HPE Discover CTO keynote by Fidelma Russo - From hybrid by accident to hybrid by design
what we're really trying to achieve here is to modernize from Edge to Cloud we are becoming a datadriven AI focused organization this is an industrial revolution and it's the fastest one yet you will not be put out of your business by AI you'll be put out of your business by somebody else that uses AI faster the hybrid Cloud solution has helped us build that resilient environment becoming a datadriven AI focused organization has allowed us to really change the face of gaming HP Green Lake has really taken the burden off of our company to provide insights that can actually change the way Healthcare is delivered for moms across the world what we're creating here with HP is a smart City a datadriven environment but it's equally about how we're taking that data and turning it into critical intelligence [Music] please welcome Executive Vice President general manager hybrid cloud and chief technology officer Fidel [Music] [Applause] maruso toos bentos at Discover Barcelona good morning everyone I hope you had a great discover and I hope you really enjoyed the fiesta last night and I truly appreciate you coming here this morning to listen to the keynote so in a world increasingly defined by datadriven insights and intelligent automation we stand at the threshold of a remarkable transformation and at the heart of this transformation stands Innovation ours and yours so this morning we're going to have a lot of fun talking about how you can unlock your Innovation using the HP hybrid Cloud portfolio you'll get a good sense of the compelling capabilities of the HP Green Lake platform some of the innovations that are happening within our hybrid Cloud portfolio and within the CTO office here at HP you'll also get a glimpse of what's driving our road map for the future and here directly from some of our customers who are using the platform from Edge to cloud and also using it for AI now as you heard yesterday we now live in a world where AI will enable us to create new solutions to some of our biggest societal challenges and accelerate business transformation across every industry and HP is playing a prominent role in this shift we're leveraging our unique ability to build Supply and support AI native architectures and we bring it all to you on the HP Green Lake platform and drive differentiation through our hybrid Cloud portfolio first let's talk about why a unified platform is necessary in the first place and what problem we are trying to solve yesterday Antonio talked about the digital transformation that many of us have been on and many years ago 20 or more all data was in a data center or a collocation facility and as technology has evolved so has the ability to process and store data generated by Edge devices and in fact there's been an explosion of devices applications and and workloads at the edge and today applications and workloads are highly distributed resulting in more than 50% of data being created at the edge today's world is decidedly hybrid and that introduces new requirements a hybrid Cloud environment delivers a cloud experience everywhere no matter where the data and applications reside this consistent experience provides Enterprises with the agility and flexibility to adapt to diverse operational needs it ensures seamless visibility and control over data and workloads and additionally the hybrid model facilitates simplification and automation of processes streamlining complex workflows and enhancing overall operational efficiency and to truly leverage and drive differentiated business outcomes in this hybrid world you will need to securely connect with your data across all locations process that data in real time no matter where it is protect the data across the Enterprise and then orchestrate your workloads across your Edge locations private clouds and public clouds at HP we recognized early on the need for one unified platform with services to run manage and secure your distributed Enterprise and we've been busy building partnering and acquiring services and technologies that Empower our customers to manage their hybrid Cloud environments so let's take a look at how One customer is leveraging the power of the HP Green Lake platform to address the challenges in this hybrid [Music] World a be food group is one of Europe's leading food manufacturers we spread across 15 countries in in Europe data is really an asset for us we have a gold mine of data in our factories with data we can optimize processes we can see bottlenecks in our logistic processes we can enhance all these processes we can track where we have issues you need platforms to be able to process all this data in today's complex world we have work CLS everywhere we have workloads in a public Cloud we have workloads on Prem we have workloads on edge and it makes it difficult for the IT department to manage all all those workloads with HP Green Lake we have really one single point of entry one single Source where all our infrastructure is managed and this covers everything it covers our our global data center but also the small data centers we have across Europe it covers the edge it covers everything and that's really huge for our people we are really in good hands with HP great so with that I'd like to join me on the stage to share their experience the head of technology and digitalization of the Bell food group Valerie [Music] [Applause] Thomas hi thanks for joining us thank you had a pretty tough Journey here got delayed but yet made it to the fiesta and made it here this morning so welcome welcome thank you so uh people may not know a lot about the Bell food group uh but you are actually fairly uh you know in many many countries and locations in Europe so can you tell us a little bit about the food group and the challenges you were all facing and why you chose a hybrid Cloud Model yeah sure so we are one of leading food manufacturers in Europe as I said in the video uh we have a 64 locations across 15 countries in Europe uh and we're producing a lot of different uh kind of food so we're starting with fresh meat uh sausages uh Spanish ham uh that we're producing in Spain in our new Factory uh we are producing also vegetables salads uh fruits uh pasta pizza so a lot of convenience products so we have a really wide range of products and uh when it comes to to hybrid clown I think we have two