 We have Arjuna Joshi with us. She will be talking about the environmental impact of digital disruption. So let's welcome Arjuna. Arjuna, the floor is all yours. Thanks Priyanka. It's exciting to be part of the Agile India. I've been attending as well as presenting at this conference. I think from 2014 onwards, so almost seven years now. And this year it's been a virtual format and the Confingene setup has come out really well. So kudos to the organizers to make sure that we continue to benefit the things of Agile even in this way. So with that, let me give a little bit introduction of who I am, where I come from, what do I do, and actually what got interested me in this topic, right? So I lead the transformation team at LTI, which is Larson and True Growing for Tech. And one of the things that I follow passionately in my personal life is also about things in terms of what is happening overall, what kind of legacy we are leaving for our children, and especially that became even more important when my daughter was born and I started thinking about what is it that I'm doing for this planet, right? And that kind of got me hooked into a little bit of trying to do my bit in terms of making sure saving the plastic piece and making sure we are at least reusing, recycling, making sure we are planting trees. So a lot of those typical things that you might hear an individual would do is something that got started, right? And then about two years back is where I came across this article as I was researching more and more on this topic about saying that, hey, while fossil fuels and clothing seem to be the top two contributors to this entire carbon thing that's going around and the whole climate change that's happening very soon and very quickly the digital emissions are catching up, right? And that is something I was like, okay, that's interesting to know because every time for me at least the way it was that, hey, I don't take a printout using a paper, I'm saving paper, but then I realized, hey, while I'm saving paper there, I'm also spending so much time on the screen that it kind of gets offset, right? And that kind of got me interested, hooked into this topic and started the research and then I realized that, hey, as a technologist who is working with a technology company, why don't we do something about it? And then from my company standpoint, we also set up a research and we started looking into this, advising, talking with our clients and taking this forward. So that's a little bit of a journey of why I got interested in this topic and how it came in. And as I spoke to more and more people, I realized that not many are even aware that there is an impact of what they do and how they do. And that's where I thought that, hey, Agile India, one of the conferences which is out there, which talks about the leading edge things, why not put this topic out there and make sure people understand and at least make them aware about the things that are happening, right? And as they say, if you are not even aware, there's nothing you can do about it. The more you are aware, at least you may take a look at what the impact you are leaving and take some proactive steps to stop it. So in the topic that I have today, I'm going to cover a couple of things. The first one is what is happening across us and what are the statistics showing about impact of digitalization. So that's the first thing I'll cover. And during the second part of the presentation, I'll also cover that not all is lost. There is still hope out there and we as technologists have to make sure that we put all our best out there to make it work, right? And I'll talk also about some of the techniques that we can use when we develop software or when we engineer software and how that helps to make our own bit, little bit in this whole climate change that's happening across. So with that, I'll get started. If I look at what has happened in the past two years, and I'm sure all of us are sick of hearing the word pandemic and the impact of it. But I also read somewhere that a pandemic has been the transformation leader for accelerating towards digitalization, right? And rightly so. Because if you look at at least in India where I'm based in, going to a doctor and taking that appointment over, let's say, a phone and a doctor trying to tell you what is, what ails you over a computer and a zoom meeting was unheard of, right? And I myself did that a couple of times during this past few years. And this is caught up, right? And it's going to increase. It's going to increase even more. So we are seeing these kind of things where it's telemedicine and talks about it. In terms of e-payments, I mean, I was surprised the other day. I had gone out and of course, typically you will go, you will buy some vegetables and all and it may not be in a, I would say in a shop, right? It could be on from you buy from street and hawkers were selling things on the street, right? And it was just a very small amount, 50 rupees, which is less than even a dollar, right? And like I said, madam, use your GP and I'll take it, I accept it, right? I was like so happy that they were not asking me and insisting to pay in cash because usually that's what's happened. So at the lowest level also, you are seeing this lot of traction that's happening on e-payments and that's also going to continue, right? So all these are the effects of digitalization that you see. Similarly, education