AWS Summit Singapore 2022: Reaching digital transformation excellence (Intel)
- Hi everyone, welcome back from lunch. The perfect time to put an Intel slot in the day, isn't it? And again, if you guys wanna come closer, it's a smaller room, so come over and let's really just idea it together more than anything, it's not really meant to be a keynote. But as per the topic that isn't showing up, which is now, as Adrian mentioned, I lead the alliance for AWS at Intel for the region of APJ. And I also lead Intel's go to market strategy for AI and HPC with AWS.
So you might be wondering why is Intel here talking about the partnership with Intel, the partnership with AWS? Why are they here talking about digital transformation and your path to excellence on cloud? With us in the industry time and time again, what we've recognized is when you have a technology that's accessible, that's consumable, and when the barriers office adoption are low, this gives rise to a tremendous amount of demand. Through that demand, rise various business models, various usage models, and ultimately, a lot of opportunity. Cloud computing continues to bend the curve on what seems possible. And with us being a silicon as well as a technology provider, we do seek out partners as well as providers of such incumbent technologies, and want to build relationships to really drive that journey for our customers, regardless of the infrastructure on the right choice. AWS's cloud transformation is powered today by Intel.
And the rest of my session I wanna talk to you about the ways in which accelerated technologies like AI are really helping us see the way in which technology is driving impact. We will also talk to you about the way in which our partnerships have helped harness the greatness of success stories. We also have a surprise couple of guests with me, and I'm not gonna take too much away from the curtain razor on it, but it's gonna be a great, great session where you will learn more about how technology like cloud can help you with your digital transformation journey. As I mentioned, just very lightly alluded, I wanna talk to you about our partnership with AWS.
What is Intel actually doing with cloud? And what value with our partnership we can offer as actual support to help you solve your business challenges for your organizations collectively. I will also talk about a technology I think that everybody about which is AI, and what you can do with the data that's available for you at your edges or at your customer's edges. And then talk to you about a fantastic use case and a success story. And I will tell you more about it, although I've given already the name of the customer out, but really it's gonna be a story about the journey that Intel, AWS, Minfy, as well as POS Malaysia took to making supply chain as well as postal services actually come alive.
So what does Intel do with AWS? As I alluded to it, we customize our silicon to be made available on Amazon's EC2's scalable platforms. So how many of you guys use EC2 today? I'm assuming most of you do, you just don't know it. I guess your solution architects build it. It's really the platform or the workhorse that AWS has that actually supports all that circle of services that you could see on my slide. Basically through EC2, you're able to transform your cloud journey across use cases and business challenges that you want to solve for. So whether it's the elastic compute environment or you're storing your data on the cloud, or you're actually trying to migrate some of your databases to AWS's world of cloud again for those databases, or you're trying to do more stuff around HPC, around AI, around IoT or Edge, there's a whole lot of these services that support that journey for you.
And on each and every one of these services, there's Intel that follows or is powering these services because we have Intel hardware instances that are supporting each of these services under the hood. Why do we do this partnership with AWS? We do feel that we have common values as well as focuses targeted towards our customers. We have now over 16, actually this needs to be updated. This year is our 17th year of co-engineering with AWS. We work with AWS's engineers, we work with AWS's product teams to ensure that the products that come out on the cloud are best optimized with Intel's technology and capability. We also see that we wanna collectively partner with AWS on our customer as well as our partner's digital transformation journey.
And I think that we do not compromise on the way in which we wanna work with our customers. And we are super passionate about those use cases and those challenges that we want to solve for our customers together. While we know that cloud computing is an incumbent technology, there's three things that Intel brings to the table that actually makes the power of cloud come alive. Number one, performance. Number two, performance. Number three, guesses? Great, you already know what I'm talking about.
Critical compute capability, and the performance of that cloud is actually going to make you get your cloud experience come to life to help you with that digital transformation journey. And that's really what I will be again covering in terms of capabilities of what AWS can drive for you. If you look at the slide behind me, we have over 250 plus Intel powered instances that are sitting across all of those platforms and services that I just went over earlier.
