From Legacy to Cloud: Success stories from migrating mission-critical applications

why don't you you know kind of introduce to everybody who is listening to you for the first time uh you know on what do you do right now at amaran what does amaran do and what your role entails yeah sure David and uh so Amarin is a Fortune Finder utility and uh we serve about three and a half million give and take electricity and gas um so basically you know we keep the lights on right so in many ways uh so what I do is I lead Enterprise shared services I'm a senior director there and in my group we have Enterprise architecture and I had a fantastic opportunity to re revamp that particular practice and then resiliency as you know yeah it's very critical for utilities you don't want your gas or electricity down right so resilience is very critical and then I I revamped agile office that is another practice in my group and integration so I do the you know my team does all the plumbing all you know with various systems and applications on-prem in the cloud and various things like that and recently you know I will ask to start looking into the testing practice you know as you know it is very critical for a successful delivery and uh we started looking into that one I'm guessing you should have at least uh at least 400 to 500 people in your team that you're managing right now no I don't we are very efficient so I including everybody I think we're around 170 180 ballpark that is where we are right now well that's a lot of a lot of people to work with too right and and we'll get into all of that when we were speaking last time you mentioned to me and I wasn't aware of this that Amarin is actually a Fortune 500 even though it just operates in uh Michigan and there's another Illinois operate in Missouri and Illinois and yeah that's right and it is surprising right you know that's what I was telling the other day right you know just operating in two states it's still a Fortune Finder company because the number of customers we serve and you know it is both for distribution transmission of both gas and electricity right we also have some generation like you know we have some coal plants which we are working on moving away as well as we have nuclears uh quite a bit of uh gas and sun solar and wind energy generation right wow this is awesome this is great to know how was the transition for you coming from uh like uh you know Monsanto bear to you know a utility company and uh of course the practices that you have to lead or the way you have to build Solutions or work with people kind of you carry that experience but in terms of domain itself what surprised you the most coming into you know Amarin yeah so utility space is uh new for me right we are all like you or anybody else we are all using utilities but we do not know to the most part how they operate other than you pay the bill kind of a thing right uh Utilities in in States at least is a regulated industry meaning a particular utility will have a territory where you have rights to you know distribute electricity or gas or whatnot so with that there is a lot of Regulation comes in so that is there's something new that's a new domain for me along with how energy works and what kind of you know technical as well as Innovation challenges we may have uh some of the critical components right like uh I was talking about reliability affordability and and as well as there is a lot of emphasis on green energy now so those are the things I'm slowly getting up to speed uh it's from technology side that different kind of challenges but at the end of it you know keeping customers what they need and giving them a very reliable affordable service is the bottom line got it all right that's amazing so you already spoke about some of the teams that you're leading of course the testing practice is something that you're going to take over but you were talking about the Enterprise architect team so um maybe uh talk us talk to us about each of these teams for somebody who's listening for the first time what do these functions entail and how do you keep the you know Integrity of what the business objectives are and drive that uh with the teams that you work with sure uh so basically like I was saying you know there are multiple teams in my uh organization and keeping all of these teams in sync is a task on its own right we need to make sure there is a right level of collaboration is going there is a right level of communication going um so Enterprise architecture as you know we have uh you know Enterprise Architects solution Architects domain Architects helping the Enterprise to create road maps create technology strategies as well as you know long-term investment planning and so on so that is what this team primarily focused on we have solution Architects which work they work very closely with these initiatives right you know projects large projects they jump in right from the Inception always you know going into live right you know production and stuff like that so they're very very critical part and uh and again I can go on with the resiliency and agile office and so on but uh going to the just of your question right how do we keep all of these in sync right you know how do we motivate them and making sure that we're delivering business objectives I think that is what we focus very much in um first and foremost my methodology is around you know you've probably heard about this like autonomy Mastery and purpose I think that is what I primarily use that was introduced to me by one one of my leaders in previous worlds so it resonates very well with me and making sure you provide some guard rails and tell them hey this is what success looks like this is what good looks like and create right expectations and goals and you know then don't get into the weeds of what they need to do right you know right right if it starts with why right and explain why we need to do certain thing and then explain them or how it needs to be done in a way that you know it needs to be aligned there is light level of collaboration engagement and so on and then give them autonomy and they make wonders