main challenges uh the first of one is technological so um we have a lot of system that need uh small latency right so on the production on the shop floor you need really small latency so uh public cloud is not a question it's not really uh possible um we need to have workflows on the edge we need to have workflows in the factory so um our design is uh that we have one Central data center where we manage all the glob mobile applications and um small data centers on the edge and uh this is this is one challenge making sure that everything works um and the Second Challenge is uh people uh finding I think a lot of people here uh have the same issue finding the right system Engineers today is really diff difficult I mean we're in Switzerland we are competing with uh pharmaceutical Industries with banking so much bigger player than than we are uh so finding the right people difficult and keeping them is difficult so if you find the right people with the with the right uh knowledge uh they don't want to operate the infrastructure right they have they want to to do you nice projects they want to have fun they want to do engineering so um yeah when it came to to renew our infrastructure in in the life cycle um we discussed with the HP team and we we saw that the HP Green Lake uh hybrid Cloud Model was actually a perfect fit for what we were doing and so You' you've explained a little bit Valerie why you chose it but what does kind of allow you to do now with the rest of your infrastructure I mean um we are producing really Fresh Products we are when we're talking about fresh meat or fresh salad it's something that comes in in the morning and that has to be delivered to the customer to the supermarkets on the same day right so uh if we have a massive it outage and the production is tops it means the the the risk is you go to the supermarket in the morning you want to buy fresh meat but there's no meat right the shf is empty so this is this is really a huge challenge making sure everything works making sure that uh the infrastructure is reliable that it's stable and um and yeah working almost every day of the year and uh yeah that's that's that's really a huge Challenge and um of course the hybrid Cloud solution as opposed to the to the old way of buying Hardware uh you're much more flexible right you're scalable uh you can and you can have the workloads where you need them you can have the workloads on the on the factory um and last thing yeah we decided to go for the manag services uh exactly because of this challenge of the of of the talents of the of the system Engineers because uh yeah we wanted to have one partner managing our basic infrastructure and making sure everything run runs smoothly yeah and I think was interesting the other day when we were chatting it doesn't really come to mind that if there's an IT outage there could be empty shelves at the supermarket so that just shows how the whole world is connected now and now that you have your infrastructure covered uh and you can manage your workloads how are you thinking about Ai and AI in the in the context of your business I mean even in the food manufacturing which is not a real digital business uh when you and when you think about it but even there AI can be a real game Cher uh it can be a real competitive Advantage so I won't disclose what we're working about but um yeah we are looking about a lot of Technologies um first of all is uh computer vision for product recognition for quality insurance where we are working with the HP uh colleagues together uh the second one that comes to mind is the predictive maintenance because yeah we were talking about outages you can have outages in the it but you can also have outages on the production right on the on the machine so if you are able to predict or to prevent these outages you you get better and uh the last of the the last is of course uh large language models yeah we discussed with the with the HP partner Alf Alpha uh we are already discussing some yeah some interesting projects that we could do yeah oh that's great that's great and so how would you what would you like to see us do in HP Green Lake with regards to a for the future that's a really interesting question um I mean i' I've worked with a lot of data scientists who we probably all did uh really brilliant people uh they have a lot of great ideas really that could can really make us reach the next level right but um as soon as their model is working it's like they're running out of the building to find new challenges to find new projects yeah they are not interested in operating the system but uh you have to see if if if an AI model is implemented into a production process it has to work it has to work 24/7 and uh I don't know any data scientist who wants to stand up at 4 in the morning because the pro the production process is not working so uh definitely we need someone to be able to do those operations so mlops would probably be a really interesting thing for us in the future okay that's that's great great input Valerie and so uh thank you again for coming and joining us here and uh thank you for the difficult journey and I wish you all the best in the future thank you fora thank you thank you so as we have seen Bell food group is using hybrid Cloud environments to manage and optimize their workloads and data and at HP we have built a hybrid Cloud portfolio that is able to support our customers in their transformational Journey it allows our customers to modernize their it Estates by offering flexible and scalable confusion Storage Solutions running in bare metal virtual machine or container configuration it allows you to simplify your it operations with a complete software portfolio that is multivendor and multicloud and we secure your hybrid workloads your data and your AI models now as Valerie mentioned one of the main workloads for the next decade or more is artificial intelligence artificial intelligence is a transformative force and the adoption of AI is an increasing imperative for companies looking to thrive in today's and tomorrow's business landscape it will disrupt nearly every industry its capacity to analyze vast data sets identify patterns and make datadriven predictions has already begun reshaping traditional business models it introduces innovations that optimize processes enhance decision making and unlock unprecedented opportunities for growth and Innovation