Modernizing Microsoft workloads on AWS | Amazon Web Services

Key Takeaways

Hello everyone, welcome to migration and modernization podcast. Today we are going to talk about migrating and modernizing your Microsoft workloads to AWS. My name is Sio Irani, a solutions architect here at AWS and today I'm joined with Yogi Barrett. >> My name is Yogi Barrett. I'm a Microsoft worldwide tech leader. So Yogi, could you tell me about what type of Microsoft workloads do you see customers want to migrate to AWS? Uh >> absolutely. Yeah. When we talk about uh Microsoft workload uh in generally any kind of Windows ecosystem that you have uh whether in your data center you are running u SQL server on Windows server orn net applications file system active directory or you have an collaboration suit like Microsoft sharepoint or Microsoft dynamics any of those basically you can bring onto AWS essentially any Microsoft program that you have it which is running on premise that can all go onto AWS whether you want to do just a quick lift and sift migration or you want to modernize both uh approaches are valid. >> Mhm. And what type of challenges do you see customers are facing when migrating these workloads to the cloud? >> Let me tell you about the challenges a customer face when migrating Microsoft workload. So the number one is licensing and the reason why it's number one licensing is like kind of a headache. Microsoft licensing can be tricky and especially to whether determine if you can port those license to AWS or not. Um and once it's you determine whether you can port it then like whether you should bring that license or not or whether you should use AWS license included instance or not right so that can be challenging for customers to determine uh another thing I can uh say that a big problem for the customer is the legacy application right so you know those old um business critical application that uh everyone relies on they are not fit to run into newer operating system. Right. >> Most of the time the developers they have probably left the company so they cannot. >> Yeah. So those old legacy applications are not cloud ready. So in that case that needs an legacy Windows operating system component and that can be challenging especially when you want to do data center exit. Um and if you have those legacy workload that business does not want to get rid of it because they are very old but critical for business program and um in that case uh that can be a challenge. And lastly what I can see here is um if you have a high performing low latency very large database uh like say SQL server database uh with a application dependency and you have like complex network setup um that needs to open the VPN and direct connect and figuring out the right um latency requirement. Um some application and database needs like IO intensive workload. Some of them are like high memory optimized workload and some of them are like high CPU requirement. So balancing out uh between memory, CPU, IOPS, uh the storage performance and figuring out the right balance sometimes it can be challenging if you don't plan properly. It needs to be planned and tested. Actually on top of those what I have seen is customers who are running active directory there there's always questions I'm I'm getting asked of those infrastructure services like how am I going to how my DNS is going to be set up when I'm extending to the cloud and on top of that I believe another thing I've seen is that learning curve right like when uh basically the teams are like oh you know moving to the cloud it's going to be hard for me to operate and I don't know where to start and all of that so those are couple of different challenges that I have seen with uh with migrating to the cloud. >> I see. So what kind of consideration you do for um migrating Microsoft workload to cloud? >> So I think uh when you look at the cloud uh migration as a journey, right? Uh you can look at the three phases that we define in a cloud migration, right? The first phase is called um SS phase, right? This is where you actually figure out even what is it that you're running in your environment. And that's uh you know if if I ask you hey what are you running in your environment and you export an excel spreadsheet give it to me and that moment you export it is already out of date because the uh especially when you're operating at scale things are dynamic things change very quickly so uh you know you need to have some sort of tool where it gives you that dynamic inventory of your workloads and dependency mapping of you know what is it that how applications are dependent to each other. So that's the first phase, right? That's the first assess phase where you need to figure out what you got. And then the second phase is the mobilize phase. This is where you're building your account strategies. You're building your control tower. You're defining uh how your migration is going to happen. Maybe training your teams and you know the different tools that they may be migrating. And of course the last phase would be uh you know doing the actual migration. Now one consideration I want customers to make is that the challenge I see is customers get stuck within those first two phases where they do an assessment they find some stuff and then they are like okay now I want to you know build an account strategy for it how am I going to do it right and they keep re going through this loop of uh you know assessing and uh you know revisiting uh their environment and they never migrate or their migration takes long because uh they just basically spend spend too much time on that. So I'm not saying that the assessment phase and planning phase is not important. It's definitely super important. But uh one thing to note is a lot of these decisions that you make while migrating to the cloud are two-way door decisions. Let's say if I pick an instance type that is wrong or it's overcommitted or under uh committed, I