations is an important part, we need to focus more on.
• Bringing different competencies together is a vital ingredient in building resilient applications.
• Cloud is about where your data resides, how you protect it and how you handle big data.
• Cloud should support the entire data lifecycle.

Cloud and AI

• «Cloud computing is the backbone of AI.»
• AI pushed for Edge AI, in addition to cloud. Reasons for Edge AI are latency, but mainly security.
• Cloud can provide an attack surface for eg. data poisoning, lack of control for training data, etc.
• AI tools can pose concerns on what and how you are exposing data.
• Awareness and education are important, when building something with AI.
• You need to at least understand your input to track your output - explainability starts with understanding of your data sources.
• There is a risk to Model Governance by on-perm due to the level of competancy needed.

Multi-Cloud vs. Single Cloud

• This is one of the questions to consider at the beginning of a cloud migration.
• Drivers for multi cloud strategy are:
   
  • Avoiding proprietary vendor lock-in,
  • Existing applications or infrastructure in another platform,
  • Choosing according to the quality of services offered by cloud vendors.
• If you choose multi cloud for automated resource management, you need to consider support platforms.

Cloud Migration

• Reason for cloud migration boil often down to gaining resiliency in the cloud, due to redundancy.
• You need to uphold Data Quality not just after the migration but also during the transit.
• Cloud migration requires strategy.
• There are great resources to help with your cloud migration, like the Cloud Adaption framework or the Well-Architected framework.
• Use observability and orchestration tools for your migration process.
• Ensure you understand your cost, and can optimize it to fit with your needs.

Transcript

Cloud Migration Strategies in Azure

Speaker 1 0:00

This is Metadema, a holistic view on data management in the Nordics. Welcome, my name is Winfried and thanks for joining me for this episode of Metadema. Our vision is to promote data management as a profession in the Nordics, show the competencies that we have, and that is the reason I invite Nordic experts in data and information management for a talk. Welcome to the second episode of Metadata, season 4.

Speaker 1 0:41

Early in the season, I think it's good to start looking at the backbone for the entire data work and the work that we could do with data at scale, which is the cloud and the age of cloud that we have been through the last years. It's a topic that really goes into the infrastructure and we're going to talk about well, we're going to talk about Microsoft Azure, but we also want to talk about cloud computing as a concept. We're going to talk about AI. We're going to talk about edge AI versus cloud. Do we have enough resilience on the cloud infrastructure for AI, or should we look into edge AI as like an alternative, especially when we're talking about latency, about migration efforts, about security? But we're also going to talk about and this is maybe one of the most interesting topics for me personally because I've been through some of those journeys already cloud migration. So how do we get on the cloud? How do we move data in a structured manner from cloud provider to cloud provider, stuff like that.

Speaker 1 1:41

So a really interesting episode I had and I found an expert. I found an expert on the topic. I found an expert on Azure to talk with me about the topic and that is Jonah. Welcome, jonah.

Speaker 2 1:52

Yes, Hi Winfred. Thank you so much for having me in your podcast today. I'm so honored to be part of this episode. Thank you for having me.

Speaker 1 2:03

And you are truly an expert on the topic. You even wrote a book about learning Microsoft Azure and I think the book is about what? 800 pages, so that is massive knowledge on Azure. It's fantastic to get you on the podcast, so please introduce yourself, talk a bit about what you do.

Speaker 2 2:23

Yeah, sure, thank you. Yeah, that was a very positive of you to say I'm a truly, truly expert, because I'm sure there are many experts around Europe in the Nordics that does the same things they do. But I appreciate the compliment To the audience or to the listener. I'm Jonah Anderson. I am from Sweden. I speak Swedish as well, so if you're going to speak Norwegian to me, definitely we're going to converse pretty well.