it's also changing how we live and work as AI Technologies Advance they redefine the way we interact with information they automate routine tasks and we make better decisions and as we integrate AI into our homes and workplaces its influence extends Beyond mere automation fostering a new era where technology truly augments human capabilities before the surge of Genera of AI in the last year deep learning played a vital role in tasks such such as image and pattern recognition and this has actually served as the back bone for diverse applications including defect detection medical image analysis and object detection in autonomous vehicles and its foundational role set the stage for the emergence of generative AI from the classroom to the boardroom generative AI has democratized access to artificial intelligence across various sectors of Our Lives education leverages gener of AI for content creation making Learning Materials adaptive in finance it refines algorithms and trading strategies optimizing decisions an agriculture benefits from crop planning optimization elevating yield and efficiency and in healthcare generative AI enhances Diagnostics through medical image synthesis these examples that we can all relate to underscore the versatility of n AI bringing Innovation and efficiency Within Reach and makes this the transformative technology of Our Generation AI from machine learning to Genai will be ubiquitous AI is critically dependent on data and it will run everywhere from Edge to Cloud it is a hybrid workload and it must must run and a hybrid Cloud infrastructure and the HP Green Lake platform is the only hybrid Cloud infrastructure that allows customers to deliver on their AI strategies while modernizing their it infrastructures simplifying their it operations and protecting their hybrid environments from end to end and as I talk with customers who are all under pressure to deliver AI workloads to their business several common questions emerge the first question that's asked is how do I ensure that I am using gen in an ethical and Safe Way the next one is how do I pick the right infrastructure once I have the infrastructure what is the right management and operations software stack and today I'd like to unpack each of these top- of- Mind questions and explore how HP Green Lake is working to solve these problems in an integrated holistic manner we must start first with how to ethically and safely use geni to adopt gen at scale Innovation is not just desirable it's imperative and companies need forward-looking solutions to address the complexities of data PR privacy then you want a hybrid deployment because we want to be able to use both public Cloud models and also models that we train on Prem we also want the ability to integrate these geni models into our business processes so in the CTO office at HP our Innovation team has been obsessively focused on creating a HP Green Lake Cloud application to enable our internal devel developers to safely use generative AI models and I want to give you a little peek behind the curtain on a project that they have been working on Project Ethan is an application that is equipped with robust guardrails it ensures compliance with data privacy reg regulations and it facilitates smooth and secure AI interactions from Simply using chash to using API to create new use cases it enables the use of various models either on Prem or in the public cloud and it removes the complexities in setting up the various environments depending on where your model is running but finally because it is on the HP Green Lake Cloud platform we deliver authentication authorization and metering out of the box project Ethan is the typal of the spere where the AI Journey Begins so now that you've decided on your governance you need to determine the right infrastructure and software for your AI workload and again this is not an easy answer and we must get grounded in the various stages of the AI life cycle in order to answer that question and the first thing in the life cycle is we need to develop or we need to choose the model we're going to use then we train or we retrain the model with large diverse data sets so that the model learns to recognize patterns and information then we tune the model so we can get the Optimal Performance for the context that we are running in the next thing that happens is prompt engineering which I'm sure all of us have heard about as the next career for everybody and this is where us as humans have a role to play and that's where we craft questions to guide the model's output and finally inference is where it all comes together the trained model it's tuned for a specific context it uses particular prompts and it gives us the answer that we're looking for now each step in this life cycle has different infrastructure and performance requirements and each step costs a different amount of money so we have a lot of choices to make and the other thing is very few of us are involved in all phases of this life cycle so your infrastructure needs are heavily influenced by whether you are training models tuning models or using models and so the answer to what is the right infrastructure well it depends on where you are in the life cycle and the choices range from supercomputers for training to servers for inference and tuning to storage for your data needs now yesterday Antonio announced availability of the Enterprise gen solution stack co-engineered with Nvidia and preconfigured specifically for Enterprises to quickly fine-tune foundational models using private data that can be deployed anywhere an AI is a data hungry workload and one of the biggest bottlenecks to training and tuning is storage and yesterday we also announced a significant update to our HP Green Lake for file storage to address density and throughput demands for AI workloads engineered to provide high performance file storage for AI HP Green Lake for file storage is optimized to keep Pace with large-scale AI workloads and with the upcoming enhancements early first half of next year we will increase the capacity density and throughput by 7x HP Green Lake for file is the next Generation in the evolution of AI Storage Solutions providing performance scalability and efficiency you need to stay ahead in today's fast-paced fast changing AI landscape so now that we've ethically and safely decided and put our governance in place and we've thought about