so I have seen it again and again uh it that particular thing works and works on all cylinders that's brilliant yeah I mean in what you kind of shared is also your leadership philosophy which is I mean great for you to kind of share you know I think autonomy to Architects and developers is really interesting especially in today's day and age where they want to be creative and they're looking at solving the problem but at the same time driving guard rails is important because the business objectives is also uh something that you have to drive towards right so that's awesome so uh it's it's a great introduction to what you're doing right now tell me about let's say it back a little bit maybe a decade or more back and how was a young Kishore you know excited to get into this field of engineering like tell us a little bit about your beginning of computer science and what led you to you know move into this direction yeah so I'm still young so I'm still Yankee sure okay I just point to be noted yeah yeah well I'm talking about a 20 year old Kishore right now so uh I I won't bore you with all the different roles in different companies but uh I'll touch some of the highlights uh yeah I had an opportunity to start on a startup in India and we were trying to do a Salesforce kind of a CRM solution it is back in the day and it's very exciting right you know it's a not a Cloud solution but it is a it's a web solution we had a lot of good features and you know a lot of good customization around it um and uh we I learned quite a bit working in that startup right from when how the Venture Capital uh investment works and all nine years of software development and customization sales so that that was really exciting journey and but unfortunately that startup did not survive and that's how I ended up in States and like you were briefly mentioning I started with Monsanto that's a biotech company uh so I was there for about 17 years did various roles um I I also had an opportunity before I moved on to Amber and I had opportunity to lead part of it and I I you know I we co-led it uh transformation which is we used to call it it 2.0 right which is around moving from service based organization to a product based organization waterfall to Agile and then moving to the cloud uh it is fantastic you know fascinating Journey there and also had opportunity you know taking existing uh applications move move them to mode one and mode two and move mode one applications to manage Service Partners uh so that was a lot of learning so basically what excites me at the end of the day is uh problem solving right yeah that is what the key for any uh I.T guy is what kind of problems we solve and how effectively we solve with the different tools we have in our toolbox either it is technology tools uh like right from Cloud to region AI right you know all new stuff coming in we just don't want to use the tool for this just for the sake of it but how do we use it to solve always business problem and make everybody's life more effective and efficient got it yeah so how what's your perspective on all these different changes that have happened like especially I know the cloud kind of burst into the scene and everybody kind of adopted to it but not everybody could move to the cloud immediately because I mean there is data specifications on being in data centers and things like that um did you first encounter you know moving to applications to the cloud at Monsanto uh what was that first project like and how did you like go about looking at something that's new and saying well it needs to be something that we need to be at yeah yeah that's a great question so even though um there is a lot a lot of uh journey in the cloud it is still not that old right um so I think the first exposure to cloud is where we have one particular application which is a legacy application at Mount Center which brings a billion dollars of Revenue so it's a huge application uh in in that terms right yeah and it has thousands of users and it it brings a lot of profit to man Center but it was not scaling to the use because we wanted we wanted to expand it to other countries like it we started in one country we are adding more countries to it so it's adding a lot more users to it right so that is the situation so then we said hey you know why don't we start looking into Cloud that is how we started that like anything else we started with the proof of concept pilot and rolling out gradually um like you know any monolith which is so business critical which brings a lot of uh Revenue to the company a lot of ice on it uh we and and we were looking for agility reliability availability and cost Effectiveness so those are some of the principles we were looking uh obviously Cloud checks all of those right you know there is like agility because you don't have to wait for your infrastructure Department to set up servers which may take its own sweet time uh it has a lot of resiliency because you know you can have multiple uh regions or multiple ways of doing it um I'm not suggesting all applications need multi-region but you have different ways of uh resiliency and availability again you can scale up much easier uh horizontally or vertically and then cost effective uh for sure it is cost effective in short term and there are some questions about long term that is a conversation for another day right yeah so that is the journey we started with and we uh slowly what we did is we migrated one piece at a time from monolith Legacy application and then we did a lot of data synchronization uh during this trans transformation people had to log into multiple systems to get their user Journey finished but eventually after a couple of years it's all moved to the cloud we get a lot of rewards for doing it and there's a fantastic user experience and you know there are a lot of Kudos and everything it is very well uh you know recognized initiatives so right then what we did is took that and replicated multiple times so because we have a path for success and how do we best customize it for future needs yeah so what I'm curious to know about is a one billion dollar project uh that's