Speaker 2 2:52

So I'm from Sweden, in a city near Hudiksvall, so I live outside the city, but it's in Mid-Norland, we call it and I am working as a senior Azure consultant at an IT consulting company called Solidify, and I also have my own company that I work for myself for my IT training and some Azure content, which is JonahAnderson Tech, which is my own brand, I do in my spare time and that tech, which is my own brand, I do in my spare time and community-wise, I am very active and famous as an international speaker speaking at conferences.

Speaker 2 3:33

I'm a Microsoft MVP for Microsoft Azure Technology and I also teach Azure certifications as a Microsoft certified trainer and also regional lead this year for Nordics, or for Sweden at least, to support the MCT or trainers community. So that's what I do, but day job. I am a consultant that works with diverse of things like development with focus in NET full stack DevSecOps with focus in NET full stack DevSecOps. I'm also kind of like acting as a role as a lead cloud infrastructure engineer, which I have a team that I mentor and teach and work with Azure as a platform in my job or project. So that's a brief, quick version about me, without saying too much like a diary. Thank you for asking.

Speaker 1 4:27

Yeah, and I think you mentioned what a hundred reasons why I could call you an expert.

Speaker 2 4:34

Yes, I should write a book about it, an infographic Good news. So what do you do in?

Speaker 1 4:39

that little free time you have left. So what do you do in that little free time you have left?

Speaker 2 4:43

Ah, that's a great question. So if I'm not on my computer, I spend a lot of time with my family, most especially taking walks with my dog. So I have a fun fact about me I have a three-year-old Labradoodle, a white one. It looks like a teddy bear, a very, very smart dog, very kind and very cute. And aside from that, I also like reading non-fictional or non-technical books in my free time. So if you visit my home office, you'll see lots of books about leadership and non-technical ones. And I also like gardening. So in Sweden we have short summer, so when it's winter time is almost over, I usually plant seeds of vegetables and flowers so they are ready for the summer in my garden. So that's my top three non-technical hobbies that I do in my spare time. So I still have a life.

Speaker 1 5:44

Fantastic.

Speaker 2 5:46

Yes.

Speaker 1 5:46

And let's try to go a bit down memory lane maybe. When did the interest for data start? So where?

Speaker 2 5:55

does it come from? It all started when, of course, with the work line of work that I do as a consultant or a developer, right. So when you're developing applications, it's not just about writing code. It is also about being able to write a code that you can use to do CRUD operations reading or update or deleting operations from your application to the database and databases is a huge platform itself, depending on what type of data you're storing into a database, regardless if it's on-prem, on cloud, so NoSQL, sql database. So that's how my journey to working with data and databases started, in different platforms with data and databases started in different platforms, and also I've been into different kind of projects, both modernization project, bug debugging projects, cloud migration projects, new projects, some mix of both.

Speaker 1 6:56

There are so many reasons also for you to write the book you have written Learning Microsoft Azure. Can you tell us a bit more about your thought process towards the book?

Speaker 2 7:06

Yeah, sure.

Speaker 2 7:07

So from the beginning, I want to share transparently, also to you and the listeners, that from the beginning, the thought process of writing a book started about three years, two to three years ago.

Speaker 2 7:23

But it was because before that, years before that, I was involved in a cloud migration project as a consultant in which I had the responsibility to migrate NET legacy applications, which is undocumented, very old the coding styles are not great A legacy app that we're like getting out of face, and we decided to move that to Microsoft Azure as a platform, and I was the one that was kind of like the only person that knew about cloud back then. So I learned a lot from that process in a way that it inspired me, really inspired me, to write my book for almost two years on my spare time, on weekends and nights, just to have that goal that I really want businesses, developers, it leaders to know what cloud is about. And it is not just moving one thing, an app to the cloud and it's all good. It's a lot of process, a lot of things to think about.

Speaker 1 8:32

Also very true.

Speaker 2 8:34

And.

Speaker 1 8:34

I think your book reflects that in a very good way. So we talked about cloud migration a bit already and we're going to talk a bit more about that later on. But if you should say, well, this is probably the most important chapter of my book, which one would it be?