deploying our infrastructure it's really just the first step as Valerie said we now have to run it and so we need to manage observe and protect not just our data not just our workloads but also our models and we ideally need to do this from a single Unified platform and in order to make this real let me bring my colleague Dennis bford onto the stage to show you how the HP Green Lake platform is the hybrid Cloud platform for AI well Dennis can you join us here for a few minutes here I am how you doing fidela how are you doing all right good morning guys how are you doing yay great so glad you could make it cuz I was a little nervous at the fiesta last night you know oh but I'm glad you are up here bright and early and we're going to have some fun we're already having fun we're learning a lot and you heard fidelma talk about project Ethan and of course that also runs on Green Lake because at Green Lake we have a button for that so let's log in to Green Lake and see just what she was talking about here we have a couple of buttons for you we're going to go through in quick succession first let's start with the application you heard fidelman say that AI is really just a new workload but boy what a workload we're not talking about your Granddaddy's squl Z anymore we're talking about huge amounts of data and realtime processing one of the things you probably all played with is chat DPT show our hands how many people here tried chpt okay oh a lot of you great so isn't it fun how you can monkey with the prompt and get different answers yeah but at chat DBT you only have one model what we have created here is a god rail driven sort of wrapper around not one model but many models I'd like to show you that we can change the model chap DBT of course we have the offthe Shelf 3.5 turbo and also the chat dbg 4 with both runs in the cloud but what about what fidelo just talked about some of the more key workloads has start you run on premises well for that we're working with Alf Alpha and what there go guys have is this luminous base model let me show you how easy it is to change models first well we have chat dt4 chat dt5 and I just set just said that let's change the model here we go chat dt4 let's grab that can you see at the bottom here it says chap tpg4 okay let's ask it a question a prompt as we call it okay so let's prompt it uh what does HP do and here's the answer straight from the cloud well we help people what oh it appears that we help people achieve their goals using technology pretty well said with server storage services financial Services the light pretty good answer so what happens if we change the model to luminous base and model is train that's actually trained for a lot more Precision when you look at that model let's go and give that a try you can see it says luminous base at the bottom let's go and say what does HP do same question same prompt what's the answer here it is very succinct ah what's the difference the difference is that this model is trained differently than the other model the other model is more has more sort of Flowery language more of a boast which way humans talk this is more a scientific answer and this is absolutely true because we are dealing with developers developers want to be using the cloud in their models they want to be able to actually log on to it and for that we have another little feature and the feature here is security we got to be able to trust what we do and to do that we have to generate the security token let me just show you that what should we call it by the way well what about calling it Barcelona 2023 okay so that's a that's a that's a label on my token I set the expiration period security feature again what if I leave I on to another project I don't want my secure tokens to lay out there for too long so I let them age out after 30 days so here we are generate a token boom can you see that string there that string is actually unbelievably long I couldn't remember it heck I can't even remember my own phone number right but here's the the point though that is what a a developer will use to prompt his or her model either on premises or in the cloud and do so securely because this unique identifier follows him or her it's Unique to them so another person can't steal their work can't go in and use a model they're not allowed to and so forth so this is security built into using a large language models so there the model there you can see here very nicely the the the key the security key is right there and I can discard it or create more and I'm only allowed to have two so that's one of them okay so now remember fidelman said this is the application but many of you here right show your hands how many people here are in offs how many people do offs Ops what oh one two three okay you guys all it right so you have to recognize we need to have some Gear Running at the location well well we actually talked to one of our customers and they said very clearly you know what that factory runs on the data in that factory that's certainly true too for new workloads so let's set up a private Cloud shall we that seems to be a reasonable thing to do we've already configured that you're familiar with private Cloud you've probably seen the demos of private Cloud before I don't have a lot of time here I could do the whole demo with this but let me just show you one thing that's very important you heard Antonio yesterday announcing a new server a new system you have a new system that's optimized for AI and what does that mean this is key to understand back in the day it used to be just a server little Dam a little storage we good to go today is predominantly GPU lots of GPU do you have the GPU and as you can see here yes we do those brand new servers the DL 380a has gpus in there oops I went a little too fast there they're AI native right we have gpus built in to the system certified together with HP and Nvidia okay so that was just the beginning server storage networking private Cloud now what else do we need well what about the data what about the fact that many of us are operating across many locations well we need a global name space and ESO data fabric certainly delivers that so here we have ESO data fabric again launched from Green Lake and we have the ability to go in and take a look at the different clusters we have and you can see right here on this little example we actually have two HCI hyper Converse infrastructure