still, those of you who've died in this part of the world, the schools have not started. They are still online and like the way we are doing this conference, it continues to remain, right? Especially for the kids of younger ages where schools have not started the education, which is still continuing online, right? And many of these things would continue and it's not just the school education, even classes like your karate classes and singing classes and dance classes. All of that has moved online, right? But this is what we are seeing now and this is already happening. This is all around us. But there's also a change that's happening for the future, right? For example, if you look at what happened in 2020, right? And this is just one statistics, which is a very interesting one which I thought I should share with the attendees here. Some of you might be aware of it. This was a very popular graphic which was doing its round early this year, right? In terms of how much within one minute, what is the kind of data that gets sent across the internet, right? Now, if you look at it, if this is the kind of thing that's happening, imagine the amount of energy that's getting consumed. Imagine the amount of screen time devices that are getting manufactured to support this ask, which is coming from us who are part of this economy. Imagine the impact that is further going to have fueled onto this entire climate change things, right? And this is just growing exponentially. There's nobody that anything can do about it. So the question we need to always keep on asking is what next and what actually can happen if this doesn't get curved or if we don't take that steps right now, right? And the way this is going to increase is not just in terms of the consumer and the social things, right? Because this is again, as I said, obvious, e-payments, telemedicine, education, messaging, social media. This is all obvious. This is all around us. But there are also some researches happening in the areas which have been unheard of where digitalization is making an impact, right? Things like digital juries. So today, instead of having a 12 member jury team presiding over certain court proceedings, there are a lot of companies who are experimenting with digital juries along with the governments of the respective countries to say, hey, can we use digitalization there to take decisions related to law and order? Now, this is very interesting because imagine for those who are familiar with the game of cricket, you have something called as third empire. So while there are two empires who are there on the field who are taking certain decisions, you go and refer to the third empire saying that, hey, okay, what was exactly the thing going on? So think of that third empire like a digital jury. Similarly, I'm sure these things happen in sports all the time. You have in football where you have a referee who is like that digital jury, right? So things like law and order, which have always been so human focused is also moving into digital jury with a lot of AI kind of solutions built into it to say, hey, look at the past cases, what are happening? Based on that, can you even say whether whatever jury gave the word it, is it even right or wrong, right? And this is getting a lot of traction. Another thing which I recently came across is 3D printed artificial organs. So we all know that organs in medical thing are always in short supply and organ donation is something which is important for all of us as well. But now companies are investing to figure out, can you do 3D printing of artificial organs where you can give a digital image of what the patient has, give certain their DNA sequencing, give certain type of blood type and all they have and based on that print an organ which is very, very suitable to them. And I'm not just talking about artificial limbs like a hand or a foot which are visible from outside but organs like your heart. So just a year ago, one of the universities successfully was able to print a heart which can be used in a heart transplant, right? Or digital clothes. So we do know that clothes is supposed to be one of the most pollutant industry out there. But there is also moving towards digital clothes which means that, hey, why do I have to buy clothes? At the click of a button I can keep on changing the clothes where if it's in a virtual kind of format, especially in the Instagram era that we have, I just click it and that image of the dress or the shirt that I'm wearing gets superimposed on me, right? And these get sold for thousands of dollars, so it's not a very small amount. You would imagine, hey, for an image of a dress or a shirt or a 3D image of I am paying so much. But you would be surprised to see that all the leading digital designers and luxury houses are kind of moving into this bandwagon and jumping into it, right? So this is the kind of thing that's happening all around us in terms of digitalization, right? And to meet this unsatiating need of all us consumers, enterprises are going and supporting and investing very highly into this, right? And I don't, it's a chicken and egg situation. There's a demand from the consumers, enterprise go and invest, enterprise invest and then that gets consumed by the customers, right? So it's a typical chicken and egg situation where this is happening every single time, right? So if you look at the global spending which is predicted, it's going