Speaker 2 8:51

Oh, that's very interesting question because, as a writer who invented or wrote my book, I like all the chapters and I find value in each of the chapters. So, before I answer you what is my favorite, I like all the chapters because, for example, chapter one is relevant. Because chapter one is about cloud computing. You can't learn what Azure is without truly understanding what cloud computing is. So that's one factor that I also like. Chapter one because it is inclusive to beginners and anybody that doesn't know about cloud. But to me, is it only one that I get to choose? All right, if I were to choose one, I think my favorite one is the one where I wrote about cloud migration strategies. There's a dedicated chapter for that. After I explained everything a lot of things about Azure, like data and analytics, ai, machine learning Somewhere in the second half of the book I have a chapter about cloud migration and strategies what to think about adapting to cloud or modernizing Very good.

Speaker 1 10:10

I had the pleasure of reading the book, thank you. I kind of also liked the cloud migration chapter. But you are right, there's some basic groundwork that you're doing in the book that is really important to understand the entire setting we are operating in. So the first chapters, when you talk about cloud computing and explain what it is, I think those are valuable for everyone on all levels of competency and expertise. I think if you're a new beginner or have no clue about cloud computing or Microsoft Azure.

Speaker 1 10:45

I think it's really important to get the groundwork. But also for experienced professionals, I think it's always good to remember what setting are we actually working.

Speaker 2 10:54

That's very true because I remember when I was beginner like years ago, beginner of Azure and I was preparing for my Microsoft certifications myself as a developer.

Speaker 2 11:05

I work with NET development, I work with Microsoft related databases, I work with Microsoft as a cloud platform. But when you hear about Azure, sometimes if you're a beginner, you don't know where to start. Because that's why I laid out my book to be cloud computing basics first, and then I kind of like grouped the categories like compute, the data and databases, the analytics, the AI, the migration and developer tools, so that they know that, hey, there's a lot of group of different things and services on Azure. But it can be used depending on the use case and what purpose you're going to use it for Like since we're focused on data in this podcast also, for example, if your application has these data types that are NoSQL, probably it's worth considering looking into Azure Cosmos DB compared to the standard Azure SQL, for example, for managed SQL services. So it really depends on the use case. But if you're a beginner, you don't know where to go. You will just like gather this information and you probably end up learning the wrong things and get overwhelmed with too many things.

Speaker 1 12:28

Information overload, I think, is the term for that right Exactly, and I think that's really easy to get if you are just trying to, by yourself, try to understand all the offerings that Azure has. And I think you paired it really well in the book because you have, on the one side, the offerings itself, almost like a service catalog. You get an understanding of what are the different services and solutions you can pick from. But on the other side, you put them into context into context of the work you want to do, but also in the context of how you do the work. I think you have a known chapter on DevOps in there as well.

Speaker 2 13:06

Yes, yeah, yes, I did so. I kind of like laid the chapters of the book. That's why all of them are almost my favorite. I laid out in the book that you're in a journey itself from this no, zero knowledge at all to someone. Like chapter 14 is about developer tools for Azure. So I kind of like laid it out you know about governance, you know about DevSecOps, you know how to develop all this, like different things, and then here are the tools. So it's like you're ready to do gardening, since I'm using garden. So you're ready, you have your tools, you know what you planned, you know what is good for the season, you know how you're going to do it and you just do the project.

Multi-Cloud Strategy and Security Considerations

Speaker 2 13:47

So, yeah, I agree, I did have you mentioned DevSecOps, so I did have, since I work every day with not just Azure development and Azure management, but also with DevOps, and I'm very into security lately, especially learning more about it and working with it in a way that I truly believe, as someone with a background of developer or development, I truly believe that injecting security into everything, including DevOps, in the entire pipeline, from software development process up to testing, up to its production, is very important Not just development, but also integrations, the things that we implement in our applications, apis and all those other endpoints that we connect to. So very important, because in my job as an IT consultant, I noticed that we are too busy delivering tasks every sprint, but sometimes the common things that are really important that we miss are implementing security and fixing technical depths, which are actually two major things that matter when it comes to application. But you're talking to someone who works in data governance?