clusters on Prem gpus in them all ready to go turns out that we also have two other instances we had an NFS server so we can actually publish and people can share the files and we also have an instance that runs in AWS wait a minute so you're telling me it's not just on Prem but it's actually a function of many things how do I visualize that and the the way to visualize that from the perspective of a global namespace is using not table view but using graph view it looks like this global names space mapped out can you see this exciting what I said and you notice to the right of this we actually have different abstractions we have volume abstractions bucket abstractions topics and tables what are they for well volume is for file for those of you who work with files buckets for those you who work with objects if you're just want an object store where you get and you and you delete right and you so you you put get and all that that kind of person you want a bucket and then you also if you deal with streams all you want is a topic and for tables for things like no SQL no SQL like key value pairs for example comma delimited key value pairs all day long these abstractions are beyond what we typically have seen in it we're used to seeing lungs we're used to seeing file systems and here we have new abstractions all Federated together with a global namespace okay so this is not just one Cloud it's hyper Cloud as you can see the global name space is across your sites and it's across all your individual locations whether they are Azure a AWS Google Cloud anything all right so public Cloud private Cloud very much a hyper Cloud setup so that took care of your data so now we have it right infrastructure data what else but what about visibility you're dealing essentially with a stack here and that stack if you can't see the problem you can't fix the problem so let let's let's go take a look at Ops whoa wait a minute what are we looking at here we're looking at a mapping notice again these visual Maps keep coming up we're dealing with a mapping here of of of kubernetes containers well we can treat this whole entire deployment through a kubernetes dashboard all integrated with kubernetes API well what about if you say well not so fast I'm not ready to go bare metal Linux kubernetes I actually want to put that in a VM so to crawl before I walk can I do that and the answer of course is yes so think about a VM when it comes to Containers it's a rber around the container stack yeah and it turns out well that runs on Hardware too and it runs in clusters and in this case you can see two clusters sharing an NFS Mount which very well could be coming from the eso data fabric you begin to see that this sort of layer cake on top of each other one thing above the other above the next and we're building essentially a whole AI stack out of individual components into one tool chain okay good VMware of course has its own dashboard again API integrated so so now we are step four here what are we forgetting guys are we forgetting something yeah remember your data your models these models by the way can take a long long time to develop and train I don't want to lose these I want to protect them what's the best protection I can get if you say Back Store you half right the best protection is actually continuous data protection and Contin inter state of Attraction works this way I take incremental rights put them in the lugas side buffer compress them and send them to a Target now that Target can be anything it could be another site it could be another building on the site it could be somewhere else as my friend used to say there only two states of data dentist it's either here or somewhere else but that somewhere else could very well be in a public Cloud wait a minute so you're saying the public cloud is nothing but a location that's right and we're dealing with a multiloc Asian world as you saw for Delma show on her map so at by the way you saw right there 10:1 compression why do we compress because we don't want to tax the target with too much storage capacity and we don't want to tax the interconnect between the two sites yeah so if you look at this we're really building hybrid clouds here so this is AI across hybrid clouds some on Prem some somewhere else yeah it could be Colo it can be public Cloud it can be anything and here you actually have a Disaster Recovery site and an ash site let's take a look at that a little more detail here we have a a virtual protection group view of the world you can see the file server it's actually protected three different ways another way to look like that if it's a VM you can actually look at it this way and look sure enough we have a VM protected as this file server VM is protected locally inside the location and it's a Dr site and lower behold is also copied to a cloud location so think of it this way your incremental backup and data protection is an any to any experience you have the ability actually to send here to there because I know that if I been hit by a flood or fire here I have my data protected somewhere else okay great so when you look at that what is actually your recovery turns out to we proven that we have roughly 8 Seconds of RPO 8 Seconds of RPO using serto any to any across your entire distributed Enterprise cool so are we almost done yeah we're sort of done with the infrastructure but what about our developers they're the ones we're building all this stuff for what do they want to do well if you talked to your developers lately they treat what you are deploying as a service that service they learned from public Cloud vendors they don't need to worry about the hardware they just want to consume it and they want to consume a tool chain that is completely open and wide so for that we have the analytics unified analytics platform and what we do here is we actually keep track on what data services have you looked at what data sources have you worked on and what Frameworks have you used to actually do the work as you can see here there sort of cast the characters of the most popular ones some from HP and some from the internet there's a lot of services on the internet if you've seen lately a rundown of AI tools available H are you kidding there's hundreds literally hundreds of tools they're not all going to survive but that's typical in an early stage of a new technology life cycle which which is where we are right now so we