could go ahead and change that with a simple API call. Or if I didn't assign a department proper AWS account for them, I could go ahead and change that later on. So because these are many of the many of these decisions are two-way door decisions. My suggestion to those customers are you know if you see you can make a decision faster go with it and you can always come back and change those uh decisions later. And absolutely that's the beauty of cloud right you can anytime shut down your instance anytime upgrade it change it right so uh one thing I would like to stress about is assessment right so when we do this assessment uh we have a tool called um optimizing and licensing assessment OA we call it as an OA with OA typically customer can save 35 to 65% of cost right and uh it's designed to help um customer save cost especially especially for Microsoft workload uh because um uh Microsoft uh uh workload requires licensing right so the OA helps in two ways right so let me say the first way is basically it's going to look at your overall environment uh look at the footprint of what you have deployed and it's going to suggest a right sizing recommendation most of the time customer environment is overprovision either you have more memory more CPU more storage than you need uh But this tool is going to go look at the actual realtime performance data. It's going to look at some um trends and it's going to do an analysis and generate the right recommendation based on what you need rather than what you have right so that is the key to save cost and second thing OA does is about licensing assessment um most of the times customer license are either overprovision or they are not utilizing license efficiently. So doing this uh OA will help u first whether to determine you can bring that license to cloud. Second uh determining in which cases you used AWS provider licensing when it makes sense to use AWS provider licensing. Um and the third thing is uh what kind of SQL server version and addition to bring how to save cost for your SQL licensing and the lastly is consolidation because consolidation can help help save licensing cost as well. Right? So I have firsthand seen this recommendation when the customers start using this OA at the beginning of their cloud migration journey they have like a proper infrastructure and licensing strategy and they save lot of cost actually it's not about what you do or how fast you do the migration it's when you migrate you need to determine about how smartly and uh securely you saving um cost from the beginning of the migration. Uh let me tell you my story. When I was a DBA back then and when my business um uh counter people asked me about what kind of SQL server addition you would required and I would say let's go for enterprise edition and now thinking back again um do I need that enterprise edition? Not really all the time. I just had it because just in case I need to use it for the future, right? Uh but going to the cloud that is very critical decision because there's a significant cost more than 50% cost difference between standard edition and enterprise edition, right? Depends on the version and and the addition. So uh you need to carefully consider you can always start with standard edition if you're not using the enterprise edition feature and you can always change it later on as you said like you can always upgrade and move it to a a higher version right >> yep yep and actually I remember uh when I was a system admin uh we used to run things on VMware and uh we we we had this big physical servers where you know the the memory and the CPU was sitting there on the server if we didn't assign it to VMs then you it would be just we already paid for it so we might as well you know assign it to to to those servers. So at the time like all of our VMs virtual machines I remember they were all over provisioned right and but that's because of the nature of how you operate on premises. >> So now since we covered the assessment and licensing and all this consideration um can you talk more about migrating those workload like how do you migrate Microsoft specific workload to AWS? Yeah. So there are different tools that AWS offers uh that helps with the migration, right? Uh first let's say let's pick those virtual machines that you may be running. And uh when you look at virtual machines that uh you're running, you either uh there are of course different virtualization platforms from the famous ones being you know VMware, HyperV, maybe you're running things on open source or uh other virtualization platforms. Now uh for migrating virtual machines as they are uh there are different tools that exist. One is VM import and export tool which is a bit of older tool and the way it works is uh basically you are specifying the discs the virtual discs that are attached to your VMs and uh VM import tool. It's going to upload them to S3. It's a whole process in the back end. I'm not going to go into details of it how it works but end of the day basically what you make an API call those discs are uploaded to uh AWS they're converted to EBS volume and we add all the required components like like drivers or things that is required for those operating systems to uh basically run successfully on the AWS hypervisor which is the nitro u and uh uh then we we create an AMI for the for you where you can now launch your instances from that AMI. Now, uh that's one way. Uh and the other tool is application migration service or short for MGM. What application migration service does it's meant for those customers which cannot tolerate any downtime. And the way it works when you set up application migration service uh for your workload and uh you know you you can set it up for your on- premises servers or if your servers are located let's say in other cloud providers. uh when you set that up it starts doing a continuous replication of all the disks that is attached to your operating system