Speaker 2 15:12

Yeah, I know, you know how it is. Yes, I also would like to deep dive into your data governance, which is one of the topics that's also very interesting, especially in Europe. Right, like the GDPR and all the governance, like policies we need to follow.

Speaker 1 15:33

Oh, yes, very much so. I think the regulatory landscape is just evolving more and more, and that's one thing to keep track of that, but also to have to write competency and people in your organizations to work within those regulations and the frameworks they're giving.

Speaker 2 15:52

I agree, and I think that's when I like the synergy of different roles.

Speaker 2 15:57

Like your competence, like your expertise in data and governance and me someone maybe like has the experience in cloud development and DevSecOps, I think the synergy of different roles within truly matters, because you exchange information, you exchange ideas, lessons learned and that will actually, in my perspective, will help build resilient and secure applications, because in our line of typical roles or job, a developer only writes code. They like to I mean, that was also my favorite Developers like to write code, develop, focus on what they need to do, qa, do their testing. But if everyone knows that, hey, we need to think about how we govern our data, how we secure our data, how we secure our applications, then it will be like a mindset or a habit that just go within the flow, like how DevSecOps should be. And it's very interesting and we're going to discuss a bit data and AI as well.

Speaker 2 17:09

So the AI world or AI trend today is something that I'm still adopting myself. I try to make use of it in a productive way as a tool, but not something that I will depend on 100% as a human being. But security is something that needs to be also thought in there. How about you? What's your perspective in that, winfred?

Speaker 1 17:34

Yeah, I think you dedicated a chapter to AI in your book as well. Yeah, yeah, I did Definitely interesting read. I think there is a lot happening on the AI side and I think nothing could have happened without cloud computing in the first place. I think really and I said it in the beginning, I think that cloud computing is really the backbone of the scale we are working in in data, and AI just puts it into an entirely different level. And AI just puts it into an entirely different level.

Speaker 1 18:04

Agree, we are seeing certain changes when it comes to cloud computing and AI, at least from my perspective. There is a push toward edge and edge AI. What can I do without being reliant on cloud infrastructure? And the reasons are security mainly, while there's a latency reason, I think that's not something that we need to talk about so much, but I think on the security side, there is a certain risk of exposure of your data. When you think of, for example, training data for deep learning application, there is an exposure and a attack surface right. Or, for example, training data for deep learning application, there is an exposure and a attack surface right. For example, data poisoning that we talk so much about lately. If you manage your training data on your own machine. You minimize the risk right.

Speaker 2 18:59

So I think it's going to be interesting to see how the role of cloud will evolve around around ai, especially when we think of security yes, I really like what you shared there, uh, and also sharing your thoughts about, like, what will happen, what it's gonna be, with all this consideration, considerations and factors that we know already. It's, uh, it actually like makes me curious to learn more. And what can we do as experts, also in different fields that we work with every day? Because we need to do something if we know that there's a risk that it can go wrong, if you know what I mean. But focusing back to what you said on cloud computing, I truly believe that cloud computing is the backbone of AI. Because of AI, artificial intelligence and machine learning and plus the power of cloud, then it's possible. But if it's just AI and ML without the cloud, I don't think it's possible. So the synergy of the three is impactful on what we are experiencing today in our modern era.

Speaker 2 20:07

But behind the scenes, I still do see some challenges in terms of security, like I can set an example in my project. We don't use AI tools, yet not specifically in my project, because we're concerned about what we are exposing out there and my company that I work for I'm not disclosing too much details, but we are conscious about data privacy. We want to keep our private data secured, even though we're using Azure or cloud as a platform. So it is important to have the awareness and education and not just build something with AI. So being able to use it responsibly, with the thought of security in mind, is important. So what's your perspective on that? Winfred, you have anything to add?