on Legacy it's a critical piece why did you guys decide that you need to move that one I'm I'm assuming that you wanted to take on a challenge of looking at the most complex problem first and if we could do it then it can be replicated more easily other place was that the thought process but that's a great question even though I said uh that is the project I'm more excited about but that is not where we started right so we started uh with uh internal applications first okay and we we did the pilots and uh uh proof of Concepts like I was talking about and then we ventured into this one but uh even when we ventured into uh this particular Mega project quote unquote uh we started with again doing a pilot and we're doing a POC and make sure it fits our needs what are the different Frameworks we need to use what are the tool sets we need to use and then uh expand in a agile way right you know you don't want to do a big bang you do one or you create a road map You Do release One release to release three kind of a thing and continuous delivery yeah and this is an interesting point of view that you're bringing that you know you are in a position where you also not just have to drive strategy but you also have to ensure that somebody has a knowledge to move these and when the cloud really came about at that time you don't have Engineers sitting there who have certification nobody's AWS certified or Google even I don't think a certification existed at that time I mean I'll have to go and check but you are actually actively learning and applying so you have this trial and error and then you have yourself before how's the project moving and so how did you tackle all of that being in that position yeah yeah yeah that's a fantastic question right yeah it is not as simple as you know uh copy paste right yeah and most of the Transformations we need to look through into three lens like people process technology like you are briefly mentioning uh people aspect of it is there is a lot of change management uh it is change management inside I.T or digital and also change management outside digital and uh you know in to some extent to the customers too right right uh so that is the first aspect you need to upskill the folks you have and also you need to have right level of Engagement with some of Si Partners to bring right level of talent to augment what is happening right and then the process side it is not like you just take a world wine and pour it into a new bottle kind of a thing right even though everybody likes old wine that's a difference in analogy uh so I think you need to look into the processes and make it optimized or you know there's a lot of transformation there either it is the processes goes into building the application and also the processes business processes inside the application like for example you need to look into what is that the devosa corps pipeline looks like what is the ACD looks like you know how is your deployment uh your deployment on-prem may be different than what you do in Cloud for example right uh like you were talking about the talent on doing those foundational items like how do you do logging how do you do monitoring and uh a system health checks and unit testing in performance testing you touch any of it it is different in a way when you build a cloud application compared to a legacy application right so you need to build those muscles you need to build that uh processes and then you need to scale them up when you need to expand from one team to I don't know maybe 10 teams now okay and then how do you scale it and still may be able to reuse them for a new application you're going to build by other teams or so on yeah yeah and what you just said here is a it's such a I mean for folks who are listening uh what Kishore you're explaining here is a true Testament of what it really takes to move Legacy applications to the cloud and to re-architect the application like just the amount of things you just shared in the two minutes each of the them is a cycle of effort that requires lots of greens a lot of efforts and it proved it kind of brings you know the effort that you have taken to put something out you know it's not like building a new application Legacy migration is a big challenge right so when you were trying to do this of course we have compute is easier right you know you need servers okay we have ec2 instance or you know uh we have instances available you need Network okay we have networking VPC available what was the biggest challenge in terms of migration was it the data was it the compute layer was it just breaking the entire application into small microservices and having agile what did you feel like that your team really did well and maybe you struggled with yeah I think can I say all of the above so yeah those are all challenges but some are smaller lines than other uh little bit longer lines right so I I think the biggest challenge we faced is uh taking this monolith and yeah and breaking into smaller consumable units right right because uh now you need to take a vertical cut not a horizontal cut if you take a horizontal cut then it it cannot be as much of transformational because you still have a lot of dependencies and everything like that um and then our challenges around when you do a vertical cut how do you synchronize the data so we went with micro Services event-based architecture we stood up for example kafkas and you know all those things so it replicates still if one process is down it is not bringing everything down with it so you can still continue to do you can have eventual consistency and then the data is even actually flowing where it needs to be and then users continue their processes um right like for example for this particular application we are talking about if application is down the trucks are rolling with seed so you're losing that Revenue right there in before you rise so you need to have a way of collecting that information and still able to synchronize eventually so right some of the challenges like that I think uh were you know we solved it but it took a little bit of a trial and error another one which is obvious