right so even you think about it if if it's a if that disc volume is over like an icecazzi it still starts you know capturing all those chunks and uh disc blocks and replicates it to AWS and then uh MGN allows you to create a test workload based off of everything that was migrated so uh after initial migration may take initial replication may take a while and then after that's done continuous replication remember during this time your original your source servers are still running we didn't shut them down and now you want to test it you go into MGN and you launch a new instances or set of instances from uh that you know shape that workload in in a normally isolated environment you test the workload once it's successful then when it's the time to cut over you basically have MGN to launch your final instances which you can now you know launch and uh you know basically then then you you're going to shut down the original source servers. >> So do you have any customer example that they are using MGN like any any kind of a benchmark that you use for how many number of servers or things like that? >> Yeah. So actually I I had bunch of customers uh which use MGN. uh one of them recently I was involved with where they had uh this uh multi-ter application where uh you know it was consist of is web servers mid-tier servers and then backend SQL servers which uh they just wanted to really get a grasp of how's it going to operate in AWS and uh uh they exactly did that they they they used MG and replicated the whole environment have their team test it once it's all successful then they went ahead and basically deprecated everything that they had on premises and the beauty of it is because it's the continuous replication uh really there they had minimal downtime while doing that right so that's uh basically uh with the MGN now um also another thing that comes up a lot is uh file servers right so uh customers run different file servers of course right it could be windows file servers it could be uh and sometimes uh often you see you know they have storage area networks like SANS uh and uh for those uh again if you're running on VMs totally fine to use something like MGM to do a you know uh block replication of all your file system and bring it to the cloud but also if you want to go with managed services uh if you're a Windows shop uh Amazon FSX for Windows file server uh where basically it's a managed service for building a cluster of Windows file servers it's compat fully compatible with Windows and SMB uh uh uh different uh versions and what you do is uh you basically you can use either AWS data sync to transfer your data over or uh using something like native tools like robocopy to copy your files. Now for those customers who are running um NetApp sand stoages uh what they can do uh they can use Amazon FSX for NetApp on tap which also gives them a native uh NetApp uh functionality on AWS and for migrating to those of course they can al always use the native NetApp tools uh to to migrate those files and now I know I talked about uh storage uh Yogi I know you're expert in SQL so how about SQL server >> uh sure For SQL server, we have different of migration options. Actually, first let me talk about the deployment option and then we'll cover the migration. Right? So for SQL server uh depends on um what your use case you can deploy onto um Amazon EC2 or you can deploy onto our managed service Amazon RDS for SQL server. So with Amazon EC2 for SQL server, it's like a lift and sift migration. Uh you can bring your license as well. Um it's pretty simple. It's an flexible option and it's just like the same feeling as running it on premises right so uh if you have a specific requirement to have like full control of the database environment and doing the same exactly um what you're doing on premises then you can run it on EC2 instance that's a choice for you uh the second option we have is running on um Amazon u RDS for SQL server right so with RDS for SQL server uh it's an fully managed service. So it gives you an advantage of um AWS managing all the day-to-day administration task like backup, patching, high availability, everything is managed by AWS and you can focus more on data and application uh strategy and at the same time give the administration overhead to AWS. Right? So that's about um two different deployment options. Now say you like a managed service and you also want a flexibility. Yeah. >> So for example I have a customer asking about hey I want to um run an custom monitoring agent like data dog or any other monitoring tools. I want to install that onto my database server or I have a custom driver requirement and I need uh I like the manage service but I need that uh as a right for my application. So those kind of customization if you want to do with your uh RDS we have a service called Amazon RDS custom for SQL server that allows you to give a managed experience at the same time you can do this customization based on your need right so it's a good of both uh world um so depends on what you want like the the best option uh you determine based on your workload based on your requirement and now when you want to modernize right so This was like SQL to SQL migration right but see you can also have SQL server migrated to u Postgress like Amazon Aurora for Postgress SQL or RDS Postgres SQL so in that case um you are moving away from SQL server uh engine and you are going to open source engine right so that's a difference over there >> now let's talk about the migration tools for SQL server right >> so for SQL server specifically actually um I remember when you're talking about application migration service. I had a lot of customers use that for VM migration, right? So they just want to migrate as it is everything on premises including operating system uh SQL server installed any applications any database everything installed as it is um they want to do migrate uh to AWS they can use um