need to have the ability to see what's going on across the tool chain how of velers are using and we can also see how much it cost we can see how much vcpu they're using how much Dam they're using and how much GPU they're using okay good so with that in mind what tool chains do we have and the answer is well we have a whole bunch I encourage you by the way to go out in the show floor and see this in more detail but let me just show you uh just the the the the Frameworks we have air flow here for example very popular with a lot of developers what about what else well spark also very popular with a lot of developers all integrated here at One Stop Shop one home base where you can actually curate and control which tools they're using and meter and monitoring it's very nice now you see here there's things like Coupe flow ml flow all things flow everything flows right but there's certainly also an HP there's the machine learning development environment and if you haven't clicked on that if you ever played with that we see this classic curve of machine learning the classic curve machine learning shows you the longer you train the less you actually get an error so this gives you a nice feedback how you run these batches by the way this is very analogous to the old Mainframe days where you ran big bat jobs and you keep doing it until you get 99.9% assurance that your AI model can actually do what it's supposed to with new data has never seen before if you're following things like Tesla and how they're doing self-driving cars this is exactly the work they're doing okay cool so let's wrap this up with these big models this is all about training big models so you let's go back one more time to Ethan and see what we can ask let's see uh let's talk about asking why do we love Barcelona are you good with that let's send that off okay so why do we love Barcelona well it's it's Rich history there a lot of nice people there's great food it's a lovely place I don't know about you but I love Barcelona so I can only say to you all thanks for your kind attention Viva barcalona so hopefully all of you got a great bird eye view of all of the Technologies we have actually not all of the Technologies we have here at HP uh but some of the Technologies we have here at HP in our hybrid Cloud portfolio and our AI portfolio to help you on your Journey so let me share with you a few updates on what we've been doing on the platform since the last time in Las Vegas and uh at Frankfurt discover in May of this year we closed the acquisition of opsramp and I am pleased to announce that opsramp is now available natively on the HP Green Lake platform so when you log in to the platform you can now subscribe to opsramp last year at Discover we introduced the sustainability dashboard and today I'm also delighted to announce availability of the HP sustainability Insight Center the next generation of the sustainability dashboard sustainability Insight Center has been integrated into opsramp to give you the ability to observe and take action not just on your infrastructure VMS containers and workload s but also on your carbon consumption across multivendor and multicloud estates this carbon Insight capability is especially important when you're working with a workload like AI the immense computational power required for training large language models often involving vast data sets and complex algorithms contribute substantially to energy consumption the same is true for inference as once these models are put into production and they start replying to queries large amounts of energy are again consumed let's just take a look at one frequently used llm it's estimated that 1,200 megawatt hours are used to train the model and then further tuning results in even more energy being consumed and it's estimated that in one day alone the model consumes 250 50 megawatt hours which is actually equivalent to 10 football matches being played at FC Barcelona at Camp new and this is where the sustainability Insight center now powered by opsramp can help discover Monitor and track the carbon footprint of the infrastructure that's running your AI workloads and it doesn't matter if you're training fine-tuning or influencing a new model further in enhancements will provide insights on the adjustments you can make to optimize your energy usage so stay tuned for more developments in this area so with that I'd like to invite one of the customers we've been co- innovating with within the CTO office to join me on the stage Sunni Bastrop from danfos and to share some of his journey on sustainability with us but before that let's roll a small video to hear about the Dan fos mission we know how to build an energy efficient future in District energy but we need to work together by breaking down barriers collaborating and sharing ideas to truly optimize resources and transform how we live with Dan fos lean heat we bring Decades of experience with state-of-the-art software cloud and AI [Music] solutions to empower you to unite siloed processes and leverage the entire District Energy System so we can use intelligent data to plan build and monitor the heart of our growing cities from people to production together we can create healthier environments and a Greener future lean on us for one holistic [Music] solution so please welcome Sun Bastrop to the [Music] stage [Music] Sun welcome to Barcelona thank you felma it's great to see you again thanks you too so uh can you tell us a little bit about uh Danos and what you do just a little bit more color than the video right absolutely so we are a family-owned manufacturing engineering company uh and we work with decarbonizing some of the most heavy Industries right so we're in Mobile Hydraulics so when you see a comb Bine or tractor out there on the field two3 of the value in some of those machines that's our product inside mechanical electrical the software systems and those machines right construction uh when you see cranes dumpers we in Heating and Cooling business so cooling everything from residential facilities like you know Conference Centers like this houses data centers heating you saw the movie right um and then we are in Power Electronics so everything that literally moves when you have a bagage spell in the airport or when you're in the elevator um and power modules for electrical vehicles and wind turbines so pretty wide portfolio so uh and and for me I'm I'm head