for most of the listeners is change management either inside change management and also uh quote-unquote business change management right there are certain ways they are able to do it and want to do it and they were doing it for years and now uh the drop down is where that is supposed to be and the drop down or a checkbox doesn't do what used to do for tens of years if not right you know years together uh so doing that change management now at working with them they need to log into two different systems to complete their Journey you know those kind of things those were challenging I mean those are habits also right yeah but at the end of it it is very rewarding right you can look back and see man you know we accomplished something it is a lot better than what we started with I mean it's so great to know that you know I know um you know just when you were seeing uh or sharing I was just thinking this problem is so interesting to solve and not many people uh go in hard on to like go and say okay we this is an aggressive project it will take time because everybody wants to solve uh the scale problem tomorrow but there is so much that goes into breaking a monolithic application and breaking it into microservice that again needs to scale I mean you have topics that you're writing into but that topic has to be uh you know scaled in a way and replicated in a way that you will never lose that data in an event of an architecture so you have to consider all those things so you spoke like a true veteran who has handled and solved these problems multiple times so uh yeah I appreciate that you know uh David another thing we are actually working on uh moving one of the cloud application one of the monolitha Legacy application to Cloud we were part of that workshops and last weekend before kind of a thing right so it is uh I was uh talking to one of the person uh who was part of uh doing the initial implementation of that particular application right uh and again it brings uh millions of uh dollars of Revenue that particular application and uh and he said I thought it's fascinating right he said hey you know when we rolled out this application for the first time we introduced Mouse along with that application for the first time so so it works the application works uh so it it but it is not scalable it is brittle and uh the talent is not there to make updates or you know business changes it is difficult uh so those are some of the reasons why we want to do it but that being said it it is Decades of knowledge and Decades of customization you need to take and take a fresh look at it and uh it takes years to re redo this yeah yeah but you know that reminds me of an episode uh or not an episode but a conversation I had when I was working at a company uh when the company was finally changing offices okay uh this the company had been in an office location for almost 20 years and then they were moving to new office a new floor much more modern setup and then when they were moving they realized that under a desk there was an as400 kept which is like an old as400 it was still running nobody knew what it did okay and they were like so concerned about unplugging that because they did not know what was running on that and we knew that it hadn't been touched for a while because whoever sat there was saying well it's been around for at least so many years you know so so you're talking about uh technology that's working for such a long time and people are really wary of like hey do we really need to change it of course we need to but just remember that we're talking about I'm sure you have seen in some others desktops and servers there are stickers do not turn off yes yes so yeah and people who put those stickers they I think they moved on and there are certain applications people do not know what it does and how it does so you need to reverse engineer those things to redo that yeah some of those challenges that's that's how I feel sometimes when I'm talking to people who are who have been working on Mainframe you know like it's like such a uh it's a great technology but once you're on it it's just so difficult for you to kind of think in another way and the effort is like wild you know but so going back to whatever just sharing is now at amaran or amaran your control also entails these kind of efforts where you're trying to move Legacy applications into a more modern Paradigm and more or an architecture and then these challenges still still remain yeah that is right you know because we need to work very closely on various things right like we talked about do we have right Talent we cannot take on multiple uh Transformations at the same time we need to make sure there is the right level of uh organizational throughput and uh right level of subject matter experts and funding and uh you can't throw too much change in the organization and still able to do a good job at it right so there are multiple factors My Team Works closely with various business and digital leaders to understand how do we lay it out that is what we call it as iltp integrated long-term planning and we create road maps and make sure hey you know we have ability to do this then it needs to stagger these five things needs to go together this needs to happen before we get to uh you know touch on this one you know those kind of things and and that's how I think that kind of team is for folks listening is so important for you to have have that kind of a team in an organization because that then drives or connects business to the actual stories that are going to go into a release right and there is an alignment and if there is no communication that can affect and cause disaster that directly affects your Revenue right so well well said I wanted to ask about um you know how do you look at Cloud strategy right at amaran uh I know you're moving towards or our own cloud as well as but um you know what is important at amaran Now with uh the cloud strategies that you have do you get involved in conversations on what kind of solution to pick do you think about multi-cloud multi-region how do you think about that yeah so we have a cloud strategy and