application migration service MGN um or they can use other native tools like uh SQL server backup and restore right so it does a uh supports the full backup um differential transaction log backup if you want to run that on IDS supports the full backup and transaction back up and restore that as well. Uh and um if you want to do other tools like log shipping, you might say yogi that's very old. I know it's old but it still works, right? So old is gold. It still works. So um that uh lock shipping um if you are using always on availability GT uh cluster you can also extend your always on AG um clusters to AWS as well and then um migrate do migration that way or you can use our u managed service AWS database migration as uh service with uh DMS you can do homogeneous migration like SQL to server to SQL or you can do heterogeneous as well from SQL server to Post SQL as well. Those are variety of options for migration. Depends on your business need and use case. You can select either of them. Actually that remind me of of a customer. Um uh earlier you talked about one of the customer like so this customer is like a SAP conquer. They had their 80,000 plus databases on premises and they wanted to move to cloud to get an advantage of cloud scalability consolidate everything and um also save some cost. So they started migrating to um Amazon EC2 uh SQL server running on Amazon EC2. Uh with that uh they were able to finish that migration project in 8 months. Uh and they used application migration service and they used AWS database migration service both of them and um at the end of the project they were able to be 70% more efficient right cost savings and efficient by moving everything to um AWS. Actually you brought a good point about using native tools to do the migration. This is something I often get where customers uh when they're trying to uh you know migrate active directory right and they ask me hey should I use like a specific tool to do the migration and uh honestly often those native tools are totally fine to migrate like for example for the active directory uh I I asked them hey you know you have your existing domain what you can do just bring up a new virtual machine on EC2 and uh just add it as additional domain controllers for your environments. Uh now of course some customers they may want to have uh specific use cases to uh use AWS Microsoft managed AD which is totally fine and then establish a trust with their uh you know on premises AD. One best practice that we always tell customers regarding migration of active directory is keeping the number of domains that you're managing to a minimum. right now. Sometimes it's possible, sometimes maybe you had an acquisition or you have specific use cases that may not be possible. But um and regarding to that uh I want to highlight one of the features that AWS managed AD launched this year called hybrid managed AD where basically what we do is we keep your existing uh active directory domain controllers. You you you have that it's the same domain. Now what we do we add additional manage domain controllers to that same domain which uh which is managed by AWS now and uh you know that that functionality uh basically gives you the benefit of both worlds right both being able to manage uh use a managed service and at the same time uh keeping the number of domains to a minimum so uh I know we talked about uh uh active directory how about uh net workloads >> uh net applications has also follow similar patterns right so you can rehost to um Amazon EC2 like as it is left and sift migration uh it requires like minimum actually no code change um and you can run it um um on EC2 when you um want to refactor those uh net application you can run onto AWS managed service like AWS elastic binto service with that it's like a managed experience and it's a little bit of like um minimal code change with that as well and when you want to refac factor you can refactor or you can rearchitect as well right so forn net application you can basically convert into docker container and run that container onto windows container or linux container either of them run on to Amazon ECS or Amazon EKS right so those are also applic options for net application and uh you can uh convert to completely serverless like run on AWS lambda that's also an option right So depends on your use case whether you want to rehost to EC2 uh replplatform to um elastic beantock or you want to refactor and rearchitect uh to run on containers right so multiple options and depends on your choice and the last one is uh the container one and serverless one that's the most optimal uh but it's it's the one require more time in in conversation and migration so now that we covered about migrating um uh windows server SQL server um active directory net application file servers um can you tell about like once customer migrate those workload to AWS how operationally it's different when you run the same workload on premises and now you have to run the same workload onto AWS especially Microsoft workload >> yeah definitely and when you look at uh operations for Microsoft workloads I think uh one of the tasks that is very common and everybody must do any Windows admin pretty much is uh making the patching is making sure your OP Windows workloads are uh you know up to date uh with with all the patches that Microsoft normally releases on second Tuesday of every month right and uh uh you know again using native tools to do to do those tasks like for patching is totally fine right but when you're operating at the cl in the cloud and uh normally you may have some of your workload in one AWS account or multiple AWS accounts or multiple regions or uh some of the some part of that same workload may be sitting on your in your on- premises environment still or other cloud providers and you need like a a tool where you can uh let's say all these workloads they're you know part of a test and dev and prod environment. Normally