to be $6.8 trillion, which is equivalent to the GDP of entire African continent and the entire South American continent put today. So in next two years, we are already ending 2021. So probably next by the end of 2022, this is the kind of amount that would be there around digitalization and the spend that's happening from a consumer standpoint. So at one stage, we are talking about reducing plastic, planting more trees, making sure climate change is something which gets controlled as much as possible. On the other hand, we as consumers and we as technologists are also contributing to that piece, right? And hence the question we need to ask ourselves is we are doing all of these things. We are investing in digital, we are excited about AI, we are, I mean, all companies today, all the tools out there in the market today have a component of AI. There's not a single tool which doesn't have it or there's not investing in it, right? So the question to be asked is, are we doing it responsibly, right? And I compare this like a thing that it's not like we should stop doing it because it's like saying that, hey, light consumes energy, so let me start using candles and not use tube lights and bulbs, right? That's not what we are saying. The question we are saying is while we are using tube lights and bulbs, can I do it in an efficient manner so that I don't have to resort to the candles by making sure the planet suffers so much, right? So that's the question we need to ask ourselves. So that brings me to a point that usually when we look at all of this, it feels very hardware centric, right? Which is the screen, battery, the devices, what does it mean from a software part of it? What does it that software can do in all of this, right? So here are just some examples to tell you what happens when software starts doing, right? So this is a research that was done by a university in Portugal and a very interesting statistics they found, right? And that statistics was that Python program consumes nearly 76 times more energy than a program written in C. Now C and Python, all software, it's got nothing to do with whether it's a hardware or whether it's a device or whether it's a screen or it's nothing to do with that. It's pure, pure software, right? So if I'm writing something in C versus a Python, there's a choice I can make. So and I can decide where I need to go towards, right? Similarly, if we are saying I want to be doing AI, single-AI model, the smallest one, right? It emits the CO2, just training one-AI model equivalent of five times the lifetime emissions of a car, which is huge. It's absolutely huge. And some of you who may be closely working with AI and those kind of techniques, you might have heard that when GPT-3 came out, the biggest argument against GPT-3 was that it consumes a lot of energy to train. And people who are using underlying GPT-3 in their own software inherit that energy because they end up using that algorithms and hence consuming a lot more energy and contributing to the digital emissions. So this was a big argument that happened couple of months back. But that argument happened. There was some noise that was created in the media. There were some people who said things about it. But today, if you ask any AI-ML person, what happens and their answer is, hey, I know GPT-3, I know how to use it, right? So in spite of all of that, what their GPT-3 today remains one of the most important software package in the AI-ML side of things, right? Similarly, if you look at data centers. Yes, we have a trend to move towards cloud, which is a great trend. But then the question also is that you move from data center to a cloud, but cloud is also consuming energy, right? So it's like just shifting your responsibility from one problem area to another problem area. But at the end of it, it's not really shifted, right? So how you configure your clouds, how you configure your data centers, that is also very important. Bitcoin mining. Yes, this is again a very, very widely debated topic. And I remember again, a couple of months back, Elon Musk said that they will not accept cryptocurrency for Tesla cars. When people go and buy Tesla cars using cryptocurrency, he had made a statement saying he will not do that because cryptocurrency underlying uses Bitcoin mining and underlying uses blockchain, which is a lot of consuming a lot of energy, right? And of course, now it's again started, he redacted some of his statements and that happened. But the point remains that if we are now going to move from a normal currency with all the banks, all the financial institutions, governments, all of them investing heavily into cryptocurrency, which is for which blockchain is the key, key building block. What does it mean? And now all of these examples that I'm talking are no longer hardware examples. All of these examples are software examples, which are in our hands as software engineers and software developers to control and make something about it so that we do it in a good manner, right? So if you look at the overall emissions that our industry, the technology industry, and when I say technology, especially the software industry is contributing to, there are four key types of emissions, right? The first of course is the hardware emissions due to servers, storage, data, devices, all of that, right? That's pretty clear, screens, everything. That's one. The second emission is the software package