The reason we have that is for getting you the choice as well as the customized experience for your application needs. So whether it's critical compute that you require, whether it's storage that you guys are exploring on cloud, whether it's memory optimization, whether it's just general purpose optimization, we've got a set of instances that power EC2 to make those happen for you. And the reason you want to see these and take advantage of these is because these are pre-optimized and pre-engineered for you.
Hence it takes away the requirement for you to have to go under the hood down to that level to optimize the features that are built on our hardware to actually make cloud come alive for you. The instances that we're super excited about that we shall also talk about support accelerated computing services like AI as well as HPC. And I'll tell you more about that.
But it's really about making sure that whether you guys know that you have Intel on under the hood or not, that if you are using the intel instances, you are getting access to that performance that I talked to you about. Some of the cool things that we've done together with AWS, things that we're super excited about are the types of use cases that we will also showcase today. The work that we're doing with Career Launcher, which is an educational service provider, actually maintained and kept education uninterrupted at the time of pandemic.
So much so that even post pandemic we have close to about 2.5 million students that are learning K to 12 grades and curriculums on this live platform. And we have multiple such examples like that even in the parts of ASEAN.
We also think that cloud is transforming healthcare. The work that we did with the University of Sydney as well as with Cairo showed us that we can actually mitigate bushfires powered by cloud. Again, the work that we're doing around HPC where simulation and and modeling of simulations actually can help you harness the capabilities of high performance computing and also have the choice of having to support that and build infrastructure on cloud, where we've done more effective modeling and simulation with parallel cluster, which is a AWS's HPC environment with select solutions to do more simulation and modeling for environments around transportation, around automotive, energy, healthcare, and what have you.
Another great couple of instances that we have launched are the M5zn instances that are the fastest Xeon Scalable CPU instances on the cloud today because of the Turbo boost as well as the NVMe storage capabilities that we have in there. And just about a couple of months ago, we launched again the next set of instances that are powered by the Ice Lake Xeon-based technology that I will talk to you about as well. So, co-innovating, co-creating, and co-launching instances that make applications on cloud come alive seem to be our regular DNA along with AWS as well as AWS's partners. I mentioned to you the Ice Lake instances.
The workhorse of Intel's hardware is based on the Xeon processor family, the Scalable Xeon processors. And we innovate and have availability gen-on-gen on AWS's environments across those software and application-based instances that I just talked to you about in a couple of slides earlier. So the Ice Lake instances, as I mentioned to you, our theme and our focus continues to be performance is to ensure that the features that are built onto our chip are actually passed over to you guys as advantage. So whether it's the high workload performance, whether it's the deep learning boost capabilities, the fact that we have AVX-512 instruction sets that help you better vectorize for VMUs and BMCs or the security features, help you get that benefit in terms of the performance. I'm not gonna read out those performance numbers, but we all collectively know higher the performance, lower the TCO, hence it saves costs and it gets you to your digital transformation journey faster. Another cool instance that we are super excited about talking is the fact that not only do we have optimization and engineering and technology on the CPU side, we also have it for the GPU side.
So all of you guys that are doing AI, that are doing computer vision stuff and natural language processing stuff and doing video analytics, you guys probably use some GPU instances on AWS today. If you want a better price as well as a better performance, you can try out the DL1 instances that are the Habana Gaudy-based instances that are four accelerators built onto a GPU instance, gives you a 40% better price as well as performance over existing GPU instances, and make sure deep learning journey come alive. So we're super excited about that as well. Probably need some water. Thanks.
And just to summarize, before we get into the fiesta resistance of today's session, I wanna walk you through the various technologies, hardware and software, again, that are powering the AWS environment. Thank you so much. It's probably all the lunch. Sorry about that. So this slide is a very busy slide, don't worry. I just wanna encrypt this and decrypt it for you.