when you do patching you do want to test uh do your test environment first and then maybe uh you know dev environment and then you know going with a prod uh now when you want to target all of these different environments at the same time that could become challenging. you need a better tool to do this. And one of the tools uh uh that can help with this is AWS systems manager patch manager which allows you to target all these instances no matter where they're located. And you can set basically different patch baseline uh that is compliant uh you know with your patch policies and uh apply patches at scale right and then get a nice report of you know what what patches got installed what failed and you know basically take action on that. So that's one thing and then uh another task that often you know comes up is monitoring right you do want to you have to keep monitoring for your servers for both performance metrics and also for logs to make sure if something is about to happen or if something went wrong you can uh you know uh fix those issues and uh of course again using native tools that you have or any third party that's totally fine but also Amazon cloudatch and cloudatch agent uh specifically allows you to capture metrics pretty much any metric that is shown in Windows performance uh counters you can capture those as well as logs different logs uh from Windows event logs to is logs or any just a pure text file you can capture all of that send it to a cloudatch uh dashboard and have a single pane of glass where you look at all these logs and metrics and and not only you can get notified if something goes bad but also you can take action to fix those issues. How about modernization Yogi? Do you want to talk about that? What does even modernization mean? >> U modernization actually um in today's context modernization is a process of taking your transforming your application and infrastructure into high value open-source or cloudnative services. Right? That's a modernization definition. Um when I say why customers should modernize is modernization can gives them um um get the better business benefit right. So for example uh they can go for innovation. It opens up the reduce the technical depth and at the same time it can reduce the lower cost of ownership especially with modernizing Windows workload. It s it saves you from uh commercial licensing cost right that's why customers should modernize because it because it unreveals the business benefit fits right which they did not had it earlier um now when I think about Microsoft workload modernization right customer ask me these questions like what do you see in Microsoft workload modernization right typically taking your windows-based application and transferring into a cloud cloud native environment right so so that it can run on cloud native infrastructure so that is what uh is modernization of windows workload so I can give you some practical example >> so basically taking your net framework application and converting into net core so now it can run from windows to Linux environment or taking your monolithic application and breaking down and converting into micros service uh application or you can also have um your application which is written right now in in any form like maybe containerize that using docker or kubernetes and then run it onto ECS or EKS that's also modernization of application or you have an um applications which is like web- based application take that and run onto managed service like AWS app runner right so this kind of an these are the examples of application modernization right now let's talk about database modernization, right? So with database modernization uh you can take um the SQL server database and now convert into different engine like Postgress SQL open source run it on Amazon Aurora for Postgress SQL. uh you might say this is like very challenging because you are completely changing your database platform and it's not an easy one like you have like years of store procedures and complex code developed and all these things but actually when you move to Aurora postgress SQL you get an advantage is like um cost savings so you get really high performance when you move to Amazon Aurora for postgress SQL so those are the um couple of I just covered the major things you can again modernize like file system and and active directory managed things like that but this is like a high level uh that's what customer consider modernization >> how about challenges like when when you talk to customers about modernization and what are the different challenges that they face for modernization >> so when we think about modernization business wants to do modernization um it technical people they want to do modernization but there are some real challenges they face right so one of them is database modernization challenge I mentioned about moving from SQL server to postgress SQL it's not an easy change right uh so because you are completely changing the engine uh so customers have been trying to run SQL server for many years so I've been managed SQL server from last like 25 plus years right so if the customers has been running SQL server for many years they have this um this complex store procedures, right? Complex business logic. Uh they have this link server and then they have um um business logic that no one even remembers what it is now, right? So to converting that it's not only a technical challenge but it's also a business uh conversation as well, right? Because the everything has to be converted now. So that is one of the challenge is taking the database as it is with like years of legacy development now converting that into another database platform and what are the tools I don't have an expertise I don't have a tools I don't have a timeline I still want to keep my existing application as it is running and now you are telling me to modernize like how I'm going to balance that so I think that is one of