recently we worked on revamping the cloud strategy you know uh changes the constant thing in it right so right right uh like I was saying before there are four things we critically look at one is agility second one is uh you know how resilient and availability and cost Effectiveness so those are some of the pillars in uh Cloud strategy we are making good progress we have a strong team which implements uh devser cops we have a strong team which works closely with agile and make sure you know we can do these things in a way which is scalable as well as sustainable and right now we're actively looking into what we call it as Cloud hosting rationalization initiative so as the name suggests we have quite a bit of applications we we have hundreds of applications in our data center and what we are doing is we are cataloging all of those applications and trying to understand what segment they fit in right are they ready to transform into Cloud application or they need to stay back or for various reasons including some regulatory concerns how do we need to retire and this is the time view right to transform invest eliminate and those kind of a view we look into and we try to put them into different segments hey this is then we need to prioritize this this is a critical application so we maybe we need to prioritize based on the availability of resources as well as funding and so on and this is not as important application probably we can just lift it and shift it right you know uh with a view of hey we are going to eliminate it in five years right so we're taking different uh aspects like we call it a decision tree and then we are putting them into various segments and then we convert them into road maps now we take those applications and put it put a business case and put a roadmap around it then what is our two-year window looks like Fourier window Fourier window looks like so we have a view into how much of investment we need and you know what is that effort going to look like and so on got it yeah and that that also helps you with the planning of what absolutely you know Talent you need what kind of Technology you're going to go dive into uh so do you do you have like Innovation uh teams that are focused on Research early on to before you fully fled out and plan what that looks like do you do that as well so basically the way we uh we're trying to do this one is we have something called uh pre-work or pre-implementation we say it so we say Hey you know for example this particular application we need to transform so we we put a roadmap before we call it pre-implementation which includes understanding about different uh opportunities or different uh ways of doing it doing like a pocs or Pilots working with various SI providers putting sows vendors discussions those are all we call it as a pre-work depends on size of the effort it may run for months or maybe a year all right there is limited value in jumping on it without knowing all the details like for example I was talking about that application we do not know how the current system works so right we probably need to do some documentation understanding current processes right and then we need to do some user Journey is we need to do some mappings around what does the future may look like right so those are all the pre-works you need to do probably you don't have to do at the individual story level but at least you need to do it at the Epic and if you understand and then it gives you a good view of hey this is my roadmap looks like then how am I going to chunk it into various continuous delivery and that is what we call it pre-implementation um and again do all applications do they need it I don't know right depends on the size and criticality of it we need to carve out some dedicated deliberate effort before we jump on it got it yeah and I know you kind of mentioned this but I wanted to dive into this a little bit more for folks who are listening to kind of learn a little bit so what drives the kind of Technology you choose for the transformation right I know you mentioned scale but is it driven by hey the company is growing we need to add more utility uh yes users there's more data required then now the experience that is required by these users is different they want to see real time hey how much electricity am I using where can I save money so what is driving the requirement for the technology that you're choosing yeah yeah that is fantastic right like I was saying before we don't want to uh implement or bring in technology just for the sake of it so it needs to tie with some business justification or written on investment or some aspects of it right so we definitely need to do that uh going back to your question on what kind of Technology we bring like depends on the type of problem we are trying to solve um so like for example a there is something called outage right when a big storm goes through your territory or you know something uh something of that so it happens you need to make sure the systems can sustain large volume of data coming in now we have so many Transformers and meters pumping data or so many people are calling and you know there are outage notifications you need to send so yeah you need to have ability for that particular application or integration because I also lead integration so now all that integration is a lot more chatty right because you need to so much of data coming in yeah yeah yeah so you need to have ability to quickly scale up and we that is one critical component and second component is a when for whatever reason they failed they need to come back very quickly that is the resiliency aspect of it right right right and now you have the uh Power is down and also your application is down it's not a good scenario so yeah you need to make sure uh you know the outage notification is also down is not a good experience for our customers uh so you need to have fault tolerant very resilient uh so we are working towards that active active active passive kind of a resiliency strategies and having a good monitoring alerting we're constantly improving continuous Improvement on that one uh how do we make things