Speaker 1 21:06

From my perspective, I think it's about balancing. You need to understand what you are exposing, at what level, and if you have a good understanding of your input data, you can, in a certain way, track your output, also through a blackboard. I truly believe that At the same time, there is a certain risk on the model governance if you do it on-prem. I feel like you need a certain level of expertise, a certain level of competency. Let's make it quite easy. I think Azure has a whole set of AI services that make it easy for you to get started. If you're building a face recognition or you're going into computer vision look at certain test data on picture recognition you have a certain set of tools that are ready to use on Azure. You don't have that on your own PC, right? You don't have that on your own machine. So you need a certain level of expertise in your company, a certain size and a certain investment to get there. So I think you have to figure out what is the risk of exposing the data compared to the investment you are taking.

Speaker 2 22:13

Yeah, I do agree In addition to what you said. Also, like you said, asher, I have a chapter about AI and machine learning and I did mention about this cognitive services that you can implement with your cloud development, with the applications. But, like OpenAI, azure, openai allows you to build a lot of things through endpoints, but if you don't know how their security risks of building your integration to the endpoints, the keys where you save it, risking it, you're also kind of like exposing it to hackers or bad people. Who has a bad intention? And also data is important. So, like you said, we need to identify how sensitive is the data that you want to protect and if everything is in the cloud, and then you have this governance policy that you can't do that, depending on the region, of course, where you are located. So probably, in my perspective, there's a solution for that.

Speaker 2 23:18

So Azure has an option for Azure Hybrid Cloud, so some parts of your data or application can be on-prem and then some parts that you want to expose public can be on the public cloud. So I believe that, for example, the Azure Arc allows the integration that some parts of the data that you want to connect to the cloud can be controlled in that manner. But machine learning and using edge AI, as you said, it requires expertise of someone that already knows and can simulate what the cloud tools ready on this side. So I think that's one of the reasons Microsoft Learn is implementing a lot of cloud AI skills lately to the community. So I think it's important you talked about hybrid cloud.

Speaker 1 24:13

There's another thing that is kind of interesting to talk about, and that is multi-cloud. We're going a bit in a different direction, but what would be the main reasons to and I'm not talking about AI in particular now, but more about cloud computing in general what would be the reasons to opt for multi-cloud versus single cloud?

Speaker 2 24:30

Yeah, that's a great question and a common question also to those that are getting started into like modernizing or migrating to cloud. I think that's one of the questions an IT organization needs to consider from the beginning of their cloud migration phase. Organization need to consider from the beginning of their cloud migration phase. But the primary drivers why some IT companies choose multi-cloud strategy it's because of avoidance of vendor lockdown. They don't want to be locked in one certain providers.

Speaker 2 25:02

Another reason that probably is true that I know is probably they have an existing applications or infrastructure in another platform, like AWS, but they want to try the AI solutions that are better in another platform. So they kind of like choose different multi-cloud strategy to solve the problems or build solutions with their use cases, solve the problems or build solutions with their use cases. So those are the top two things, at least I know, that are very common on that perspective. And then, since I work with DevSecOps or the infrastructure management, I do coding with infrastructure as code and I know that if you only have one platform for infrastructure management and automation, you can use Bicep as a standard. But if you are choosing multi-cloud strategy for automated infrastructure management in building your resources, then you would probably consider like other platforms, like using Terraform, which is a very good support for Azure resources, or Azure RM. So those are the things that are different layers and interesting to me, not just AI, but it's good.

Speaker 1 26:19

No, no, exactly. Yeah, ai is kind of the sherry on top right, yes, sherry on top. There's so much more to talk about. There's one thing in your book that caught my interest, and we talk about platform as a service. We talk about infrastructure as service, but you also talk about data as a service in your book. Is data as a service really a thing to reckon with?

Data Lifecycle and Cloud Migration

Speaker 2 26:42

That's an interesting question. We don't hear it a lot because of the, but I did write about a bit about my book, especially when I spoke about cloud computing. Like you can have like functions as a service, you can have compute as a service pass, but they are just like the standard there's. Also you can build something as a service and I truly believe if you have a powerful data platform, you can have it as a service that suits a specific use case. But in my perspective I'm not sure. Maybe I am in a different field, I'm not sure if it's getting popularity or it's being over dominated with something else.