I wanna talk to you about the fact that we're not just hardware optimization, but we're also software based optimization. You guys are AWS or cloud users, practitioners, connoisseurs or trying to get into it, obviously all the innovation that you guys do is on the software layer. We wanna tell you that we got you even on the software side. So if you look at, and I'm gonna go bottoms up. If you look at the hardware, you see that we have GPU and accelerated experiences that are available on AWS today. But in addition to that, when you've got cloud experts, when you've got data developers, programmers, analysts, you're also working on the foundational layer.
You guys are probably working on libraries. Intel software team, which is the third largest software team in the entire world across all organizations, help you get access to Intel distribution of Python, which gives you 80 to 90% plug and play performance. We have DAAL, which are the data analytics accelerator libraries that sit on every pipeline of your machine learning code and optimizes that for you.
And then we have DNN capabilities built into our libraries as well as our compilers. The reason I say this to you is that you get access to that when you try to optimize on AWS's AI services. If you're pulling in SageMaker, if you're using DLAMI, you'll be able to get that performance benefit just by picking up the Intel optimized version that's available for you guys at Marketplace. Moving up a layer, how many people work on TensorFlow, MXNet, PyTorch? Quick show of hands. Okay, maybe you forgot.
Maybe there's a lot that's happening at the engineering level that you guys don't know. But basically the frameworks are the way in which the models today are being trained as well as inferenced. We work with every one of those open source frameworks to get you again, that performance that I was talking to you.
And then we've got toolkits on the edge that actually make those use cases that you want that data that you have in your individual industries, drive that to insights using the entire fleet of hardware and software optimizations, including those toolkits that can make significant successes or use cases within those industry come alive for you. So I've talked to you guys about a whole lot of intel technology, it could be a whole lot. The key thing that you wanna take away from this is the fact that Intel's silicon, Intel's hardware, Intel's software that's powering AWS cloud today gets you that experience on performance for your use cases for your digital transformation journey. But this is not something that we can do alone. We are technology providers.
The only way that you guys get access to this technology is through our ecosystem. We rely very heavily on our consulting partners, our technology partners, and maybe some of you are those, maybe some of you deploy those, maybe some of you use those services. And when you're working on your IaaS, your PaaS or your SaaS implementations, these partners that work with us, understand our technology, train on our technology and build on our technology, come there and make this alive for you.
So one such journey is about a special customer that we have on stage or not yet on stage that went through this journey with us. Once we identify use cases that we want to really care for, once we identify segments that we care for, between Intel and AWS, we go out work with the customer, the customer brings us a use case and says, "I wanna be able to do better image recognition or image detection on cloud. Make this happen for me." Intel and AWS's teams gets together the best competency partner and helps build a pilot experience for you. Through that pilot experience, you guys learn, you guys train, you pass over the benefit, and you get that benefit that are partners Intel and AWS have to offer to you 'cause we have that replicable reliable learning from actually implementing those probably with other customers in that segment.
And then not only do we do the training, but we co-engineer using the competencies of our partners to make sure that that use case is doing what it needs to do for you. And this is really the journey that we hope that we can try and build with you guys along with AWS as well as our joint partners. This journey has been undertaken by one of our coolest customers that are in the ASEAN region. We have POS Malaysia here today. I talked to you about cloud computing, I talked to you about digital transformation, I talked to you about the way in which you can use accelerated technologies like AI to make your journey on the cloud for your business challenges come alive. Super excited to show you first that journey as a visual and then also invite on stage our customer, followed by the partner that actually made this experience happen where we will do a dialogue and then hopefully open this up to questions.
So let's quickly do our visual. - [Narrator] POS Malaysia Berhad is the national postal service provider and has a network of more than 3,500 couch points across Malaysia. POS Malaysia has been on a 12 months transformative journey to revamp its parcel delivery business. This included transforming legacy IT systems and operations to become more nimble and agile using cloud technologies, applications and databases.
- We have a number of challenges in terms of collection of who's who of legacy IP systems that we need to transform. We have talent gaps within our product and tech divisions that we're working to address. We also have a significant amount of risk associated with on-prem data storage that we've had all these years.