better right we are learning from certain past mistakes and make it better we don't repeat the same thing again yeah I think and also learning from mistakes is just so critical to it's just a part of building a better solution is what I feel is an active active that's the world I live in you know I want everything active active all the time I don't want to think about Disaster Recovery like when we you know cockroachdb right is one of those Solutions where we I always felt like why do I need a dependency in my architecture or in my design it should be a Peerless infrastructure it should be I can write and read all the time I don't need to just kill reads I want to scale everything right and and that's where I feel like these architectures and these kind of design patterns kind of make sense and when you were saying that active active uh you know are you considering or do you are building your applications with uh you know kubernetes infrastructure where self-healing Auto healing capabilities have to build into the infrastructure and application transformations yeah they are definitely and also we have uh hosted in AWS and other clouds and we also have some SAS products like you know we use Salesforce we use Oracle we use other Maximo and other some of them are on-prem most of them are hosted in Cloud right right and they in turn they bring some of those resiliency out of the box right and uh all the custom Solutions we have like you just said we they're running a container with the right level of resiliency built into it like for example we use mulesoft for integration okay yeah they have Cloud solution which has resiliency built into it and we also use quite a few apis on-prem so we are building active active for that so we have two data centers and for whatever reason API is down or one particular segment is done it automatically routes to the other active active region so users do not even know something of this kind happens right right so it is not just uh to that level right now we need to look into underlying database right it needs to be active active in some shape or form and if you're using message queues if you're using file systems you know everything needs to have that kind of a posture so the the user process is uh you know able to sustain these kind of of fault tolerance and having that high level of resiliency end to end right right yeah of all the Technologies you said honestly Oracle is the only one that I don't like yeah and part of the reason is part of the reason is and nothing because of working at Cocker's TV nothing to do with that but it's just that my personal experience with like new modern paradigms and new architectures that I'm building I have to go in and when I hate Golden Gate and designing replication that way uh and I used to work at another company that had a PLS architecture so uh I don't want to think about failover at all or blue green deployments right and uh so uh that's why I I started using uh I use actively uh you know Technologies not just on a database level or I try to build it in a way that we have a playlist architecture and it's done in a way that I don't even think about Disaster Recovery or rather only thing I did think about is making sure I replicate the data to maybe another region and keep it there just in case my region is not available like that's the max but I don't want to get involved in uh like Golden Gate and things like that so except for Oracle I I love everything else that you said that that makes sense right you know if you keep that problem outside and it is solved already so you don't have to solve it again right you know why we invent the wheel if it is given to you already right exactly also also Oracle is just so expensive it's just yeah yeah I hear you that's how you know maybe we need to have a different conversation after this call but uh but I think what you said uh I mean there is so many amazing insights that you shared about just the design pattern itself that you have to consider when you're thinking about scale right and it's great to know that uh you know uh even utility companies and not I mean we talk about amazing stuff that Uber is doing amazing stuff that Twitter is right or say of you know doordash all these modern new companies but we have to realize that these patterns and these Transformations are happening within companies who are Fortune 500 like amaran uh that have been historically known for building you know monolithic applications and things like that so I'm really glad that somebody like you were sharing that perspective with you know the listeners here so that's great to know yeah yeah so uh the user expectations are gradually going up right you know like you just pointed out either it is Amazon or the worlds or you know now it is X not a Twitter I guess so yeah excellent and uh are you oversaw the world so when users like you and me when we are using those services when then we go to a power company and use their website your expectations are the same so whether so we need to meet those and we if not exceed so we are doing all these Transformations we are doing doubling down on these Technologies where where they're applicable and uh you know amaran is in top quartile when it comes to customer satisfaction so we do that every year and we got awards from tidia uh you know the customer satisfaction we're in uh and that is very uh you know what they say is re uh enforcing and that we are going in the right path and we are you know providing it to our customers what they are looking for right right and you were talking about this whole idea of like you know looking at resiliency of course is important to you right but at the same time how do you look at the whole build versus buy right you know I mean there is a lot of things that you can I mean and you're in a position where you you have to at least highlight talk or make decisions around hey this is something we need to buy uh you know why do we build and put engineering effort into something versus when something already exists right so how do you get into give me an example of that as well as how do you think about those