...

Transcript

Walter

00:03 Hi everyone, welcome to the latest edition of Walter’s World, the podcast series from Astadia. My name is Walter Sweat and I’m the Astadia CTO. For those of you who don't know, Astadia is a global service provider, a company that focuses on mainframe migration and modernization.

00:25 Today I'm delighted to have one of my friends, Steve Read, who's the Principal Program Manager for Microsoft joining us. Steve is working with cloud migrations for organizations and I am delighted to have you here with us today Steve.

Steve Read

00:41 It’s a pleasure to be here, Walter.

Walter

00:43 It's always fun to get to talk to you.

Steve

00:45 Thank you, same here.

Walter

00:47 Thanks so much. Steve, can you describe for the audience what you and your team do at Microsoft?

Steve

00:54 Absolutely. My team is known as global engineering and it's a pretty large group within Microsoft, we actually are a part of Azure Engineering.

01:07 The team that I am on is tasked with going after what I would call the strange and unusual workloads that did not run in a cloud platform, that we collectively refer to those as legacy. So, you know, a good example is IBM mainframes z/os Unisys mainframe comes to mind all the MID range is that are out there, be its power platform AS400, legacy Unix, these are you know non traditional in that there are non non x86 based and a lot of times what we find in our travels is that, you know, there are folks that actually want to get out of the data center business, but they still have these pesky mainframes and these legacy systems and a lot of the time they don't even know that they can run those workloads in Azure.

02:02 So what my team does is tasked with is helping those customers and, as you know, this is a partner-based echosystem that makes this all possible, so we really get involved with working with the partners to make sure that these workloads and their solutions and how they architect this on Azure are the best, and that these workloads run as well as possible.

(02:30) So we get involved in a lot of tuning and things like that, so basically helping customers and partners on their cloud transformation journey from legacy to cloud, that's what we did.

Walter

02:43 Ok, fantastic, thank you, Sir. Steve, you are one of the first people that I knew to talk about digital transformation, quite a few years back. Could you share your thoughts about, you know, because I think people consider this as digital transformation is it really an actual possibility for people who are running on the mainframe today?

Steve

03:05 Yeah, great question. It is and I love when we get in conversations with customers and they had no idea that they thought that the mainframe was going to be, you know, with them for some very, very, very long time, and that they were going to have to always have some kind of a footprint on prem and that they could build these elaborate cloud architectures, cloud native architectures that we're at some form or fashion, going to have to integrate with the mainframe and quite frankly, you know, that's in my before my current role, one of my previous roles of Microsoft, that's what I did was integrate you know a lot of the web based architectures out there with the mainframe.

03:54 That's a very mature business, but you know with digital transformation and the possibilities that are out there, it's entirely possible to take all those workloads on the mainframe and move them to a cloud platform, and you know I know what we're going to get into in a little bit, but you know what the just start peeling the onion on the benefits of when we start doing that right so it's just you know.

04:18 Obviously there's cost reductions that come into play, but you know just the cloud native and the power and the you know the feature functionality of the cloud it's really start to make that a very compelling path for customers.

Walter

04:33 But to me that's the really exciting part, the fact that people can really do things differently than they've been doing them for the last 2030 or some cases 40 years, being able to leverage cloud native so that they can be more responsive to changing business needs. It makes it to me very, very exciting that people can now have more options than they've ever had before.

Steve

04:58 Absolutely, absolutely. A good example that comes to mind when you're mentioning that, that I find to be a big compelling driver is DevOps. When I got into this, I really did you know, to me, I was like oh you're going to want to do this because it's a lot less expensive, typically 10x less expensive as far as you know, the month over month cost of writing something and in Azure as compared to you know, on prem, for you know, a given mainframe, but what we're finding is that DevOps is becoming one of the primary drivers in you know, I know you and me, being old CICS developers, you know, we had TSI SPF right which was the state of the art back in the 80s.

(05:48) I actually still find I still find customers that are still using that. And when you contrast that and talk about what you can do with the DevOps pipeline it really starts to it really starts to open their eyes, as far as why you know that's a big compelling reason to actually do this.

Walter

06:05 Absolutely, we both have been through the scenario we had those reams of books behind your desk we had to go through it and figure out what to do, the flexibility that devops reports people today is just monumental.

06:21 Absolutely.

Walter

06:23 Talking about modernization from a mainframe perspective, and I know different people look at it differently, what did you consider application modernization to really mean, then, is it just devops or what are the components that you see people taking advantage of.