Right now we are almost through our migration journey where we've essentially move most of our service onto the cloud, and that's really critical part of it because it helps us get more agile and nimble. - Well, it was an absolute pleasure to start this exciting journey with POS Malaysia. One of the major challenges which we were told to start off with was to transform not just the applications or systems into the new tech, also to look at standardizing things the way IT works. So this was a very unique project for us for cloud migration where we were asked to add value even before the migration started.
- To help POS Malaysia to embark on their digital transformation journey post our discovery, we started with identifying the business drivers for the three core business applications. We engaged with them on four different streams. First, we started with taking over application support for about 30 plus applications.
And then with the due diligence and the data work we have got post our discovery, we migrated the scoped applications onto the AWS cloud. As a phase three or phase two, we started modernizing these applications after migrating them onto the cloud. - When it comes to solving and improving business outcomes for a vertical or an industry like transportation or automotive, Intel has continuously tried to innovate as well as bring the ecosystem of experts together, especially with POS Malaysia driving and harnessing the partnership with Minfy as well as the experts on POS Malaysia actually helped Intel ideate on its purpose around transportation. We recognize that the partnership around Minfy as well as the domain expertise that POS brings, will help us make a difference towards this business. - Recently we already engage with machine learning where AWS SageMaker, it allows us to give estimated time of arrival of parcel for the customer. We already engaged on AWS as well for EKS segment.
- I think we are all incredibly optimistic and excited about the future at POS Malaysia. We are transforming a 200 year old organization into one that fit the purpose in 23, 24 and beyond. We're completely looking at diversifying our products and services, expanding our portfolio, what we offer, and most importantly, aligned to our objective of becoming a community services enabler, which is really critical for us to play a significant role in Malaysia going forward. - At Minfy, we are on a bionic journey to build a better world for our customers. We focus on our customers, their needs and try to address customers' customers needs as well. Minfy is committed to helping both enterprise and public sector customers in ASEAN, such as POS Malaysia facing unique challenges to accomplish complex missions with limited resources.
Minfy as a cloud native system integrator is able to harness the power of Intel and AWS and bridge the gap between technology and end users. That's our vision of a better world. - So without any further ado, you've already seen this gentleman in the video, I would love to invite POS Malaysia's CDO and chief strategist, Sumesh Rahavendra, on stage.
Welcome, Sumesh. (audience clap) (upbeat music) - Hi, good afternoon. Thanks for having me on. What Akanksha really meant was when she said, "I'm gonna introduce you to my coolest customer," she tried calling them, they were not available today, so she called me.
So I think firstly, thanks for having us here. Thanks for having us here, really appreciate it. And, yeah, ready to answer your questions? - No, why don't you kick us off, yes, please do sit down. Happy to do that. You can stand as well. No problem, as you wish.
Would like for the audience, Sumesh, really to understand what POS Malaysia does and what your digital journey has been so far, and what you thought about the various ways in which, or the reasons why you wanted to go and adopt some environments or experiences on the cloud. - Yeah, sure. I think for most people who are not aware of POS Malaysia, POS Malaysia is the responsive post office of Malaysia. It's a 200 year old organization and essentially is an icon in Malaysia. Over the last decades, it's been privatized as a publicly traded company with still deep roots to mail and parcel primarily as business services among a whole host of other things.
So we've been on this journey over the last 12 months or so. Personally, I joined post Malaysia about 12 months now. And when we kicked off our journey, we'd obviously given a lot of thought, but we were sitting on, I think at that point it was something like 140 applications and a complete monolith structure, all of it on-prem in servers around the country and in our head office, and most applications being super legacy ones. And I was just checking with my team if we had any applications run on Cobalt. And so essentially, it was a highly legacy environment. So, our journey has been a very challenging one, but also very interesting one because being on-prem, being filled with legacy systems was a significant business growth inhibitor.
And the one thing that we believe in and I very strongly believe in is, you should never have a digital or a tech strategy for the sake of having a tech strategy. It should support what your business needs are. And at the point our business need was, we needed to provide a significantly better customer experience, which we were not able to because of the way our systems were structured. Our business priority was to significantly reduce our costs, which we were not able to because we had a high cost base with all of our servers in house and inefficient development and pipelines and all the weird and wonderful things. And the third thing the business needed was, we needed to get agile, nimble, build better technology for our customers, which again, we were not able to.