emission, right? Which is about how software is getting developed and used. And as digitalization increases, the emissions around software packages, what is going to increase? Hardware not necessarily so because since past almost 15, 20 years, the hardware manufacturers are focusing very, very heavily into reducing and reducing the energy consumption and making their hardware efficiency. But for software standpoint, we've just gotten started or I would say in a way that we haven't even got started in certain cases because people are not aware. If you just stop a developer and just ask him, hey, what is your software package emission of the software you've developed? I mean, that person would not have a clue on what, what am I talking about? They wouldn't be like, okay, what is this? Right? So we haven't even got started. And the pace at which digitalization is increasing this piece across all of these four is increasing exponential. Third, of course, is human capital, which means if we are traveling, and then of course the physical assets, which all of us are familiar with. And typically when somebody says green, we go and try to optimize the physical asset emissions by switching off the light in a room which is not being used, right? So that's our typical individual behaviors and we feel good about it without realizing the ones that we are contributing directly is the software package thing. So what can we do about it, right? That's the big question. What is it that can happen from software standpoint, right? So while developing your software itself plus how you maintain your software is crucial. So that means that every single software engineering practice that we do, we need to relook at it. Now there's also an argument that's out there in the industry that says, hey, let's do carbon offsetting. So if you look closely and I kind of started reading a lot about the ESG reports that these large enterprises and companies go forward with. And there's always a statement about net zero carbon emissions, right? Now, the net zero carbon emissions, the underlying assumption for the net zero carbon emission is that some of this would be achieved through carbon offsetting. What that means is I may continue to use the screens which are there, which are highly energy consuming, but I may go and plant some trees, or I may buy some credits from companies who are reducing their emissions, right? That is called carbon offsetting, which means I don't necessarily curb the things that I'm doing, but I spend money to make sure somebody else does it well so that they can plant trees where it's needed or they can reduce their own things because I may not be able to do it in my area, right? And that way then I say that, hey, though I'm doing these things, somebody else I'm paying them to not do it and hence I'm making sure I get some brownie points for what those guys do. And that concept is called as carbon offsetting. Now, this is a technique which is used very widely by all the companies when they do their declarations on their net zero emissions. But the reality is that carbon offsetting is not enough, because if you look at the pace at which the things are going, if you really want to offset everything, all the emissions which are there, there are some studies which says the entire earth has to be covered with its green footprint, which means we as humans cannot, should not even stay. Everything if you can plant on the earth, make it green, then only you might be able to do it, which is not practical, which is not humanly feasible on an average of forest takes more than 100 years to grow. And just doing tree plantations by planting one set of trees is even more harmful to the environment because soil erosion happens when it's just one type of trees which get planted. If you have a mix, then there's also an ecosystem that you are creating. So carbon offsetting is okay, not enough, we need to do much more, which means we have to reduce it right at the level where we start creating software itself. And that's where some of the things that I'll now start talking about. For example, change in the architecture itself, right? Now I know that this is something which people have talked about off late, saying that hey, go to serverless computing, do containerization and we keep on hearing this in a modern technology or an application modernization context. But we don't realize that this has bigger impact on the way things get done, right? For example, the whole element of serverless computing is to make sure that you are using the resources only when it's truly needed, right? Or when you are, let's say, creating databases and all of it nowadays is all databases, data insights, centralized, large data lakes and all of that. Number of times you would be moving your data element that also affects the time and the energy that you consume, right? So right at the word go where you are architecting your system at that point itself instead of just taking considerations of performance of the system or just taking considerations of technology for the system, tech stack, right? The architects need to start looking at how do I architect it for a greener IT out there, right? So that becomes a crucial factor and in fact I would go so far as to say and if that suggestion goes through that when you are talking about Togaf and enterprise architecture and those kind of things, when you are talking about these architecture