of it and I also lead our digital uh organization so a good mix of classical it and digital transformation so big machines decarbonization not much machine learning in AI would that be a true statement or am I completely off the mark I think we've been working with AI and machine learning for some time right and it's a it's a lot of use cases you you saw the movie here right um but it it's also it's really deep into our value chain so it's difficult to do AI on a system like this you have the Heat Factory you have the Grits uh you you have the residential at the end so we have hundreds and thousands of of sensors here and and the data that we pull in is something that we cannot compute with our mind right um we couple that with external data sources it can be weather data so that we can preheat or precool and and by that we can actually cut a significant amount of energy in those systems and at the same time we can deliver Superior Comfort to our customers and yeah I think other examples would be autonomous vehicles uh that we work on with off-road uh cases right or it would be supermarkets where we balance also the cold chains in in supermarkets and what do you uh how what's your view on gen and how applicable is it to Danos uh as other Industries are contemplating you know improving their businesses with it how do you see both your business opportunity and using it you know within the it organization so so for us it it gives a ton of advantages right we we would not be able to run some of those Advanced systems uh without but I also think for for us as as it leaders right it we bear a huge responsibility we just heard heard about the energy consumption that you mentioned right and when we look at some of those systems applying AI we can really do a big difference and we've been collaborating also on optimizing the energy ecosystem around data centers together where we with some of the Technologies are actually able to cut up to 50% of the cooling energy required in a data center compared to traditional uh cooling Technologies and we can use the excess heat coming out of those data centers right that it doesn't go to waste and then we can make sure that all of that is integrated into to the dashboard and and that we can really optimize the systems that that we work with so there is a lot of use cases for geni and do you think uh and I'm assuming that you are also working with other businesses to help uh make sure that you know the danfos solutions are integrated into solving their sustainability challenges with geni absolutely so for us sustainability is a steam team sport right that goes across the full of our company and we put up requirements for our partners as well right so when it comes to the effectivess in within our company we work with our partners to make sure that we take Optimum usage out of their platforms and then we have the other use case for AI which is what we just saw right it's where it gets deeply integrated into our Core Business and here we need to work much closer with our partners because we need to understand that you know our mutual businesses the ecosystem of our customers and and it's deep deep deep Integrations right and it's long partnership so it's important for us that when we collaborate on these things that we do that with partners that we really trust and with Partners where we have a common value proposition right in the market so it's an interesting Viewpoint uh Sun I think uh what you're saying is as we go down this journey of geni it's really important that as we're kind of carving new ground that we work with partners that we trust and that we've kind of got long deep relationships with uh because we're we're going to be forging new things together absolutely absolutely and what what are you last parting words uh what what advice would you give to the audience or what do you want to see more from HP as we go over the next uh couple of years well already today we are very satisfied Green Lake customer right so uh we have a number of of Green Lake instances running um and I think with the announcement that we heard from Antonio yesterday on the AI part and also what Dennis just shared with us it's very natural for us that we'll extend the Green Leg usage for us in the company into the AI uh space because there is nothing more natural than placing those systems just next to our Enterprise systems that's where we have all of our data uh and that's where we need the speed uh between those systems so I think we'll be doing a lot more together on AI going forward oh that's great well it's great to work with you and uh we really do actually appreciate all of your feedback that you give us you know every day probably uh and working with our teams in the field and in the CTO office and across the company thank you thank you thanks great partner Sun so as you've heard from Dennis and you saw in the uh demonstration HP Green Lake has a comprehensive set of SAS offers to manage observe and protect your data and your AI models so with those offers you can choose to either subscribe to them individually ually but we've also added an option to integrate them into a single bundle with a single subscription so we can reduce your cost and complexity so with that I'm happy to announce the HP Green Lake hybrid operations software suite and the hybrid operation software suite brings together our esmeral software opsramp with the sustainability Insight Center integrated and our data protection sueg for disaster recovery and backup this is a worldclass hybrid Cloud platform-based SAS portfolio and enables you to manage observe and protect your data and your workloads from Edge to Cloud so as you can see and you've heard here today deploying AI requires a combination of infrastructure and software ideally tuned to the life cycle and the part of the life cycle that you're in it's actually pretty complex and this is why we're introducing our private Cloud solutions for AI by extending our private Cloud portfolio to introduce Solutions specifically designed for AI workloads we are bringing together infrastructure and soft software tuned for AI and this private Cloud solution for AI is based on our newly announced Enterprise Computing solution for Genai our newly enhanced HP Green Lake for file storage HP Ops ramp and zerto cyber resilience Vault and all of this is available through HP Green Lake Flex Solutions delivered as a