So you combine all of this and obviously the first thing among many others was the need to take all of these massive legacy applications and do a immediate lift and shift to the cloud without even getting into modernization simply because the flip side of that was we had a high burden coming up in terms of hardware costs that we needed to replace over the next year. And so we've gone through that journey for the most part, and hopefully soon we'll be complete. But beyond that, the journey only starts now. We are now on the cloud, but we need to re-look. We've got this 140 applications down to about 68, but even though 68 is probably 40 too many.
And so we need to re-look at the complete architecture, re-look at how we wanna build or modernize our tech stack and actually deliver what our customers want us to. - Amazing. Thanks so much for sharing, Sumesh. So tell us a little bit about specifically use cases like parcel delivery, customer experience and behavior for the future. You wanna make sure that your customers are happy, you're managing supply chain.
What are the extended technologies within cloud where you feel applying those into a cloud environment will help you get faster to your goal? - I think firstly, parcel delivery business, although seems really simple, is not quite as simple. It's definitely not building a rocket to Mars, but it is a fairly complex business in the sense that any given parcel goes through anywhere between 11 to 14 checkpoints through its journey to be delivered. And that causes potential complexities along the way.
I think one use case that perhaps I can already talk to, which was only possible because we moved to the cloud, and because we set up our data pipelines, and because we set up our data lake on the cloud is more recently, and it's a little bit crazy because as a customer, when you order something, you wanna know when it gets to you. I mean, that's almost a basic need. And in Malaysia, I'm not sure how many of you are from Malaysia, and I'm pretty sure it's similar in other Southeast Asian countries. When you order something online, your estimated time of delivery is anywhere between three to seven days. And in 2022, that sounds crazy to me.
Any other service that you order, you know when you're gonna get it. You order food, 37th minute it's gonna be here. And today in Malaysia your ETA is three to seven days. So one of the things that was very high on our priority list in terms of being a differentiator for the customer experience is actually telling you when are you gonna get your parcel. Like Robin said in the video, on top of our data today, we built an ETA engine and today with a plus minus three as accuracy, we can actually tell the customer when they're gonna receive their parcel.
And that technologies like that is only enabled because of this. I'll give you another example. We have really big ambitions to get into being community enablers. Because today we sit on a massive footprint of over 3000 retail locations in Malaysia, and we have over 650 post offices around the country, and that's a massive footprint where we can actually start providing a lot of essential services to communities and citizens in Malaysia. So one of the things as an example is we're looking at getting into health services.
We're trying to get into, because getting into something like health services means we are able to collect data that becomes essential to, as an example, the government. Having data on people, having chronic conditions data, be it hypertension, diabetes, blood pressure, and making sense of that data, identifying patterns, identifying predictability, enables the public health care system to become more robust, become stronger. So that's a future use case we're looking at.
We sit on top of 10 million addresses in Malaysia. There are only 14, by the way, 17. We sit on top of 10 and soon to be 11 million addresses.
Today, the concept of address in most countries is broken. And so how do you build, as an example, products like address as a service? Because we deliver to these 10 million addresses constantly. So we know exactly what their geo locations are, what their (inaudible) longs are, how do you build address as a service? And technologies like these are enabled by the likes of work that Intel, AWS, Minfy do, and how they set us up for success by building and getting the base technology. And that was my commercial and I'll take my commission now. (audience laugh) - No, I think the way that you put it should be something that's very beneficial to the audience here as well, Sumesh. We were very excited to work with you, I know AWS was.
It's great that you have almost an Amazon persona, but you're using AWS's services, so that's amazing that you're doing that. What's the future for you? What does it look like? I know that you're very passionate about community building, about expanding the solutions that you have on AWS that's powered by Intel into your existing ones, but you probably are going outside of healthcare, you're probably gonna get into financial services maybe, or telecom or anything. What's the future for POS Malaysia from a digital transformation journey on cloud look like? - I think not to get too ahead of ourselves. One thing at a time. So firstly, let's get our base systems right, is the most immediate future. I think we're not living in a age where you can go sequentially.