problems yeah sure see the uh our we have our decision tree or Mantra right you know the way we try to do it is the emphasis is to go to cloud and if there is an exception that's fine but uh the default quote unquote is you know why not Cloud right yeah that is the first one and also why why we need to custom build when we can buy it and utilities there are quite a few Utilities in States and maybe hundreds of utilities if you add all the countries in the world so it in many ways the problem is already solved right so we go that route of hey can we just bring a solution which is like a SAS solution that is the route we try to take right right in some Niche areas where we probably need to do custom build right right um and uh going back to your question on hey you know a way example of it right for a we recently implemented uh Erp and I think that is a good example where I don't know how much value it is for you to custom build a Erp and the same uh we have a lot of Maximo is one of the examples right where we have lot of utilities use it and I think some non-utilities use it as well so so limited value building that kind of a functionality and custom build so wherever it makes sense bring a best possible uh industry leading solution and try to make it as a cloud hosted sales solution if possible and so that's where we get into a lot of challenges around integration we need to now have a robust Plumbing to connect all the data coming in from various clouds and SAS Solutions and also on-prem Solutions and how do you make all of this work and do not make it as a single point of failure imagine your plumbing gets choked now you have everything else working but no water flowing right so so that is something very critical for us and we we double down on making sure we have right level of emphasis on integration so um I wanted to ask you when you were saying that is that you have all of these things that you have to do transformation how do you uh like wet out which technology makes sense uh for this like there has to be a waiting process right as to do you do you have you built that into your engineering practice or your architectural process principle process that you have built if we we have quite a few patterns established and we have standards established uh we have a process we call it technology work assessment so TWA so we any new any new application or any new product when it comes through it has to go through that process we add bunch of requirements they're not always just architectural requirements we give architectural requirements data requirements resiliency requirements as well as cyber requirements we provide those and then we weigh them based on that it has to fit those particular requirements and also we add in patterns hey you need to have single sign-on for example right right uh so it it is weight based on those criterias and then we try our best effort to meet all the requirements in some cases you need to throw in an exception right for certain reasons like you know there is a regulatory reason or a Time boundary then you need to get something done you come back and register it later so there is a there is a waiting process before we bring some s4ws or something like that or a procure tool or a technology we go through that waiting process and it is effective so far yeah and the whole idea is that the process should enable Innovation but rather than slow it down it's something that's right you're also thinking about okay yeah I I agree and uh that is why going back to our previous conversation a little bit you know we need to provide good guard rails that way people know hey you know this I understand I just don't jump off and add no authentication and create a huge cyber risk or uh I expose all my data but if you provide that kind of a guardrail I think you know it gives a lot of Liberty for folks to innovate um and that being said you need to constantly check and adjust if your guard rails are in a well-placed location or not or you need to move them around and make sure it aligns with what's happening in the ground right oh that's brilliant yeah I wanted to ask where does Kishore go for uh keeping himself updated with all the technology wildness that's going on I mean here's here's why I'm asking that question 15 years ago you had like 100 softwares that you need to pick and you need a where award okay and you would hear about it in conversations and now we have 10 000 Solutions almost software is like every cloud company has uh the same or similar Solutions on you know their platform then there are third-party SAS products uh you know who are also solving this and then there are all those other people who are innovating in the open source space so how do you sure like keep up with all of that and you're in a position where you have to be like well somebody says something like at least I should know about it you know what do you do yeah to be honest with you I I I there are a lot of key buzzwords I still need to be a little bit reactive right I need to go and find out what they mean because you know that little bit contextual also right you know um so I I agree it's a constant learning um yeah I think you know one of the areas I go into is uh Gartner uh so we go there for research and try to understand what is the quadrant they are in and you know some other we try to do some little bit of research there before we feel comfortable and it needs to be we set up some analyst calls and try to understand what is the industry Trends and uh there are some in industry consortiums for example epri is an industry Consortium we have some forums there we go there and understand hey if any similar utility has done something like this or what are the industry Trends happening you know um and some essays uh like you know accentures of the world Delights of the world they also bring their perspective of what's happening in the industry all other implementations they're doing um and uh last but not least I started using uh chat GPT and Bard Google bard and uh yeah you know they give a good perspective too you know obviously you need to take it with a grain of salt but yeah