patterns, can you bring about the architecture patterns which help you to figure out which patterns are more suitable to reduce this footprint? I know there is a lot of fuzzy world out there, a lot of studies need to be done but if we as a community don't come together and start doing that study and sharing those results, it will be a lot more time until we reach here, right? So if there are some architectures on this call today for the stock, I would urge them to look at at least one parameter when they are designing architecture, not just the tech stack, not just the new architecture patterns but also bring this as a third parameter while you are doing it, right? The other piece is cloud. So yes, you are moving to clouds. We have those public AWS, GCP and Microsoft Azure as the top cloud providers. Out of these, GCP is one of the ones which went first on green energy. Microsoft has gone to a certain extent but still getting there and so is AWS which is now it's a little much farther away amongst all the three. But while they may be moving to green energy, the question is when we are moving our workloads to cloud. Many times we do use public cloud but also there's this whole thing of hybrid cloud that's coming up, multi cloud that's coming up, right? So you need multi cloud tendency that you are building in. You are bringing in the hybrid cloud tendency where you have a combination of private and public that's coming in, right? Now when that is happening, how are you architecting your cloud nodes to make sure that this is done in an efficient manner, right? For example, many a times when I speak with the cloud architects and cloud engineers, the way they do it is let's say they give a direction to the cloud provider if it's a public one. That they are consuming that look at one of the things in terms of the nodes that are used or from where the workloads are decided and then there's a BCP strategy which looks at let's say the loaders in US then look at the BCP somewhere else in South America and so on and so forth, right? But every hops that you do on a node is going to consume a lot more energy for your software. Of course it brings in a latency factor but also the energy factor. So when you're moving to cloud and architecting for the cloud, how you are doing it and is this one of the factor that you are advising to your clients? Is this one of the factors you are advising to your teams becomes very important? The third of course is coding language preferences. So few years back there was this whole dark mode, bright mode, white mode kind of conversations that was happening on screen, right? Which means that all apps have to be optimized to this. One of the things what was the reason behind it was that white mode is a little stressful to the eye, constantly bright light is coming in. Whereas if you use darker background, it's not so much so. But there's also one more aspect to it is that from an energy consumption standpoint, how does it happen, right? So if a person who likes a bright mode gets that pop-up saying that, hey, it's not just for your own eyes but it's also there's a impact from an energy standpoint. They might look at switching and remaining on to the dark mode, right? Or when you are looking at the coding languages that you are using, the one example that I showed you before, that C versus a Python. I'm not saying that move everything to Python but at least when you are choosing those languages, try and see which ones you are doing, right? Or when you are doing HTTPS, HTML, these kind of coding that you do, especially in the web apps side of things, there's a lot more efficiency that you can bring in. Now, the other piece is of course on the AI standpoint, I did speak about GPT-3 and the impact that AI has but AI is here to stay, it's there everywhere, everyone is asking for it. So the tagline of green which usually is reduce, recycle, reuse applies completely to the AI, right? One is be smart about where you are applying AI. You don't need to do it everywhere the way things are happening today. Like if it's not AI, it's not cool enough, right? That should not be the case. So first is reduce where you have to use it and use it only when necessary. And then reuse and recycle, which means that every time when you have spent energy and time training an AI model, can others within your enterprise at least use and consume it? So making sure that reusability gets a lot more focused on by enterprises becomes very, very crucial. I'm just looking at my time as well. Thank you Priyanka for that. So that also becomes a very crucial point on how when you are designing software systems, you, if you are in a position where you are leading some of these decisions within the company, you should take a look at these. And also if you are a developer who's working on this, you should stand up in that sprint planning meeting or your PI planning meeting and say, hey, why are we doing this? Have we considered impact of looking at some of these things, right? But the biggest question which happens is that all this is all awareness, but where are the data points and how can you use those data points to show that something is truly missing, right? So the way you calculate your technical debts for the software that you have developed using a SonarCube or a CAST, are there any tools which tell a developer that, hey, this is your green debt or a green IT debt, right? So, so far the