Speaker 2 27:27

But in my field of work in Azure, I work with Azure every day and I work with databases. So I did help migrate once, a few months ago we were moving data or upgrading our infrastructure, moving away from an Azure SQL managed instance. You know managed instance it's like virtual machines of lots of database servers, thousands of them that can Azure autoscale up and down depending on the performance needs and requirements. But we recently moved away from that because of our use case scenario. We wanted to be able to have the features like auto failover, the ability to auto replicate, the ability to auto backup, which was more available in the past form of Azure SQL. So we did move to that. So I can consider actually Azure SQL in a PaaS model, and especially the serverless, as kind of like in a category of data as a service in my perspective in the cloud competing arena, even though maybe it's not labeled that way, but it's almost close to that category.

Speaker 1 28:48

Yeah, I think there are different perspectives on it, and you can come into data as a service from a very technical perspective, but you also can come in from a data management perspective, trying to implement certain strategies, certain operating models. Yet I haven't seen it really work, so that's why I was interested in your perspective on it.

Speaker 2 29:11

Yeah, I haven't seen it a lot as well. It's there, but it's like the word data as a service is not being raised the flag more than SaaS, maybe, or AI as a service, so we will see how it comes up. But data is part of what we do and it's what we build every day. So I used to work as an SEO consultant. Before I was a programmer or a developer and I know there was this term like data is king. Like when you say at that time, long time ago, google, you could google everything and it says, oh, data is king because the google has everything. But now is the other way around. It's like co-pilot maybe has all the things that we need to ask about. So it's uh, data is very important and we put it in every day from our users in our applications, and how we protect it, how we govern it is really important, especially where we store. It is also important how we protect. So it's a very interesting topic we are discussing right now.

Speaker 1 30:20

Oh, definitely, and I think, especially when we talk about and this has been a term that's been floating around for decades a data-driven organization, whatever that means. But the data-driven organization is relying on data. I mean, that's the only given we have. The question is, just at what points are you relying on data? And I think that, no matter how you define your data-driven organization, I think it's important to have a certain structure and place where the data can reside on. And then we are back to the cloud computing, right?

Speaker 2 30:52

Yes, that's right when the data resides, on how you protect it and how you handle big data, because I know in my book I did write about big data, because you have a bunch of data but it can be very huge and big data and kind of like AI data. It's kind of like everything connects to each other. Maybe they just have their different focus, but everything relates to each other. And how machine learning and AI learns or learns to do things it does to help us, it's the data that matters. How genuine, how clean is your data? Because if we inject or put in wrong data there, then it will learn the bad way or the wrong way as well.

Speaker 1 31:37

And just to close a bit of the loop, to the data as a service. I think that what we're talking about now, this is really where you can see an application of the data lifecycle and I think the data lifecycle from collecting data getting into your systems to storing it, managing it, structuring it, using it, retaining it and even ultimately, deleting it that entire lifecycle can be supported by certain services that Azure can offer.

Speaker 2 32:04

Yes, it does so. For example, azure Storage does that the blob storage. You can even like, recycle or set some like tier levels to it, like the HUD archive and different levels depending on how you want it, and you can protect it using SAS keys and connection strings. And also related to that in terms of governance as well, like depending on where you are in Europe. We have GDPR. Then we also have to consider, like, what kind of data you're storing in and how long you're going to store it, because there is a law and I think if we break that law, I think it's going to cost a lot of money to certain companies if they're not so conscious about that.

Speaker 1 32:49

And I think that consciousness or awareness that actually starts before you get data into your systems should start ideally. Yes, that's right. Which brings us to cloud migration. Yes, really, and you talked a bit about it already throughout. But before we talk about the what and the how, maybe we should talk about the why cloud migration, and we talked about some of the advantages on cloud computing. There are different scenarios that you can see where cloud migration happens, and it's not just the oh, I want to move from on-prem to the cloud, but it actually could be between different cloud providers. It could be a merchant acquisition situation where you're trying to consolidate data. So maybe we should start talking a bit about the main reasons for cloud migration that can lead us to maybe some of the most common patterns that can lead us to maybe some of the most common patterns.