and that is a that is an interesting new space to go as well yeah no I I agree I think uh one of the biggest things that when you just said it suddenly hit my mind is that uh I was talking to somebody else about it as well as like there's so many papers that are getting released research papers and sometimes you feel like I don't have time to read 30 paper research paper to get get that knowledge now you can pull that into uh chat Jeopardy or uh or wired and it condenses it down and you have enough information there for you to have an executive conversation with somebody or uh have a conversation with your architect around hey why don't we look at this as an idea of implementing something right uh so I can feel like it's very helpful you know I am doing that myself you know uh and I I I feel I'm glad to know that people in your position are also feeling that's something valuable for you to try so that's good yeah yeah I think so you know that's why it got uh 100 million users in two days or something that's what they said right and yeah yeah yeah yeah I think it definitely has value on how to use it but you need to double check and validate I think is where it is right now yeah and and one of the ways I'd really do it is I will ask the same prompt on tragibility I will ask it on 3.54 as well as on Bard and the

difference really is the output you know the output from intelligence or the way it makes me feel gpd4 does a great job but uh it doesn't have internet data unless because they had the browser plugin available and they took it off because of an issue but bar gives me more realistic timeline and so I agree with you like validation is the biggest uh Factor when it comes to AI llms yeah I mean they're still maturing right we need to give them a little bit of a slack they're working on it yeah I'm assuming you know because I'm paying 20 just for that uh internet again that is for the chat GPT 4 isn't it so lgbt4 uh yeah and I use a few plugins as well like I use the PDF plugin um I also personally use the instacart plugin which is which I feel is a great use case where uh we make our meal plans on instacart so what we do is we ask to produce a meal plan for us and then I bring the instacart plug-in it creates all the ingredients I need for the week and just we just order it directly from Tom Thumb or uh an already yeah so that's awesome yeah that's a pretty practical use case so yeah I'll tell you one more practical use case which I recently did uh we went to Iceland for vacation yeah and then I went to one of this region AIS and I said hey create itinerary for me and they did it for multiple days and I I tweaked it little bit here and there and it worked you know I just used that it array exactly I mean I mean the way I see these is like it's enabling us to save time and to refocus that time towards something that might need more attention you know and I thought and somebody like me and I feel like you who wants to save time and make the time more valuable I think definitely is a place to uh you know invest in you know they're trying to agree I think we all become like a a region AI uh experts very soon yeah yeah I think so I think we will be uh well uh I I know we're getting towards the end of our you know uh podcast and I mean first of all some of the things you've shared is perspectively so important for folks to consider when they are moving from Legacy and re-architecting everything for a modern uh you know architecture or an you know agile thought process or microservice and things like that uh for companies who are uh you know are folks who are trying to build infrastructure like this are looking to move in this direction what is your advice to them uh and what what what would be your I know you've already mentioned a few things but what do you want to say that this is what you have to stick to to kind of make sure that you don't make those mistakes that probably you have made in the process so oh yeah you know everybody has cars in the back right yeah for sure I think uh in it is not just a technology problem I think that is what I I probably emphasize again it is a people process technology you need to look through all three lenses I think uh making an assumption that technology only can solve everything is probably a pitfall right you need to make sure like we talked about you have do you have right level of skills uh from the people perspective and how do you uh uh upskill them or you need to bring go to somewhere and get an augmented right skill set you know do we have right level of subject matter experts in technology and also in the business process and do we have a product mindset if that's what you're looking for and uh are you able to take that big piece of uh delivery and can you chunk it into smaller pieces and show success show delivery continuous progress and value are you able to realize that value as soon as possible so I think you need to look through all three lenses I think I think technology in many ways is easier problem to solve in a corporation than uh people and process perspective that's what I would recommend and encourage people to take a look when you're embarking into this and take one step at a time you don't have to do the whole thing like a big bang approach it is not usually the best way to do it that was great advice for folks listening in uh you know kishore's uh you know is active on LinkedIn I think that's one of the places where people can follow you is there any other place that you recommend people can connect with you or like learn from you in terms of you know material that you post or repost um the link is the best way and happy to collaborate with uh folks and uh learn from them as well as I can provide my thoughts as well and happy to do that thank you so much for so being so candid and for all the amazing uh you know ideas that you shared on the Big Ideas in app architecture podcast it has been a pleasure to have you here and again I appreciate it thank you David you know I can't thank you enough for uh hosting me and thank you for this opportunity thank you and good luck with them [Music]

2023-09-13 10:41

Other news