Steve

06:41 yeah that that another great question because it is a very overloaded term that means different things to different people. In in my world there's the well known, five or six ours slide that somebody made you know, however, many years ago and I, you know I won't bore you or your viewers with that because they probably know what it is. I think what we like to do when we talk about application modernization, we like to understand what that means to the customer.

07:22 A lot of times what it means is you might have a financial system, this is very, very popular and I'm sure it probably resonates with a lot of folks watching this that. You have a financial package once upon a time that was written on a mainframe, and a lot of times the best landing zone for that is SAP or some kind of cots package right and so that's modernization that's running in, it's running in Azure you know it's really more of a data poor or you know, but that's one. The other is I've been involved with some you know, especially it's very popular and vince have to do a rewrite so let's just take this thing it was state of the art and kept the lights on, paid the bills, for you know 25-30 years it's time to rewrite it in something like Kafka or you know, one of these modern architectures that are out there, so that's definitely one, the other is, I just want to run in the cloud, I want to take what I have, I want to bring it, I don't really want to change that much, and I still have some COBOL developers, so I want to you know evaluate that estate, I just want it to be less expensive and so you know, on the other side of the spectrum it's rehost, right, so less risk, you know I’m pretty much going to always like to say with something like that people go home on a Friday, they come back on Monday and the users don't know that they're actually running in Azure.

08:45 Right, it's a it's a magic trick.

08:50 And then, and then the other one that I'm finding more and more is, you know, alright, so I'm losing my COBOL developers and there's not a plethora you know a renewable resource as they used to be back in the 80s and 90s, you know there's less than less COBOL developers and there's more and more Java and C# developers.

09:10 I was just talking to a customer the other day, and they were like look I don't want you if I'm going to do this, I don't want to tell my developers that they're still going to be developing and COBOL when they all had to learn it and their native language is Java or C# so I want to be in a modern language, I want to have devops, that's okay to keep the existing code base, but I don't want it to become all, I want to convert that COBOL to you know Java or.net or something like that so.

09:42 With the way of answering your question, I think it's you know it's all of that right, it's a spectrum and it's like we got to plug in and go all right, what exactly are you looking look into you know, to get out of this modernization journey.

Walter

09:57 I could not agree more, you know we talk all the time that there is no one perfect answer for every possible people organization that's looking to move from the mainframe because no two mainframe environments, are the same no two pain points are the same, and just trying to get people to agree on what those five or six are is really mean is challenging enough but providing that flexibility absolutely is key.

10:22 Yeah, couldn't agree more.

Walter

10:24 Steve, I know we've had the opportunities over the years to work together in both public and private sector opportunities, do you feel that there are any unique concerns that maybe state and federal government face that companies outside of the government don't face?

Steve

10:45 Absolutely, absolutely, the number one I mean, obviously with last year and 2020 covid 19 and everything it created a big, a lot of interest let's just put it that way. I at Microsoft focus a lot on state local because I used to work in that industry, before Microsoft actually and actually covering it when I was in Microsoft or in the in a sales role, and I can tell you that you know they, there are some a couple of stories that made the press that we probably all I mean your audience could go back and remember the one that I think that comes to mind for me, was in New Jersey, which is kind of like the poster child of you know, these unemployment insurance applications that were out there that you know let's face it they've been great, they've been running they they've been doing what they're supposed to do that, all of a sudden, we had this you know additional load that they were never really engineered to handle.

11:49 And that's when you know you started making always got to make the paper for good reasons or bad reasons, and this is for bad reasons.

11:59 So, you know that the obvious thing there is that's what I would argue that's one of the founding principles of the cloud right, was to address those types of workloads that you know they're run in and then it's like they're supposed to and then Christmas happens or some kind of event happens, and we need to throw additional capacity computed that.

12:18 And that's not that easy to do with the mainframe, so I find that that is one of the biggest drivers where folks actually you know, it wasn't a problem before, but then, when that started happening customers started reaching out to us and going, you know what can I do to solve this type of a problem, so that consider also with budgetary constraints is you know these things, what they're paying for these mainframes and on prem with the the facilities and everything like that it's not cheap right, and so you can save money, and you have a very compelling story to mitigate the risk and save money that's a that's another driver I find for state and local government and federal customers.

13:09 And then the other thing too is they have part that almost all these customers have a foot in both worlds right, so they have a foot in the cloud world, they have a foot in the on prem world. But most of their data and their Ai and their analytics is in the cloud. So if we can do anything to essentially move those workloads closer to their data lakes and there are places where they you know they you know try to predict the future or you know do, what do it information management is all about, then, then they realize that that's a good thing, so I find this to be the main drivers that and, of course, like I said before, you know there's probably there's not a never ending never ending supply of COBOL programers.