the challenge customer face and second one I see is with application modernization. Let's say Microsoft.NET application, right? So, uh if you're running a typical Microsoft.NET framework application and that application you are trying to convert, uh it's not designed to run for cloud. So, it's not going to run as it is because it's looking for that legacy DL or legacy x86 um executable or it has some internet explorer integration, right? or some older Windows component. So even though customer wanted to convert and modernize that legacy net application the challenge is here they don't have a better tools they don't have a um the that timeline is so long and the same thing the skill set is as well right so those applications has been written like 20 years ago by someone who doesn't is not there anymore in the company right so those can be real challenge and the other one I think is a people challenge as well. So, Windows administration or doing and working on a Windows system needs different skill set. Now, you are asking those Windows admin people or Windows um users that now you have to um learn container, serverless or lambda things like that. So finding a right balance between the skill set like getting the windows skills at the same time getting the cloud skills it's a challenge as well. >> So actually one of the challenges I've seen is just business leaders in general uh they they they don't want to uh change something that is already working. So when you talk to them about modernization they're like why would I even change this if it's already working right? So uh that's another challenge I've seen. U now um how do you overcome these challenges? Uh great question right so uh glad I'm here uh working with at AWS where we have a great tools to help customer um modernizing and overcoming those challenges. So let's talk about the first one with database modernization challenge right um I think earlier in 2016 AWS released a service called um um AWS uh schema conversion tool right so with we call it as an SCT uh with schema conversion tool it's an standalone tool that you can download and you can point it against your database schema and it's going to generate and report and at the same time whatever it could convert to postgris SQL it would do a conversation Typically we see like 70 to 80% success with this tool but then remaining work you have to do manually convert especially if you have a complex store procedures using a lot of SQL native feature right so that was the tool that uh customers started using it and then when you have to migrate the data we have an AWS database migration service um that can do heterogenous migration as well so migrating from SQL server to Postgress SQL in this case um customer use that AWS DMS tool. Um so until now recently I think last year we uh announced um um AWS DMS capability. It's a generative AI based capability to do um modernization. Uh so basically use the schema whatever was schema conversion tool was doing. Um instead of doing that like this functionality is now part of the AWS DMS. It's a generative AI tool that can take your SQL server code and convert into Postgress SQL. So earlier the code snippet which were not converted um and left for manual effort now the AI based tool uh uses that LLM and use runs get Amazon beddrop and then it does the conversation basically. So this tool um AWSDMS is um SCT is still there. It's like it's still like you can still download it's like an agent based it's executable based tool but the DMS is like a web based tool uh console experience and um you can point your database source and target and it's going to do the conversation for you. So what I'm trying to say here is AWS has a tools and it's using generative AI to help you accelerate your modernization for database right so to solve the DB modernization challenge this is the one thing that we are doing and second thing about application modernization now when I think about Microsoft application is like obviously net modernization right so lot of customers still has application written in net framework. Actually when I think about my career journey when I started at that time it was net 1.1 right and over the last two decades it's been evolved and last year November Microsoft released net 8 LTS uh version. So from net 1.1 which runs on net framework um Windows operating system uh and going to net 8 LTS which is like a crossplatform.net net it can run on Windows and Linux like any platform. So this overall journey customer was still fine developing using net framework or or even net core. Uh but the issue is that um as I mentioned net framework only runs on Windows. So when you modernize to net core um you can take few advantages basically number one is like now you can run on Linux get up to 40% cost savings of licensing right uh second thing is um you can get by default like 1.5 to two times better performance without making your application code changes you can get better performance uh um then third thing is uh you can get 50% better scalability right so all these benefits um are good for customer. So customers started to modernize that on their own right. So they started making looking at the code, making the changes, analyzing the dependencies, looking at the nugget packages, looking at the APIs and transforming the code, uh validating the code, porting again and doing this in a cycle. But this whole press uh process is errorprone and repeatable. Um do you have only one application in your enterprise? >> Oh no. Yeah. So enterprise customers has hundreds of application right. So in that case doing this process again and again um was very timeconuming and some customer never even ended that or they just left in the middle. So to help our customers um last year reinvent we came up with a service um Amazon q developer transformation for net and that allowed to port your net framework to net 8 core um and it can do for at scale uh so that your