So I think we are building a lot in paddle and getting our base systems right is one thing. But the second piece is how do we gear for the future? And to your question, we wanna, because we have access to 95% plus of Malaysia's population and we can actually physically reach them, there's a lot that we can do in terms of enabling services for communities and people. So things like health is one vertical, banking for the under banked is a second one. How do you create opportunities where you can create financial services and products where people who don't necessarily have access to investments or saving or lending products, get access to it. So financial services, digital identity services, we already have a subsidiary that's working on it, so building digital entity products, and then at some point, hopefully, we start looking into a national digital identity product from Malaysia as well. We have a number of aspirations across the organization.
We have aviation division. So, we're looking at how we can build products in the space of cargo handling as an example. So lots in play, but we got to get started somewhere. - Agree, well, very motivational, super inspiring team. We will probably not have any time for questions, but I know Sumesh as well as the rest of us are going to be here.
So why don't you come over and talk to us post the session. Thank you so much for having- - Thank you. - Your time with us here. Wonderful having with you. All right, thank you so much.
(light hearted music) (audience clap) So the people that actually made this experience and this journey for this use case happened with Sumesh is Minfy. They're a very strong partner for Intel as well as AWS. I'd like to invite Vijay Jain, founder of Minfy on stage. (upbeat music) (audience clap) - Hey, thank you, Akanksha. - Welcome. - Good afternoon everyone. - Welcome Vijay.
- Thank you. - One question Vijay, nail it out of the park because we don't have a lot of time. Tell us about what Minfy did with POS to make their use case actually happen and what was the choice of, what was the reason why you had AWS and Intel be the power of technology for them? - Sure. I think before I answer why did we both choose cloud as a platform. So I think I'll look at the objectives. So when we met POS 12 months back as Sumesh mentioned, I think there were three key objectives.
So one was to drive innovation, second was to drive lower cost, and third was to improve overall staff productivity. When we look at this three key objectives, I think a lot of choices were in front of us, but cloud emerge has the most prominent choice to solve all these three objectives. And absolutely that was the whole reason cloud came into play. And how did Minfy address these needs and helped solve it? So we took three phases to it.
The phase one was primarily assessment. So we engaged with the POS business teams, IT teams and mapped the key objectives linking to the business outcomes. And second phase was the migration phase, third was modernization, and fourth of course as the innovation. However, as Sumesh mentioned, it was not sequential and we are running it parallel streams.
As of now we're done with the migration journey. We have already seen 40% overall savings once the migration is completed, that is phenomenal. The second objective for POS was to improve overall staff productivity. That's going to happen once the modernization phase kicks in.
When the apps the customer facing apps becomes cloud native, I think then the world employee staff productivity is going to get impacted. And the third objective of POS was to build a culture of innovation within POS. And I think once the migration is done, the modernization is done, once data's on the cloud, as Sumesh mentioned, once the data leak is done, I think then innovation is at play. - Wonderful. I'm sorry, I don't think we have time for more questions, but again it would be wonderful to dialogue with the team over here in the audience.
Thank you Vijay- - Thank you. - So much for being here on stage with us and sharing. - Thank you. (light hearted music) (audience clap) - Okay, you can keep it. All right, so audience, hopefully it gave you an idea of the way in which digital transformation with the power of partnerships and the availability of use cases that you want to explore with cloud can actually make that journey come alive for you. Quickly, to summarize, it's all about collaborations. Second point, remember those three points, which is actually the second point.
It's about performance for you on the cloud. And the third is the fact that you can use this technology, you can bank on this technology and go out and innovate for your products, for your customers collectively with the power of partnerships that we have over here. So I wish you guys the rest of a good AWS summit. Thank you for spending this time. I hope you find it valuable.
2022-11-16 08:38