Walter

13:59 That has been proven for sure.

Steve

14:01 I think there's less and less each year, it's a great language you know and it's been around for a long, long time, but I think you know object oriented languages, you know modern languages just have some incredible advantages.

Walter

14:16 Without a doubt. You mentioned the cloud just in general, before I like to ask you specifically about Azure, ss mainframe customers start to look to the cloud, why Azure, what are some of the advantages that they would find Azure to provide specifically?

Steve Read

14:36 Yeah, another good question, and what the first response that I have to that is before it was not even really related to mainframe as much, as one of the reasons Microsoft has had a really good story when it comes to cloud is because of the great work that's happened with Microsoft in Office in 365, and there's a lot of customers out there that have already made that journey to 365 and I wouldn't say, most of them are happy, they realize the benefit,  Office is just a very mature great you know solution, but the price of admission to get there is you have to you have to do the you know the foundation work and that typically involves taking your identity stores your you know everything that was running your typical you're you know you're on premises footprint in the form of that active directory typically I mean I can direct it was probably the biggest one and you had to move that to azure right to Azure active directory, so you had to you had to migrate that identity up there, and when you did you actually sort of created a conduct, a pipeline, so that now all this identity that you, you know it's it's basically the crown jewels you could argue of your IT, is now seamlessly integrated with cloud platforms right a cloud platform like Azure. I mean you can integrate with anything with that, I mean Azure active directory is just got such a great integration story.

16:26 And so, having said that, the mainframe plugs into that right, so in the mainframe we have rack up top secret things like that, and you know, thanks to the great work done by you know the partners that seamlessly integrates with you know typically something like Azure active directory a lot of times you know books have their own sort of variation of the old APP or what might be out there, but these you know that this in and of themselves also integrate with Azure active directory so I guess it's a long winded way of saying that one of the most compelling reasons that I found you know mainframe or not just you know for from a cloud migration standpoint or modernization standpoint is the identity story that Microsoft has had and it's very strong and I find that it's kind of a wind in the sales for a lot of these transformation journeys.

Walter

17:19 It's really a cohesive point for being able to look at moving to the cloud and being able to leverage that, so I can see where that is very, very important.

17:30 Steve, in terms of technologies, we hear a lot now about containerization, I'd like to get your viewpoint on you know, is containerization a goal that you feel most mainframe companies could and should explore or their reasons that they perhaps should not explore that.

Steve

17:50 Yeah that's a great question that I have an opinion that's maybe not that popular. Containerization at the end of the day is a deployment model to me and it offers some inherent benefits, but with the price of some additional complexity. It solves some very key problems, namely scalability and things like that, and just you know it manages itself so it's very from that standpoint, if you have workloads where you're going to benefit from that, then you know containerization is a great technology. What I find is that you know I kind of view this as a journey containerization is a is a sort of a step towards microservices right, where you get into like truly decomposing monoliths that are out there, and you know containerization is a good first step. I you know you can go into a vm take your your monolith and go into a vm. that's kind of a you know that's been done there's nothing new there that's been done, as you well know, that's been done for like last 30 years probably it's been very popular.

19:08 And then containerization is taking that a step further, so that now we can actually you know reap a little bit more of the cloud benefits, and I would argue that the next step is really going into you know true microservices where you have you know you're  now in a micro services architecture and you've taken that monolith hopefully you know the keeping the core business logic intact.

19:31 But you know, in a truly more cloud native, the more you can go cloud native I think you know towards that that journey, then the better your company's going to be with that solution that workload because you're going to be more agile and have the ability to respond to you know things that go bump in the night and cause issues.

Walter

19:56 Not surprising that I feel you were spot on with that, I kind of worry sometimes that people look at containerization just as a panacea it's a cure all and it is absolutely a great step. But it isn't right, in my opinion, for every application I think people need to kind of look at what their own environment is like and determine where that benefit comes from. Like you, I think that microservices looking at doing things differently is the real key to be able to leverage that you know deciding if I have a bad versus an online application, where, can I start to pull those pieces out to make them more viable and available that's every bit as important as just saying what here's containerization let's go tracing.

Steve

20:45 Yeah, you know I always thought that technology and just in general is very similar to the fashion industry and containers are like you know turtlenecks are in everyone wants to do containers and a lot of times you'll have a conversation with by car what, what do you really want to do, and you maybe sometimes it's a good fit, just like you're saying, and maybe sometimes there's a better price but.

Walter

21:09 It's important to have those conversations, you are right.