Data Migration Best Practices and Pitfalls

Speaker 2 33:52

Yes, in my experience the reason why I was involved in the cloud migration was because of out-of-date virtual machines. So basically it was mostly the reason of both moving from on-prem to the cloud, plus modernization, which can be two factors that can be combined, and of course, the rest of the other reasons or drivers that you mentioned, like merge of two companies or moving from one platform to another. But the primary reasons that I know in the business perspective or IT companies perspective why they move to the cloud First, the ability to handle this resiliency in the cloud, which is not easy to handle when you're on-prem. The geo redundancy, the feature to be able to auto-replicate your data from one region to another, performance-wise in different global scales and being competitive. And the ability to auto-scale when you need it, like when your application demands. It is really a good advantage.

Speaker 2 35:02

And, of course, upgrading databases. Sometimes it's not easy to do it on-prem, like, hey, I have my data is growing in my database. On-prem it's not easy, but in a cloud you just go manage it and then you can have a super hyperscale database that can handle a lot of performance and compute on the cloud. So I think those are the major, both application-wise and data operations-wise as well.

Speaker 1 35:30

And I think that if you are in a migration setting, there is certain steps you're taking prior to the migration, during the migration and after the migration. What is interesting for me, what is interesting for me from my perspective, is certain data quality dimensions that you can uphold during that process.

Speaker 2 36:10

So how can we ensure that data is available databases from on-prem to the cloud because on the on-prem database, which was not maintained, well, you do see this like dummy data, test data from different developers and testing and then you cannot just move all those like scrappy data, like the unclean data, into the cloud, especially if you're doing like serverless or like cosmos db. You need to restructure it in a way that it is like filtered and you're not you're. It's not just about moving things to cloud and it will work. It will be actually worse when you just do it. So it really requires a lot of strategy and when it comes to like strategy and getting started, I have problems. Two major favorite tools to use or to recommend. First is the cloud adoption framework for Azure, which actually guides organizations and how they can plan their strategy in cloud migration, and also the well-architected framework, which has the amazing pillars, which also includes security and what to think about architecture, data and everything in terms of preparedness and on the cloud platforms like Azure, for example.

Speaker 1 37:35

That's a very good tip. Definitely yes. One thing you mentioned it artly already and I think this is both for structured but very much for unstructured data. There's always like a thought of well, should we clean up our data and when should we do it? Should we do it pre-migration or post-migration? Maybe it's easier there are more tools available once you are migrated to the cloud than to do it on-prem, where much of the processes can be manual in the cleanup. So any recommendations here oh, that's really interesting.

Speaker 2 38:13

Interesting because it really depends also because if you're moving to the cloud from on-prem, you want to improve it, you want to upgrade somehow. I would probably consider if I was the one that makes the decision should we clean up our data first before we move? In my experience so far, I would definitely try my best to clean up as much as possible the data that I have on-prem before I migrate it and, if possible, rethink and re-evaluate the data relationships that you have between tables. Because probably at the beginning of the architecture, beginning of the application creation, maybe the thought process of this is the tables we're going to have, this is the data that we're going to have.

Speaker 2 39:03

But over the years, when the application goes into the production, data comes in, new columns comes in and then developers are busy building things and DBAs are busy doing their SQL queries but they don't realize that the data and the tables and the relationships are getting too complex and the relationships are getting too complex. But that probably has the potential and opportunity to be rethought, re-architected, the entire database structure itself. That probably is potential for maybe using NoSQL instead for this data type or SQL, so kind of like kind of make it, micro databases, if it makes the application faster, like improve it if it's going to, if it's possible on cloud.