service now you may have noticed that we included zerto cyber resilience vault as part of the solution cyber resilience continues to be top of mind for Enterprises and the negative side of AI is that the bad guys will use AI to increase the speed and effectiveness of their attacks and the vault is your protection against that so let's run a quick video to demonstrate its capabilities ransomware protection is vital but prevention alone is not enough if an attack happens how quickly can you recover introducing the zerto Cyber resilience Vault an isolated zero trust clean room combined with an immutable Data Vault that saves the day when all else fails even better Zero's Vault uses best-in-class Solutions every step of the way HP Electra HP rant HP Aruba networking and of course zerto the leader in resilient data protection bottom line rapid Air gaap recovery that protects detects and responds to the worst ransomware [Music] attacks so my next guest is with NT data he is deep it knowledge and experience and he's helped modernize some of the largest global service providers in the world he's helped lead digital Transformations across Europe and he's now come full circle back to his hometown in Barcelona please welcome Alberto Valero to the stage to share his experiences with us and impart some wisdom on his journey to where he is today in his career Alberto hello how are you you very good thanks very see us so uh you have uh we had a great conversation last week and uh you have been um you've had a great career and you're a thought leader in a Visionary in the cloud transformation Journey with many large service providers which means that you've you know seen it you've been there done it and seen it all and so you know we've talked about a lot here today but as you reflect on what you've seen and what you see going forward can you tell the audience what you think is the single biggest challenge for cios today sure felma but first of all let me say to everyone Ben Binga Barcelona I hope we are treating you well and giving you the warmth of our people so now you just need to give me your warmth back thank you now I feel better you feel better yes that's good so basically H it's the complexity is the lack of operational integration and I'm talking about operational data so cios are a struggling with the disparity of tools delivery models Silo operations uh hybrid uh environments hybrid reality of evolving portfolios which are actually evolving towards Cloud becoming more complex containers microservices so that complexity is the biggest challenge for for uh cios today so you have looked at a lot of tools over your career and you came you know we met uh as we were looking at the acquisition of Ops ramp and and tell me why you chose Ops ramp you know in in many of your uh your different um you know parts of your career with uh service providers what does what does it give to you how do you how does it help you and and also like how do you see it different to other tools in the marketplace sure so we could tell that opam was a cloud native solution very powerful in the cloud and he had a clear mission to simplify the complexity organize the chaos and it has three distinct H traits number one it's a true aiops Cap N to end capability there are very few ones that can have that end to end capability from from observability going through big data and machine learning at the core and ending with insights and automation it might not be the best at each and every step of the process but it's definitely the one that covers the full spectrum that is required to provide realtime insights at the end that's number one number two it's about the speed of adoption so Ops Ram is very easy to on board new customers is especially good for multi client environments and that's very important for M manag service providers like entity data and it's also very quick to deploy and the time to value is hard to believe so it's a bit too too too good to be true is what we need to tell the customers every time we we face that resistance we just need to show that actually it gives the results in weeks instead of months and number three and most important they can do attitude of the partnership so obam has always been committed to make you succeed in the marketplace they are not just thinking about oh we need to sell to you but they are thinking how can we make you more successful and that can do attitude it's uh unparent it's great to hear it's always actually about the people and the culture in the end we that we'll figure out the technology so what advice would you give to cios on gen hm J everyone is talking about that is the great hype of today and yes and everyone is thinking that is going to make magic or all of on its own well let me tell you something my friends here in it operations we cannot take shortcuts and you need to go or we need to go together through the logical steps of evolution before we can get the most out of H generative AI so first of all take control of your operational data ingest all the data manage send all those mountains of data into your data Lake of Ops Ram then you can apply the machine learning algorithms let them do the magic you will then get the the Holy Grail of ncio which is to actually sit down under a palm tree enjoying a p colada and looking at the V real time Vis visibility of the full entire H digital uh state that they have to manage and then once you get there you'll be able to actually multiply the effects with generative AI when it comes to the chatops space with natural language processing that's great so uh clearly you have a lot of experience but a lot more to impart and uh I wish you well back here in Barcelona and thank you for sharing your insights with us thanks Alberto thank you byebye so the overarching theme is clear AI is not just a technological leap it's a cultural shift it's a shift that impacts how we work how we innovate how we secure and share data and how we run and transform our businesses before today you may have thought artificial intelligence was going to solve all your problems but in fact it's clear you need a trusted partner who can help you you solve the complex problems that AI brings the HP Green Lake Cloud platform helps you manage observe and protect your hybrid infrastructure data and models from Edge to Cloud HP Green Lake is the hybrid Cloud for all thank you and Viva Barcelona gracias
2023-12-09 22:26