answer was no, but from past one and a half year, I'm seeing a lot of good things where some of the existing two providers are investing into it. There are also small niche players who are entering the market into it, right? So these are just some examples. The thing that you see on the left hand side is just an example of the research that was done by the Portical University where they compare different languages on how they consume energy. As you can see, C's right there at the top compared to a J-Ruby and a Lang and then a Python, which is here, right? Which are more and more heavy, whole pythons, all of these today's languages are out here. But there are companies like these which are investing a little bit into it. And there's an interesting thing that I quickly wanted to show you what I did was just for this conference. I went ahead and looked at what is the impact of the webpage of this conference, right? So there are some of these tools out there which can give you this right away. So this is what it says, right? So it says, compared to other pages, where do you stand? So it says it's a high consumption page, most of it on the CPU networkless, so on screen. And then they have some statistics around compared to the Agile India webpage, which are the ones which are less consuming, right? So this consumes almost 91 milligrams of CO2. What does it do, right? And then some suggestions on what can happen. So such kind of tools exist. You can always use it. Also these, sorry, let me just, okay, sorry about that, yeah, okay. So also when these tools, what they do is, for example, you can hook it into your CI CD pipeline. So the way it gives you tech that you can get a green data as well, right? So there is a lot of research that's happening. It will continue to happen. And while all of these at a developer level we can do, at an enterprise level, we need to be doing a lot more. So companies need to, of course, make sure awareness happens around it. Then there should be your NFRs, non-functional requirements today, which are around performance, security, and then your memory and those kind of things should start including NFRs around green IT as well. You cannot not have those. So this becomes very important when you're developing a software system. You, of course, need to have roles who can tell and help out the teams. It's one thing to set policies, but if those policies are not enabled by right roles, those remain on paper, right? And then, of course, continuous tracking is necessary by these tools. Now, again, if you look at what is happening in the market, there are some tools. For example, Microsoft has some tools which looks at the Azure consumption that you do, which you can extrapolate to the CO2 impact. Or then the things like the cast or the green frame that I showed you, which may attack certain part of it, which means some tools are attacking only the web page. Some tools are looking at the server side consumption. Some look at the cloud consumption. But there is nothing one solution which is looking at it from an end to end. Because if you look from a software architecture standpoint, it's not just the front end which is there. You have a front end, you have a middle, you, of course, have your databases, then you have your cloud and hosting and infra and all of that, right? So that ecosystem where there's a consolidated index that comes out today doesn't exist. While research is on, there's a lot that we need to do. And that's one of my focus areas as well, where I'm looking at all of these different things and how we can bring some of these things together is something that I'm actively involved in to figure out, right? So if you look at it overall, not many things can be done. Not many things have to be done. And this is a movement that we need to start to make sure that we are being digital and we are also being responsible. So with that, that was a quick view of what I wanted to share with you all today. Hope it was useful. And if there are any questions, I can take those now. Thank you. I'm just still reeling from the shock of all the things that you shared, I would have never realized that software could cause such energy consumption gaps. But from me personally, I don't have any questions, but I'm just like, you know, still absorbing everything you have said. And I think the more we share this, the more people will actually become aware because yes, a lot of myths exist in terms of polythene and polythene use and plastic use, but all of these hidden energy consumption, it's just are totally hidden. Sure. So Ritu Parna has a question, is there any organization who's leading on the software emission front? Yeah, Ritu Parna none, everybody's in the same boat, everybody's just started. So all of them are similar. If you if you look at it, every single organization has signed up for the net zero piece anyways. But if you ask me personally, and this is my opinion, has anyone truly achieved it? None. They all are in the same boat. So there's a lot to be done, research is happening. So it's for all of us to contribute. Some may be marginally ahead than others, for example, their website development may have brought into this. But that's a very, very small part of the organization. Okay, with that, thank you so much, it has been a such an eye opening talk today, and it will definitely make me look into it more. So thanks a lot. Thanks everybody for joining in.