21:14 Steve, I’d like to get your opinion, cause I think this is important to organizations, you know the mainframes it's great environment we both have a long history with it and it has worked well for an awful long time as organizations look to move to the cloud, how secure you feel that they really can be about the security levels they can see moving from the mainframe to the cloud?

Steve

21:41 Yeah, I get that question a lot and you know, once you know the details, but we have some really good collateral and things that we can talk about as far as how Azure data centers are set up, the levels of monitoring within Azure, the security ring fields that we put around Azure to make it secure. Once you start going through that, then you start to realize that you probably could never architect an on prem data center or a mainframe to be as secure as a cloud platform like Azure right because there are so many levels of course there's nothing stopping anybody from you know, create the vn and putting some public endpoints on it right, I mean. You can you can definitely you know for loaded gun, you can be a lot of damage, if you don't know what you're doing but. And that's where my team comes in, we have you know expertise around that so that, you know, we get that question a lot like hey this is, you know hey this has never been hacked this mainframe never been hacked and we don't want it to be hacked in when we run it an Azure and.

22:57 And so, our job is to really come in and talk about the layers of security that we put on top of that, so that you're not going to get hacked and then, and then you couple that with one of the things that you know, so we can we can make these workloads as secure or more secure actually in Azure one of the things that we can do that is really not something that the mainframe is really good at is, we can also our bcdr story or business continuity and disaster recovery story is so much better.

23:31 I'm sure you've seen this will have a conversation about what's your RPO and they'll talk about days, well, we got to be up and you know we got a two day window, I mean wow. We can come back and say you know we'll have you back up and 30 minutes you know I mean you're not you know if you do go down you're not you're not going to lose any transactions and and when we start coupling that with security, it just becomes you know, like I said very, very compelling story.

Walter

24:03 I have always said, if you if you treat the maintenance of the cloud, with the same intensity that you treat the maintenance of the mainframe and you really utilize the tools that are available, you absolutely can see at least the same level of security as you say, if not more worried about on the other hand, what about scalability and performance?

24:28 From your experiences are people seeing equal to or better performance and the cloud and what about the story about scalability as opposed to what they see on the mainframe?

Steve

24:40 Oh absolutely, absolutely. I probably shouldn't mention there's some I actually helped some customers on some benchmarks, we have something called Z ref.Aand there's one that's out there, that is, you know I think pretty well known, and I think you know my numbers might not be completely spot on but on one single vm which I believe was a 32 a machine they weren't able to get to something crazy like 28,000 minutes. It might have been 64 and a half to be out to be back and like if you just look at the rough benchmark and you'll see it so from up and that's on one vm so and that's a huge bank that's a globally financial institution so. So the story, as far as the capacity of the cloud is when I really even talk too much we still talk about it a little bit, but it was like when we were first when I first joined this team that was like well, this this thing it's fast going to be as fast as my mainframe and when you show them what you can do, then it's pretty it's you know it's a pretty good story, I think, from a scalability you know, one of the benefits of the cloud is you know, obviously we have the ability to satisfy the daily you know routine running of an application a workload and then, when something like a you know again like something like coven hits or we have like extreme cases where unemployment insurance or you know that this system that was only used to you know certain amount of usage all sudden at three x right.

26:26 In Azure it's just so easy to say, give me I need three x compute boom in there you go so and then, and then the flip side of that of that benefit is probably the most compelling which is and when I'm done with my high watermark a good back down to where I was.

Walter

26:47 That is key without a doubt.

Steve

26:49 And I didn't have to know servers had to die in the process, I didn't have to throw up a bunch of hardware to do that, or you know get a bunch of different cpus from IBM to make that happen, I just added the capacity got through, and then I took the past the way.

27:05 And I think most people a lot of folks get get that that that scalability that elasticity of the cloud is is is huge it's huge benefit.

Walter

27:18 Something that I've always thought is important, data. From your experiences, what can organizations expect in terms of what they can do with their data once it's off of the mainframe.

Steve

27:31 What can they do with it? Yeah, that's a huge one I mean the story, as far as what you can do with your data I always say like you know you what you want to do with your data is you wanna, I was at a hospital once and I remember one of the things, and this is on an IBM mainframe and I could we could tell you what the census was in the hospital and we could probably tell you what the census was going to be, you know over the next day or so and they had the plan nurse nursing, you know resources, and you know everything that goes into that right when you have patients, but what they wanted to know, was what it was going to be like next week, right, sort of predicting the future, and how can you do that, well, we couldn't back then, this is the early 90s.