Speaker 1 39:51

So there's one more thing I wanted to talk about during the migration process, and that is monitoring and migration orchestration. So how do you ensure that you move the right data at the right time during the process?

Speaker 2 40:07

That's an interesting question because that was one of the things that was missing actually in my migration project, because at that time, many years ago, we didn't have the cloud adoption framework and the well-architected framework. So I really had to learn the hard way, which became my book. So I did the lift and shift, doing the modernization, refactoring and re-architecting. So I did all of that. But to me, if I were to give an advice based on what I know now and what I learned from that migration experience, I would consider making use of monitoring tools. That's built in within what Azure provides, like the application insights for NET. You can use it even if you have your applications on-prem. That's a good way to track your applications and definitely there are other observability and monitoring tools available also that support this, so you know what's going on during the process.

Speaker 1 41:13

Very good. Maybe a last question to cloud migration, and this is maybe many of us have been in the situation where certain things fail during your migration process. That I mean, you see certain patterns right of pitfalls that people have or things that go wrong continuously and over and over again, but are there some pitfalls that you would say be aware of those?

Speaker 2 41:40

Yes, I believe there are a lot of them that I'm sure the tools that I shared Cloud Adoption Framework would provide. But my top three pitfalls or try to avoid at least, is based on what I learned is that first is don't move to the cloud without truly understanding what you're getting into, meaning you need to prepare so the entire business like strategy, migration strategy, and you don't have to move all of your applications to the cloud. One of the advice of the cloud adoption framework is that choose your first migration project like a pilot project. If that pilot project works small one, that's good, Then it's a good start and also being able. Number two is having this cloud readiness and cloud education mindset, because you're moving to the cloud because it's needed, it's required to be competitive in the market. But if you as an organization don't really believe in the power of cloud computing and what it can do for you and your business, then it's useless, because you really need to believe and educate your team members about that.

Speaker 2 43:05

And then, number three that I think I would recommend is just take time to do some cost optimization or cost management, Because one of the things people get confused or get really worried about cloud is the cost, the cost and also security. So cost first. So think about what your architecture would be in the cloud and, if possible, do the TCO calculator or Azure pricing calculator as a tool and the assessment tools available for Azure, like Microsoft Assessments, are good for that. And I have number four security. Of course, Think about security, which is part of the well-architected framework and the cuff part section. So everything is important and I really enjoyed it.

Speaker 1 43:59

Throughout the entire episode. You managed to tie everything together through security. When we talked about AI, when we talk about migration, when we talk about cloud computing, security is really connecting those topics and I think that's a good thing that I set it up this way.

Speaker 2 44:14

Yeah, I think so, because everything is connected and I think my entire book, even though it speaks about different topics, they connect to each other to build something great and robust. And when we speak about AI, we have data, and when we speak about data, we think about security and hackers, the bad guys. They are trying to hack you and get into your system because they want the data. So it all goes back to the important things that we're talking about. Thank you for a well-organized discussion that was really also very spontaneous, and go with the flow.

Cloud Migration Tips and Resources

Speaker 1 44:52

Thank you so much Before we finish do you have any key takeaways or a.

Speaker 2 45:02

What I shared is that, before you move to the cloud regardless if it's migration, modernization or just like a development learn about cloud computing, the fundamentals aspect of it, and build that mindset and culture within your organization. So that's my first part. So, if you're choosing to move to Azure, read my book, not because I'm selling it or sharing it, but I think it's the reason why I speak about it all the time in the entire Europe. So that's my call to action. And then, if you are not sure where to go, always contact or reach out to the cloud providers, because they always have this initiative, like MSPs, and even Azure migration program, especially to Microsoft partners or IT organizations. That helps those that need help. And aside from that, that's all. Actually, I mean connect to the community, talk to experts, listen to podcasts like this. Then you'll learn new things, that's all.

Speaker 1 46:16

Thank you so much and thanks for sharing your insights and knowledge.

Speaker 2 46:19

Yes, you're welcome and thank you so much for having me in your podcast. It's been great.