larger application uh footprint can be converted and earlier this year in May of 2025 we released the AWS transform.net net service. So with this service um the advantage that we have if we make that part uh completely separate and we make it for developers uh to give better experience to use themsel right so we have two experiences uh one of them is a web experience with web experience like uh where do you store your repository >> github gitlab >> github gitlab yeah exactly so you point the web experience to your code repository like github gitlab bitbucket create um Azure DevOps and uh it's doing the scan of that all repositories looking for the packages looking for all the dependency interdependency between the projects and solutions and it's going to port all of those applications if it's applicable to net 8 right >> so so question this as transfer for net is it what is the cost for this >> um it's completely free actually so when what it does it it does this assessment does the transformation it does the assessment and it also does the validation and it generate the whole report. It also generate the Linux readiness report as well to see whether now you can take this whole application and directly port to Linux and run on ECS or EKS. Right? So with AWS transform uh that's a web experience and we also have ID experience. So if you are an individual developer and if you just have one solution and if you want to port that you can just use an ID environment and then port that application to net core. So basically as I mentioned it's it's completely free of cost. Um and um >> how about any limitations? Uh >> so with with this tool actually uh you can um uh transform 1 million line of code per month >> at no cost. >> Per month at no cost. >> Got it. >> Yeah. And I I also get a lot of questions about what kind of uh version supported. So like the source is net 3.1 um um and then destination is net 8. uh and you can go um uh why not net 9 uh because net 9 is not an LTS yet so we only look at the LTS like long-term supportability net version >> cool um I know you covered uh the tooling for modernization I want to also talk about a little bit like the people aspect of it right like how do you help people when they are going through the modernization and how AWS can help with that so I want to talk about a little bit just about the EBA or experience-based acceleration for modernization. Uh basically it's a program where uh you know we work with customers on uh you know who want to modernize right and uh we we we ask the customers to go and pick that one application that you want to modernize right of course there's so many of them but just pick one of them and also we we ask them to gather different teams who are involved in doing that practice right basically you may have bring in people who are from the the actual developers who wrote the code from security teams, from leadership, decision makers, uh maybe even a partner if it's involved and AWS experts all in one room and over normally it's a couple of days uh program where uh we help customers uh customers modernize that one application and then all the learning they take that and they apply to other uh other applications which they are planning to modernize and we've been we've been seeing very success with this program >> actually I have seen uh success uccessful customer modernization using EBA right so um I can also share another customer story actually uh this is one of the customer that I I'm working with them like thms and routers uh they had their legacy application um and it was working perfectly fine right so it's not that there was any issue with the application but it was written in net framework and when they decide uh to modernize uh and to port that to net core uh it was taking them time of like they had like a months of project plan. Um it was difficult to scale and it was um u basically hindering the innovations because they wanted to move out fast and because they were doing all this manual work um they were trying trying to find some automation. So they also looked at other cloud provider and also uh tried their tools but that were not helping modernize at scale right. So when they started working with AWS transform.net net um they see the real use case of the tool right so um they see with AWS transform fornet within first month they modernize 1.5 million lines of code right uh giving them four times more velocity uh they were able to save 50% of the cost from the technical depth because of moving from .NET framework to NET core and they were able to reduce the deployment like overall conversation time um the application modernization time from months to weeks now in in a sprints right so that way the customer actually still modernizing they are not done they are still modernizing like 1.5 million lines of code every month right so if you what I would like to say like if you have a large doesn't matter where you stand right now in modernize and journey whether you have a large application code base or if you have smaller applications but at one day you need to work on it to get modernized right >> so what I'm saying that just try it once so that you know where do you stand right now like at least do an assessment and get the assessment report done it's it's free of free of cost right so it's nothing hurts to do an assessment >> y cool thank you >> I would like to ask uh I think we covered the migration modernization operations now lastly can you cover the cost optimization like how do customers save cost for Microsoft workload >> definitely and you know cost is a topic which we couldn't slip skip right so uh one thing is over the past couple of years which we've been running Microsoft workloads on AWS and uh we have helped customers uh you know optimizing their costs uh through different methods really if I wanted to list all of them it would take a whole day maybe uh but what we did We actually captured all