28:20 You know everything that we did we would you know, back then, we would download it into something like excel or something like that, and we do some kind of regression and we go out it's probably going to be something like this, but you know as.

28:33 Well, that scenarios now actually we can we can actually give you a pretty good idea what it's like so we're always say we want to get closer to predicting you know what what's  it going to be like.

28:46 I can maybe tell you what's going to be like tomorrow, the next day what's it going to be like seven days from now it's kinda like the weather right. And now we're really good with that, and the reason is because of the data, and so we can take that data and put it into you know an analytics platform like Azure we can we can do crazy stuff with it and that's one of the things that we can do, and the other one is I've seen it and the one that sort of is familiar to me is fraud management right where we want to you know, we have all this data, we want to be able to look at this and go how how much this is is real and how much, how do I know there's not some fraudulent activities going on, and when we have the ability to analyze the data in real time like we can in the cloud, we can start to analyze those and find those and that's just you know amazing to be able to do that because we obviously cut down on waste which is really, really important, especially in government.

Walter Sweat

29:50 Just being able to look at data as a different tool that's the exciting part to me. You know, getting compute off it's great and oftentimes that's monetary you see the benefits, but the benefits of working with your data in different ways that's the real differentiator I've always don't.

Steve

30:09 Completely agree.

Walter

30:11 Steve, I think you know, there is no magic button that you push where these things just automatically happen obviously that pastor opinion because I think this is important to everyone is they consider their alternatives. From an architectural viewpoint, what are the biggest challenges that you feel people should be aware of, as they contemplate a migration from the mainframe.

Steve

Yeah and that's what my team actually really focuses on, you know. A good examples things that I've been involved with is mainframes are really, really good at processing transactions really fast.

30:53 You know they you know this day I mean so it's the transaction throughput for these reservation systems are these you know these medical systems, you know they have the process, you know, thousands, hundreds of thousands, maybe you know millions of records they're really good at that, and you know mainframe imprints typically know composed of two types of workloads online and batch right.

31:21 And so, some of the one of the things where we really sort of had to you know thread on them and get into the weeds and figure out how are we going to match this right, how are we going to match it's really impressive transaction processing speed, and so the I wouldn't say the biggest architectural challenge that we really you know, have worked with and have solved, I should say we we've there's one where you know we started out, I think we were 10 X slower than a mainframe batch window, and then we actually just from you know, working on it, you know the team here at Microsoft, we had the unless the data ninja team, which is great, they got it to faster the mainframe right, but it was tinny and we couldn't change any of the code it really came down to how do we, how do we make sure that the database performs as well as possible, and how do we mitigate latency right so when we have processes talking to another process in the cloud, one of the benefits is it's really it's really easy to spin those things up, but we got to make sure that when we do spin them up that there's close as possible and that the latency is next to nothing.

32:34 That I find to be latency transactions being things like that, if you just come in and you and you, you know what we are doing our team is we're documenting best practices, as we do this with reference architectures like hey if you're going to do this, you need to make sure that you enlist that you know these types of features and functionality that are native to Azure to make sure your workloads are performing as well as possible, so we do that with customers and then obviously with partners and partners solutions that are out there, we work with them on tuning in hey let's make sure that this this thing is going to run faster than a premise that's always good for that, but, but I do find that you know if you don't know which cloud services to leverage, to decrease that latency and that data throughput and you know you could end up with a non performance type of workload.

Walter

33:28 And it's important for people to understand that it can be done it just takes the work right.

Steve

33:33 Yeah, it does it takes it takes a wisdom and know how and been there, done that.

Walter

33:41Some battle scars.

Steve

33:43 Yeah you have to you have to know what to look for.

Walter

33:46 That’s right. Well Steve it looks like we have unfortunately come to the end of our time for today's session. First, I want to thank you for taking the time out of your busy day to join us it's been a delight, and I just like to throw this out if people wanted to learn more about what you and your team do, how should they best reach out to you.

Steve

34:07 Thank you for that question and based on your guidance, if you look at my picture there, you can see the email address legacy2azure@microsoft.com, that is, the alias for my team, we all, everyone on my team is basically get that email, and let us know you know what your question is, and let us know how we can help.

Walter

34:31 Fantastic! Again Steve, thank you so very much.

Steve

34:34 Walter thank you.

Walter

34:35 It's a pleasure and for the members of the audience, thank you once again for taking the time out of your busy days to join us for the Walter World podcast. If you want to keep up with upcoming ones that we have and we have some exciting ones in the future, please just go to www.astadia.com and check out our podcast series. Thank you all so much and hope you have a wonderful rest of your day.