those different techniques, services and uh things that customers can do to reduce their cost and optimize their cost while keeping their performance high into a library called um Microsoft on AWS cost optimization short for macro which is available on AWS prescriptive guidance and different type of workload. It has different recommendation but there are a couple of them which I want to highlight here. The first one is uh OA which we talked about right which uh we we talked about it in the concept of migration where during the assess phase you want to know what you have and use the OA but actually OA can be also done for your existing workload. So if you already migrated and you have workloads running on AWS you can still run OLA against those workloads and optimize for both u for both uh performance and cost. And uh then uh the other uh uh service which I want to highlight here is AWS compute optimizer. So uh you know uh a lot of times when uh you uh you decide on what instance type your instance needs to use you come up with something you estimate something right but you wouldn't know exactly what is the exact optimized instance type unless your workload is under load right but even after uh you know although you look at the monitoring and performance metrics uh still uh you need a better tool to do that right so as custom computer optimizer when you enable It looks at your performance metrics. It looks at your logs if you enable those and using AI, it learns your application behavior and those instances behavior over time and after a while it gives you suggestions of what instance types you can switch your instances to to save on to keep the performance but also reducing cost. So the last uh the next one that I want to talk about is AWS license manager license switching feature. So license manager itself it allows you to track and enforce licenses for Microsoftbased licenses or other third some of the other third parties. But one of the features is its license switching feature which allows you to switch licenses between bring your own license or license included. So uh think about a scenario where maybe you have some licenses being freed up somewhere else in your environment and you want to bring them to AWS. you can actually you don't have to rebuild those workloads. You can h uh use AWS license manager to switch from license included to BYU or vice versa for your workloads. So Yogi, we are almost at time. Uh what is the what are the takeaways uh from this session? >> All right, let's wrap up this by some takeaway summary messages. First of all, I would like uh you to don't skimp on the planning, right? Doing that optimizing and licensing assessment is a key. It will help you rightsize um your environment. It's going to basically like give you a health check on your whole infrastructure and save cost. >> Second thing is plan your strategy wisely. Not one sizefit all approach, right? So certain things you can do left and sift migration and then certain application and database you can modernize it right. So plan your waves, group your application and database and plan both your waves accordingly uh based on your business need. So that's second thing and third thing is um we talked about so many great tools and options that you have right uh so go based on what you need right you can also use native tools or you can also use uh different AWS tools so determine your requirement determine your skill set and based on your business need and skill set go tool that your team is most experienced with I would say go with that and um fourth thing is about don't rush for the modernization right modernization is not an overnight journey. It's not going to take you from like uh from here to next destination in a day, right? It takes it takes little bit time, right? It's it's a strategical process that you have to go through. But at least start thinking about it. Start thinking now. If you have not started thinking about it, you should start thinking about modernization right now. What kind of benefits it's going to aail you and what good things is waiting for you in cloud. Right? If you have not tried the AWS transform uh forn net and if you have a legacy workload I would say try that out. There's nothing wrong in trying it's a free tool for you. Uh you can run an assessment and look at the report and see how your environment looks like. At least it gives you an idea where do you stand right now and uh lastly uh we have a programs and tools to help you right. So uh as sas as you mentioned about the EBAs um we can always start with an PC application and do an EBA uh and then the whole AWS experts team is with there to help you get started with your modernization and migration journey. Uh we also have programs like um map uh migration acceleration program. Uh we have a modernization acceleration program. We have a DB freedom. We have a different programs that can help you give some funding as well that can help you ease away some cost uh for migration. We have plenty of partners who are also um who also has an migration and modernize and competency. They can also help you get started with this journey. Right? So what I would say u doesn't matter where do you stand in your journey whether you already started your migration and modernization journey or you are just taking a step into it. It's never too late. I said let's just start now and then you will see what benefits at then you will gain when you move to the cloud. >> Yogi I know we're at time you you just talked uh touched on map. Can you talk about map a little bit? >> So map is an AWS program uh that can help you accelerate your migration. Basically it gives you an funding um and resources and methodology that you can use to uh migrate to AWS. >> Thank you so much Yogi. I really appreciate all the information and I want to thank everyone for joining this podcast. We hope you found this informative and thanks for watching. >> Thank you.